ACCUMBENS-PROJECTING LATERAL HYPOTHALAMIC MELANIN CONCENTRATING HORMONE NEURONS INTERACT WITH OVARIAN HORMONES TO MODULATE MOTIVATED FOOD-SEEKING By Lauren Marie Raycraft A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of Psychology – Doctor of Philosophy 2023 ABSTRACT Food intake requires a complex interplay of signals from central and peripheral systems as well as external stimuli, which are integrated across multiple timescales to coordinate feeding behavior. Individuals make numerous decisions about food each day, including what to eat, how much to eat, and when. Traditionally, research examining the timing of food intake has done so on a 24-hour scale, examining the influence of appetite-stimulating and satiety signals on circadian rhythms of feeding behavior. However, individuals can keep track of time on multiple scales including in the milliseconds to minutes range, a form of timing known as interval timing. This perception of brief intervals supports associative learning and the formation of predictive relationships (C. V. Buhusi & Meck, 2005). Thus, interval timing is critical for learning and decision-making. However, despite the role of interval timing in decision-making and the frequency of food-related decisions, few studies have examined the relationship between appetitive signals and interval timing. Appetitive signals, including the neuropeptide Melanin concentrating hormone (MCH), are predominantly produced in the lateral hypothalamic area (LHA), a heterogenous brain region characterized by its role in energy homeostasis. I previously provided the first evidence that LHA neurons that produce MCH (LHAMCH neurons) influence time-dependent food-seeking in a manner that depends on LHA subregion, sex, and estrous cycle stage. In particular, excitation of anterior LHAMCH neurons selectively prolonged motivated food-seeking in females tested during diestrus, the period of the rodent estrous cycle when levels of circulating gonadal hormones are typically lower. This suggested a role for the nucleus accumbens (NAc), a ventral striatal region critically involved in reward processing, and circulating gonadal hormones like estradiol. This dissertation extends these findings by examining the role of LHAMCH neurons that project to the nucleus accumbens (LHAMCH à NAc) on time-dependent food-seeking. Specifically, I separately examined the effects of chemogenetic excitation of NAc-projecting neurons from posterior (LHAp; Chapter 2) and anterior (LHAa; Chapter 3) subregions of the LHA in intact female rats. While chemogenetic excitation of LHApMCH à NAc neurons failed to produce behavioral effects on time-dependent food-seeking, excitation of LHAaMCH à NAc neurons influenced responding selectively during diestrus. Interestingly, however, these effects were in the opposite direction than expected: chemogenetic excitation of LHAaMCH à NAc neurons reduced food-seeking after the omission of an expected food reward. Finally, I directly examined the influence of estrogen on LHAaMCH àNAc neuronal excitation by ovariectomizing rats (OVX) and testing them with and without estradiol replacement. Contrary to expectations, chemogenetic excitation of LHAaMCH à NAc reduced post-criterion food-seeking in OVX rats treated with estradiol, rather than without. In short, the removal of peripheral estrogen through ovariectomy does not recapitulate effects of LHAaMCH à NAc excitation during diestrus. This data indicates that motivational effects of LHAaMCH à NAc neurons are sensitive to circulating gonadal hormones, including estrogen. Furthermore, these data indicate a role for LHAaMCH à NAc neurons in guiding decisions to persevere or attenuate effortful food-seeking after the omission of an expected food reward. This dissertation is dedicated to my Granmom, Marjorie “Marj” Raycraft (10/16/1928 – 3/17/2023). Whip-smart and well-read, she instilled the value of learning in all of us. iv ACKNOWLEDGEMENTS As it turns out, the acknowledgments are nearly as hard to write as the dissertation itself – especially given how many people have supported me throughout graduate school and the years leading up to it. First and foremost, thank you Alex, for accepting me into the lab as a naïve and enthusiastic undergrad, and then back into the lab as a PhD student. Often, it’s the small, unassuming moments that irrevocably change the course of your life – for me, this includes the decision to take your Neurobiology of Food Intake course, and that day I stayed after class to ask if I could work in your lab. Thank you, thank you, thank you, not only for saying yes then, but for welcoming me back into the lab as a PhD student and then investing six years into mentoring and supporting me as I chased this dream. I would, quite literally, not be a behavioral neuroscientist without you. Thank you also to my committee, including Drs. Alexa Veenema, Kelly Klump, and Gina Leinninger, for your steadfast support throughout the chaos and uncertainty of graduate school. Knowing I had a great committee in my corner was reassuring in even the toughest times. I am also grateful to have made lifelong friends in the Johnson lab over the last six years, including especially Jenna Lee, Kate Sapkowski, and Nicollette Russell. These women worked side-by-side with me through tears and dance parties and just about everything else you can think of. I love you all and am so grateful to know you. I would also like to thank Maria and David, Chase and Shay, and all of the other support staff in the building and animal facilities. These individuals not only made this v work possible, but also provided kindness and a bright spot – and even delicious baked goods! – on long days in the lab. Thank you to all of my BNS friends, from my initial cohort – Ben Fry, Jenny Gerena, and Henry Yang – to the cohort who adopted me – Jessica Lee, Samantha Bowden, Taryn Meinhardt, and Allie Costello – for providing solidarity, support, and humor as we navigated grad school through moving buildings, a global pandemic and more. I am also grateful to have found mentors and friends in BNS post docs Katie Yoest, Christina Reppucci, and Jenna Lee. Beyond MSU, thank you to each and every mentor I’ve had over the years, especially my first – Dr. Ján Pékarovic, who inspired my initial passion for research. Thank you to the friends who’ve indulged me in continuing school for so long, especially those that have stood by me since high school – Mallory Busso Wollerman, Alex Liou, Colleen Unsworth, Kyle Mulvaney and Becca Jacobs – and college – Taylor McEvilly, Erin VanBuskirk, Sarah Schulte, Megha Patel and Amanda Ahrens. Thanks also to Keenan Scribner and Erica Tooker. I’m so grateful to each and every one of you. I am phenomenally lucky to call Dodie and Jay Raycraft Mom and Dad, and Meghan my sister. Their influence on me – and my education – can’t be captured in words. I am also lucky to have the support of my (future) mother-in-law, Janet Kerber. I could not have done this without them. I love you all dearly. Finally, thank you to my fiancé, Mitch Lindstrom. You have had unwavering confidence in me since I first entertained the idea of applying to graduate school, and have been my rock throughout. Thank you for believing in me, even when I didn’t. I love you. vi TABLE OF CONTENTS CHAPTER 1: Introduction ................................................................................................ 1 CHAPTER 2: Accumbens-projecting Melanin Concentrating Hormone neurons that originate in the posterior Lateral Hypothalamic Area do not influence motivation or time perception ....................................................................................................................... 32 CHAPTER 3: MCH Neurons in the anterior LHA interact with estrous cycle stage to influence motivation in a time-dependent manner .......................................................... 66 CHAPTER 4: Estrogen is necessary for LHAaMCHàNAc neuronal effects on post- criterion responding ........................................................................................................ 95 CHAPTER 5: Overall Discussion ................................................................................. 120 BIBLIOGRAPHY .......................................................................................................... 128 APPENDIX ................................................................................................................... 145 vii CHAPTER 1: Introduction The ability to perceive time is integral to our daily lives. Time perception enables us to respond to changing environments and make predictions about future events, such as food availability. Time is perceived on multiple scales, including within the seconds to minutes range. This distinct form of timing known as interval timing supports associative learning and the formation of predictive relationships by providing temporal contiguity (Balsam et al., 2010; C. V. Buhusi & Meck, 2005). Interval timing can contribute to the ability of an individual to learn about their food environment through the acquisition of predictive associations between environmental stimuli and food. As such, interval timing has the potential to greatly impact decisions associated with the acquisition and ingestion of food. Despite this, few studies have examined the relationship between interval timing and feeding behavior. The lateral hypothalamic area (LHA) is a regulatory brain region ideally situated to integrate a variety of physiological signals with information from higher-order brain regions involved in timing and reward processing. As such, it has been described as an “integrator,” “hub,” and “motivation-cognition interface” for its role in motivated decision making (Berthoud & Münzberg, 2011; Bonnavion et al., 2016; Mogenson et al., 1980; Petrovich, 2018). A subset of neurons within the LHA synthesize the appetite- stimulating (i.e., orexigenic) neuropeptide, Melanin Concentrating Hormone (MCH). Like the LHA, MCH has also been described as an “integrative peptide” for its role in synthesizing information from various homeostatic and hedonic neurochemical signals to coordinate motivated behavior (Diniz & Bittencourt, 2017). Notably, LHA neurons that synthesize MCH, herein referred to as LHAMCH neurons, influence learned food intake in 1 tasks that inherently rely on the ability to perceive time in the seconds-to-minutes range (Noble et al., 2019; Sherwood et al., 2012, 2015; Subramanian et al., 2023). Previously, I examined whether neurons in the LHA that produce MCH could influence time-dependent food-seeking in an interval timing task. The peak interval (PI) paradigm is a duration reproduction task in which animals are trained that lever presses result in reinforcement only after a criterion duration (e.g., after 20s) has elapsed. Using chemogenetics to selectively excite LHAMCH neurons during the PI paradigm, I revealed that LHAMCH neurons potentially modulate time perception and/or motivation within this food-seeking task. However, the effects of LHAMCH neuronal excitation depended on the extent of LHAMCH neuronal excitation within the LHA (i.e., targeting of a more anterior vs posterior subset of LHAMCH neurons, LHAa and LHAp, respectively) and sex of the animals. Excitation of LHApMCH neurons putatively changed time perception, an effect that was subtle in male rats but pronounced in females. In contrast, excitation of LHAaMCH neurons had no effect on time perception, per se. Instead, excitation of LHAaMCH neurons robustly increased responding after the omission of an expected food reward in female – but not male – rats. This prolongation of high rate responding after the criterion had elapsed indicates increased motivation to continue food-seeking. In other words, LHAaMCH neuronal excitation increased motivation to continue food- seeking in female, but not male, rats. Given that the rodent estrous cycle is known to modulate food intake in female rats, we next examined whether the influence of LHAMCH neuronal excitation on time- dependent food-seeking was modulated by estrous cycle stage. The rodent estrous cycle is characterized by fluctuating gonadal hormones which typically peak during 2 proestrus and estrus (P/E) and then fall rapidly during metestrus and diestrus (M/D) (Goldman et al., 2007). Strikingly, in the PI paradigm, the effects of LHAMCH neuronal excitation on temporally mediated food-seeking were dependent on the estrous cycle stage in which rats were tested. Excitation of LHAaMCH neurons produced a robust increase in temporally mediated motivated responding only when rats were tested during M/D. This effect reveals a role for LHAMCH neurons in coordinating motivated behavior based on information about the timing of food availability (i.e., the temporal context) and in manner that depends on estrous cycle stage. However, the potential circuitry underlying these effects has not yet been identified. As a key mesolimbic brain region involved in motivated behavior, including in aspects of both food intake and interval timing, the Nucleus Accumbens (NAc) is a potential downstream target of LHAMCH neurons (Kelley, 2004; Kurti & Matell, 2011; MacDonald et al., 2012). Indeed, the MCH receptor, MCH1R, is densely expressed in the NAc and infusion of the MCH peptide to the NAc increases feeding (Georgescu et al., 2005). In addition, the estrogen receptor ER- α is densely expressed in the NAc with MCH1R, providing a potential mechanism through which fluctuating estrogen levels could influence MCH activity (Terrill et al., 2020). As such, projections from LHAMCH neurons to the NAc might underlie the effects of LHAMCH neuronal excitation on temporally mediated food-seeking in female rats. To explore this possibility, this dissertation examines the effects of chemogenetic excitation of anterior and posterior LHAMCH neurons that project to the NAc (LHAMCH àNAc) on time-dependent food- seeking in the PI task. In addition, I examine the role of the estrous cycle in mediating the influence of these LHAMCH àNAc neurons on this motivated behavior. Finally, I 3 directly manipulate estrogen to isolate a potential role for this hormone in moderating the effects of LHAMCH àNAc neuronal excitation on motivated behavior. Together, these studies will explore a role for LHAMCH àNAc neurons in modulating time-dependent motivated behavior and describe whether these effects are influenced by estrous cycle stage and/ or estrogen. To place these results in context, this first dissertation chapter will provide insight into the traditional view of food intake from a circadian perspective as well as describe alternative models of interval timing. I will then describe the role of the lateral hypothalamic area (LHA) in food intake as well as the Melanin Concentrating Hormone (MCH) neuropeptide system and its interactions with the nucleus accumbens (NAc) in motivated feeding behavior. Finally, I will discuss the interactions of the estrous cycle with MCH and describe how these effects are accounted for in these dissertation studies. These data provide a novel framework through which LHAMCH neurons may coordinate behavior based on information from a broader temporal context. 1.1 Feeding behavior is coordinated across multiple timescales (i) Circadian Timing Traditionally, research has primarily examined the timing of food intake on a 24- hour scale, examining the influence of circadian rhythms on food intake (Bass & Takahashi, 2010; W. Huang et al., 2011; ter Haar, 1972). Circadian timing relies on a “master clock” in the suprachiasmatic nucleus (SCN) of the hypothalamus, which integrates light cues via the retinohypothalamic tract (RHT) to synchronize the transcription of so-called “clock genes.” The lateral hypothalamic area (LHA) receives input from the SCN to coordinate arousal and sleep-wake cycles via neuropeptide and 4 hormonal signaling (Arrigoni et al., 2019; Brown et al., 2015; Goodless-Sanchez et al., 1991). Thus, damage to the LHA disrupts the sleep/wake cycle (Mistlberger et al., 2003; Pfeffer et al., 2012) in part due to the influence of MCH expressing cells. Accordingly, during REM sleep, MCH production increases and a subset of MCH neurons become active (Blouin et al., 2013; Hassani et al., 2009; Jego et al., 2013). Furthermore, acute optical stimulation of LHA MCH cells promotes REM sleep (Jego et al., 2013). Although MCH has been critically implicated in the sleep-wake cycle, findings suggest that this does not extend to an influence of MCH in modulating food intake over protracted time frames. Animals will evoke an increase in motoric behavior in the period preceding predictable food availability; this food-anticipatory activity (FAA) is under the control of a circadian-like time-keeping mechanism that prepares organisms for meal intake (Challet, 2019). Although the brain mechanisms underlying FAA are independent of the SCN, they are influenced by neurons in the LHA (Mieda & Yanagisawa, 2002); however, this does not appear to require the MCH system as deletion of MCH1R has no influence over FAA (Zhou et al., 2005). Thus, while MCH is important for regulating sleep-wake cycles, it’s influence in guiding appetite through timing likely involves mechanisms that control shorter duration timescales. (ii) Interval Timing Interval timing refers to the ability to perceive time in the seconds to minutes range and is functionally and molecularly distinct from circadian timing. Rather than rely on the SCN, the perception of time in this brief range involves a distributed network of corticostriatal circuits, the basal ganglia (BG), and the substantia nigra (SN) (C. V. Buhusi & Meck, 2005; Meck, 1996, 2005). In the lab, interval timing is studied through 5 the use of duration reproduction tasks, which train animals to reproduce a criterion time through operant responding. One such procedure is the Peak Interval (PI) paradigm, a classic interval timing task. The PI paradigm is modified from Fixed Interval (FI) procedures, in which animals learn that instrumental responding (i.e., lever pressing) results in reinforcement only after a particular criterion duration has elapsed (Balcı & Freestone, 2020; C. V. Buhusi & Meck, 2006; Rakitin et al., 1998; Roberts, 1981). During FI trials, animals can respond on the lever at any point in time, but only responses that occur after the FI duration (e.g., after 20s) are reinforced. Thus, animals typically respond intermittently throughout the trial and accelerate their response rate around the time of expected reinforcement (Skinner, 1938). Thus, lever responding is characterized by a “break- then-run” pattern, where low rates of lever responding abruptly increase in anticipation of reinforcement (Balcı, 2014; Schneider, 1969). This transition from a low to high rate of responding is conceptualized as the “start” function and reflects an increase in reward expectancy as the time of reinforcement approaches (Balcı, 2014; Gibbon, 1977). Acquisition of this “start” function depends on the dorsal striatum (MacDonald et al., 2012). While FI trials enable responding prior to reinforcement to be examined, reward delivery and trial offset prevents the ability to examine responding at or after the criterion time. Thus, Peak interval paradigms are unique from FI paradigms in that they also include probe trials (Balcı & Freestone, 2020; C. V. Buhusi & Meck, 2006; Rakitin et al., 1998; Roberts, 1981). Following the acquisition of FI responding, probe trials are intermixed during PI training. Probe trials onset in the same manner as FI trials and are 6 initially indistinguishable from FI trials. However, during probe trials no reinforcement is delivered, even after the criterion time has elapsed. Instead, probe trials typically last at least 3x the length of the FI criterion and randomly offset without reward delivery. Thus, after sufficient FI training, probe trials can be randomly intermixed with FI trials in order to examine how animals respond across time without the contamination of reward delivery. During probe trials, animals exhibit a break-run-break pattern of responding, with a “start” transition into high-rate responding before the criterion (similar to FI trials) and a “stop” transition back to low-rate responding after (Balcı, 2014; Church et al., 1994; Church & Broadbent, 1991; Gallistel & Gibbon, 2000). In other words, during probe trials animals increase their response rate to a peak (peak rate) at the time of expected reinforcement, then attenuate high rate responding after the perceived time of reinforcement has elapsed. The actual time at which peak response rate occurs is peak time, and under normal conditions this will approximately equal the criterion time (Balcı, 2014; Gibbon, 1977). When responses are plotted across time as a proportion of peak rate, the resulting proportion of peak rate function forms an approximately normal distribution centered around the criterion time with a slight negative skew. Early in PI training, when probe trials are first introduced, this negative skew is considerable, with high rate responding persisting long after the omitted food reward. The negative skew of the proportion of peak rate response function is reduced as animals learn to attenuate high rate responding and the “stop” function is acquired. Thus, the right-hand side of the proportion of peak rate response function narrows as the “stop” function is acquired. The “start” and “stop” functions are viewed as distinct features of responding that reflect 7 independent, time-mediated decision processes and that are acquired during different phases on the PI task (C. V. Buhusi & Meck, 2009). Consistent with this idea, unlike the “start” function that is mediated by the dorsal striatum, the decision to stop responding once the criterion duration has elapsed reflects control by the ventral striatum (MacDonald et al., 2012). Once the “start” and “stop” functions have been acquired, the overall width of the response function that is generated varies proportionally with the length of the criterion being timed. This in part reflects errors in timing that increase in proportion to the length of the duration being timed (Balsam et al., 2009; Gibbon, 1977; Malapani & Fairhurst, 2002; Meck, 1996, 2005). This property forms the basis of Scalar Expectancy Theory (SET; also known as Scalar Timing Theory or the scalar property) and is reflected in a proportional broadening or narrowing of the response function with longer or shorter intervals, respectively. (Church, 1984; Gibbon, 1977; Malapani & Fairhurst, 2002; Meck, 1996). The scalar property is an important feature of timing, as it enables confirmation of changes in time perception. While the “start” and “stop” functions represent discrete behavioral states that can be separately influenced, changes to time perception proportionally alter these functions and shift the response function. For example, if time perception is sped up (i.e., clock speed increases), high rate responding will “start,” peak, and “stop” earlier than under normal conditions, resulting in a proportional leftward shift of the response function that coincides with a reduction in peak time. In addition, the width of the “run” period (i.e., the length of high rate responding around the peak) follows the scalar property and thus would proportionally narrow when clock speed is increased (Gibbon, 1977). 8 (iii) Models and Mechanisms of Interval Timing Conceptual models of interval timing posit that an internal clock keeps track of time in the seconds-to-minutes range. This clock is molecularly and functionally discrete from the SCN “master clock” of circadian timing. Unlike the circadian master clock, which reliably syncs to light as a zeitgeber and accurately tracks long periods of time in the ~24-hour range, adjusting timing slowly (such as through a gradual phase shift), the internal clock of interval timing can time any number of meaningful events rapidly and flexibly (Meck, 1996). Indeed, it is posited that the internal clock can time multiple arbitrary intervals simultaneously, with seemingly no limit aside from those imposed by attentional processes (Balsam et al., 2009; Gibbon, 1977; Malapani & Fairhurst, 2002; Meck, 1996, 2005). However, unlike the circadian clock, which is incredibly accurate over the course of many hours, the internal clock of interval timing is susceptible to variance, which increases in proportion to the duration being time. Thus, the internal clock varies from the circadian clock in at least three ways: (1) the internal clock is more flexible, (2) the internal clock is less precise, and (3) the internal clock displays the scalar property (Gibbon, 1977; Meck, 1996). While there are multiple models of interval timing, most of them share three key components: a clock or accumulator, a decision-making comparator mechanism, and a memory component. The Pacemaker-Accumulator model is perhaps the most influential model of interval timing (Meck, 1996). This information-processing model posits that the firing of dopaminergic cells in the substantia nigra pars compacta (SNpc) results in the release of striatal dopamine (DA) which in turn acts as an internal clock or “accumulator.” This accumulation of striatal DA serves as an indicator for when 9 meaningful events occur. An attentional gate or “switch” closes when salient events occur (e.g., when a discriminative stimulus is presented) and the amount of DA that has accumulated before another meaningful event (e.g., food reward delivery) can thus be measured. In this example, the duration between when the stimulus is presented and when actions (e.g., lever presses) result in reinforcement can be stored as a reference memory of the amount of DA that accrued during this interval. In the future, when the same stimulus is presented, this reference memory for time is recalled into working memory and compared to the amount of DA currently accruing as time passes. This is the “comparator” component of the Pacemaker-Accumulator information processing model. When the amount of DA that has accrued matches the reference memory, a decision is made and behavior changes accordingly; this is the decision component. For example, the decision to start responding at a high rate is made when the reference memory for time approximately matches the current perception of time. Likewise, the decision to stop responding at a high rate occurs when the current time exceeds the reference time. Given that the internal clock relies on accumulating DA, changes to the rate of DA accumulation can alter clock speed (i.e., the perception of how quickly time passes). As such, drugs that increase striatal DA (e.g., methamphetamine) result in a decrease in peak time that coincides with a leftward shift and proportional narrowing of the response function (Matell et al., 2006; Meck, 1983). Together, these features indicate an increase in clock speed. In contrast, drugs that decrease DA accumulation (e.g., the D2R antagonist haloperidol) delay peak time as well as produce a rightward shift and proportional broadening of the response curve, indicating a decrease in clock speed (C. 10 V. Buhusi & Meck, 2002; Meck, 1983, 2006). Together, these studies provide evidence for the dopaminergic nature of the internal clock. On the other hand, the memory component of the internal timing system appears to depend on intact cholinergic signaling, as disrupting acetylcholine interrupts memory formation and/ or recall, depending on the timing of disruption in memory consolidation or retrieval (Meck, 1983, 1996; Meck & Church, 1987). While the pacemaker-accumulator model of interval timing provides a succinct theoretical framework, one criticism of the model is that its molecular underpinnings remain unclear and – at times – in contrast to its theoretical components. For example, how the pacemaker-accumulator process begins (i.e., to what is the initial duration compared?) remains murky. In addition, the ability to simultaneously time multiple intervals challenges the idea of a single internal clock. The striatal beat frequency (SBF) model integrates new evidence regarding the neurobiology of interval timing with the conceptual framework provided by SET and pacemaker-accumulator models (C. V. Buhusi & Meck, 2005). As in the pacemaker accumulator model, SBF ascribes a clock function to both the SNpc and striatum, but in this case timekeeping relies on coincidence-detection by multiple neurons (C. V. Buhusi & Meck, 2005; Matell et al., 2003; Matell & Meck, 2004). Fundamental to the SBF model is the assumption that neurons oscillate at a given frequency – albeit not always synchronously – and that their coincident activation can be used to measure time across multiple intervals (Matell et al., 2003). In other words, the perception of a given interval is associated with a broad neural activation pattern, during which some neurons are activated and others are not (C. V. Buhusi & Meck, 2005; Hinton & Meck, 2004; 11 Matell et al., 2003; Merchant et al., 2013). From this perspective, the striatum acts as a perceptual filter, integrating the pulsatile activation of SNpc with meaningful events and stimuli via coincident activation in the striatum. These activation patterns are learned through Hebbian strengthening; thus, the memory for a learned interval in this model depends on long term potentiation or depression (LTP or LTD, respectively) within corticostriatal circuits (C. V. Buhusi & Meck, 2005; Matell et al., 2003). The SBF model has gained traction as electrophysiological and fMRI studies have revealed corticostriatal activation increases around the time of expected reward (C. V. Buhusi & Meck, 2005; Matell et al., 2003). It is also supported by studies of reward prediction error, where DA activity has been recorded during trials in which initially neutral stimuli are learned to predict reinforcement (Fiorillo et al., 2003; Hollerman & Schultz, 1998). While DA initially fires robustly to reinforcement delivery, learning quickly shifts this neuronal response to the earliest stimuli that predicts future reinforcement (Fiorillo et al., 2003; Hollerman & Schultz, 1998). Thus, DA activity encodes the time of expected reward via predictive cues. In the PI paradigm, stimuli that indicate trial onset act as this predictive cue, and the DA pulse that occurs may serve to initiate timing during the trial (Matell et al., 2003). Despite the accruing neurobiological support for the SBF model, it still cannot fully account for interval timing processes. For example, although SBF attempts to account for biological variability and the scalar property of timing by adding sources of variance to the model, the neurobiological basis of this variability and how to best incorporate it remains unclear. Nevertheless, the SBF model may provide the neural underpinnings of the conceptual model described by the pacemaker-accumulator (C. V. 12 Buhusi & Meck, 2005). Furthermore, when considered together, these models provide a framework through which corticothalamic oscillations acting as the “clock” work in concert with striatal MSNs, which in turn filter or integrate timed signals within a broader context to modify behavior (Matell et al., 2003). Specifically, striatal activation is relayed through basal ganglia output nuclei to the thalamus and motor cortex to coordinate timed behaviors (Matell et al., 2003). Regardless of the model ascribed, there is significant evidence to suggest that interval timing mechanisms involve a distributed network of thalamo-corticostriatal circuits. Given that the hypothalamus also contacts corticostriatal circuits involved in timing, I hypothesize that cells within this region may be capable of influencing time in the seconds-to-minutes range. Integration of the temporal context in the seconds-to- minutes range could inform a variety of food related behaviors, such as predicting when food will be available or when to initiate or terminate a meal. Thus, it follows that hypothalamic neurons that coordinate food intake behaviors might use this temporal information to appropriately facilitate feeding. As such, let us discuss the anatomy of the LHA and evidence suggesting LHAMCH neurons may guide appetitive behavior through temporally mediated processes. 1.2 Lateral Hypothalamic Regulation of Feeding Behavior (i) The Lateral Hypothalamic Area in Food Intake The neural control of food intake is primarily localized within the hypothalamus, a diencephalic region critically involved in energy homeostasis. Classically, the hypothalamus was conceptualized as a “feeding center” after its stimulation was observed to produce voracious food intake (Hoebel & Teitelbaum, 1962). However, the 13 functional heterogeneity of the LHA has been apparent from its discovery. For example, in addition to promoting feeding, stimulation of the LHA also augments a variety of behaviors including drinking and gnawing (Valenstein et al., 1968), copulation (Caggiula & Hoebel, 1966) and aggression (Hutchinson & Renfrew, 1966) as reviewed in (Stuber & Wise, 2016). In fact, the behavioral output of LHA stimulation appears to depend on the state of the animal, the experimental parameters, and previous learning, which led to the view that this stimulation could generally augment arousal and modulate reward processes (Stuber & Wise, 2016). In addition, these studies also revealed a direct role for the LHA in reward, as rats would continuously lever press for LHA self-stimulation (Berridge & Valenstein, 1991; Hoebel & Teitelbaum, 1962; Olds & Milner, 1954). These additional findings provided the first indication that classification of the LHA solely as a “feeding center” did not account for its complex and varied roles in motivated behavior. Instead, these findings alluded to the diverse functions of the LHA supported by its underlying physiological complexity. (ii) Anatomical and molecular complexity of the LHA The LHA forms a bed nucleus for two large bundles of fibers which pass through it rostrocaudally: the fornix (fx) and the median forebrain bundle (mfb), the latter of which is instrumental in reward signaling (Nieuwenhuys et al., 1982; Saper et al., 1979). The mfb contains dense bundles of dopaminergic axons extending from the brainstem and midbrain to the forebrain, including fibers separately comprising the nigrostriatal (substantia nigra to dorsal striatum) and mesolimbic (ventral tegmental area to ventral striatum and basal ganglia) reward pathways (Nieuwenhuys et al., 1982; Qualls- Creekmore & Münzberg, 2018). The LHA is thus ideally situated to act as a relay station 14 capable of integrating central and peripheral energy balance signals with higher-order brain regions involved in affective processing, decision-making, reward and timing (Berthoud & Münzberg, 2011; Bonnavion et al., 2016; Petrovich, 2018; Saper et al., 2002; Stuber & Wise, 2016). Despite its easily identifiable location relative to the mfb and fx, the LHA is complicated by a lack of clear and discernable anatomical bounds (Hahn & Swanson, 2010; Saper et al., 1979). In the rat, the LHA extends rostrocaudally from about -1.30 mm to -4.80 mm relative to Bregma (Paxinos & Watson, 1998) and is typically described as being constrained by anterior and posterior boundaries in the preoptic area (POA) and ventral tegmental area (VTA), respectively (Berthoud & Münzberg, 2011; Hahn & Swanson, 2010; Stuber & Wise, 2016). Recent advances in neuroscience have enabled further delineation of the cytoarchitecture of the LHA. Notably, tract tracing studies have been used to identify and describe more than twenty LHA subregions (Hahn & Swanson, 2010, 2015; Swanson et al., 2005) while molecular profiling of the various cell types identified within the LHA – of which there are many – has highlighted its heterogeneity (Bonnavion et al., 2016; Mickelsen et al., 2017, 2019). Within the LHA, one of the best characterized subpopulations of cells are those that produce the orexigenic (i.e., appetite-stimulating) neuropeptide Melanin concentrating hormone (MCH) (Bittencourt et al., 1992; Qu et al., 1996). Like the LHA, MCH has also been described as an “integrative peptide” for similarly integrating the homeostatic and rewarding features of feeding behavior (Diniz & Bittencourt, 2017). In the following sections, I will discuss this system in more detail. 15 1.3 LHA Melanin Concentrating Hormone (i) The MCH peptide MCH is a cyclic 19-amino acid protein, the production of which is driven by the pMCH promoter, which also encodes neuropeptides EI and GE (Broberger et al., 1998; Pissios et al., 2006; Qu et al., 1996). As an orexigenic peptide, central infusions of MCH increase food intake (Baird et al., 2006; Della-Zuana et al., 2002; Gomori et al., 2023). In addition, genetic overexpression of the pMCH promoter in mice leads to obesity whereas deletion of the peptide or its receptor both result in hypophagia (Ludwig et al., 2001; Shimada et al., 1998). Recall also that MCH plays a role in arousal by increasing REM sleep to support energy conservation (Monti et al., 2013). In line with this, mice with a genetic deficiency of the MCH receptor, MCH1R, are hyperactive and lean (Marsh et al., 2002). The orexigenic and rewarding actions of MCH are thought to occur through the nucleus accumbens (NAc), where infusions of the MCH peptide also increase food intake (Georgescu et al., 2005). In the following section, I will describe the anatomy and physiology of the MCH system, with a particular emphasis on its role in temporally mediated food intake and actions in the NAc. (ii) LHAMCH Neurons: Anatomy & Physiology Within the LHA, MCH neurons are capable of synthesizing both GABA and glutamate, however the majority of these neurons are likely glutamatergic, as they lack the vesicular transporters vGAT and vMAT necessary for GABA release (Bonnavion et al., 2016; Mickelsen et al., 2017, 2019). Indeed, at least in the lateral septum (LS), stimulation of LHAMCH neurons leads to the release of glutamate (Chee et al., 2015). Based on their co-expression of additional peptide markers, Mickelsen et al. (2019) 16 suggest that glutamatergic LHAMCH neurons can be further categorized into two subclusters, defined in part by whether or not they co-express the anorexigenic signal CART (i.e., CART+ or CART-). The CART+ subcluster robustly expresses CART as well as Tacr3, and Nptx1, which encode tachykinin receptors and neuronal pentraxin, respectively (Mickelsen et al., 2019; Stelzer et al., 2016). In contrast, the CART- subcluster instead moderately expresses Scg2 and Nrxn3, which encode secretogranin II and neurexin 3, respectively. These proteins are involved in receptor function and cell adhesion, as well as the sorting and packaging of peptide hormones (Mickelsen et al., 2019; Stelzer et al., 2016). In addition, LHAMCH neurons express receptors for GABA, glutamate, glucocorticoids, NPY, melanocortins, and leptin, although whether the expression of these receptors varies between subclusters of LHAMCH neurons remains unclear (Bäckberg et al., 2004; Harthoorn et al., 2005; Huang & van den Pol, 2007; Lee et al., 2021; Mickelsen et al., 2019). Finally, early immunochemical work provides ample evidence that LHAMCH neurons also express α-melanocyte hormone (α-MSH), corticotrophin releasing hormone (CRH) and growth hormone releasing hormone (GHRH) (Bittencourt et al., 1992). However, the extent to which the synthesis and release of these hormones by LHAMCH neurons varies by subcluster – if at all – remains unclear. Nevertheless, the molecular heterogeneity of LHAMCH neurons speaks to the robust potential of these neurons to interact with a variety of brain systems to modulate behavior. (iii) The MCH Receptor System In rodents, the MCH peptide exerts its actions through the MCH1R receptor (Hawes et al., 2000; Saito et al., 1999; Tan et al., 2002). Although a second MCH 17 receptor (MCH2R) has been identified in humans, it is either absent or non-functional in rodents (Tan et al., 2002). The MCH peptide binds MCH1R with a high affinity and selectivity: nanomolar concentrations of MCH strongly activate MCH1R, while neuropeptide EI, which is also produced by the pMCH promoter, does not (Saito et al., 1999). MCH1R is a G-protein coupled receptor (GPCR) first identified as the orphan GPCR known as SLC1 or GPR24 (Hawes et al., 2000; Saito et al., 1999; Tan et al., 2002). GPCRs are typically grouped into subclasses based on their α-subunit and denoted as Gi/o, Gs, or Gq. The MCH1R can be coupled to multiple G protein subtypes, including both Gi/o and Gq. Activation of MCH1R by MCH inhibits cyclic AMP (cAMP) production, induces a transient increase in calcium concentration ([Ca2+]), increases mitogen-activated protein kinase (MAPK), and increases inositol phosphate (IP), a marker of increased phospholipase C (PLC) (Chambers et al., 1999; Hawes et al., 2000; Saito et al., 1999). The effects of MCH1R activation on cAMP and MAPK production can be blocked by pretreatment with pertussis toxin (PTX), which inactivates Gi/o proteins, indicating that these MCH1R effects on intracellular signaling are Gi/o mediated. On the other hand, PTX pretreatment does not fully abolish the effects of MCH1R activation on the stimulation of PLC, indicating that this mechanism is mediated at least in part by Gq receptors (Hawes et al., 2000). Thus, MCH1R functionally couples to both Gi/o and Gq receptors and can therefore exert both inhibitory and excitatory effects on cellular activity. To date, little is known about what determines which signaling mechanisms are initiated following MCH1R binding. These intracellular mechanisms may vary by brain region, function, or any number of other properties that 18 determine what molecular components are present in a given MCH1R+ cell (Hawes et al., 2000). MCH1R is expressed extensively throughout the CNS, as demonstrated by in situ hybridization and RT-qPCR of MCH1R mRNA as well as by protein immunoreactivity (Bittencourt et al., 1992; Hervieu et al., 2000). Bittencourt et al. report that MCH immunoreactive (MCH-ir) fibers are found in nearly “every commonly recognized cell group and cortical field” (Bittencourt et al., 1992), including throughout the diencephalon, mesencephalon, and rhombencephalon (Hervieu et al., 2000). In addition, MCH signaling also occurs through volume transmission in the cerebrospinal fluid, which indicates that the MCH peptide can also influence brain regions which are not directly innervated by MCH-ir fibers (Noble et al., 2018). Within the CNS, MCH1R expression is particularly dense in the hypothalamus, including both within the LHA and ZI. MCH1R is also evident throughout the olfactory system, the hippocampus, and the amygdala. It is densely expressed in the caudate putamen, substantia nigra, and striatum, each of which are important for interval timing (C. V. Buhusi & Meck, 2005; Hervieu et al., 2000; Meck, 1996, 2005). The expression of MCH1R in the ventral striatum is of particular interest, given the role of the nucleus accumbens (NAc) in motivated behavior, timing, and food intake. 1.4 Interim Summary Now that I have described the interval timing system, the LHA, and the MCH system, let’s reconsider the preliminary data indicating a potential role for LHAMCH in time-dependent food-seeking. As described above, I previously examined the influence of LHAMCH neurons on time-dependent food-seeking by chemogenetically exciting these 19 neurons in the anterior or posterior LHA (LHAa or LHAp, respectively) during the peak interval paradigm. Using male and female rats, I found that LHApMCH neuronal excitation putatively increased clock speed in male and female rats, an effect that may have been driven by responding in females. In contrast, chemogenetic excitation of LHAaMCH neurons did not influence time perception per se in male or female rats. Rather, in female rats this manipulation robustly delayed the “stop” function, resulting in prolonged high rate responding after the omission of an expected reward. In contrast to the LHApMCH neuronal excitation, these effects occurred selectively during M/D, indicating an ability for LHAaMCH neuronal excitation to increase motivation only during periods of the estrous cycle when circulating gonadal hormone levels are typically lower. Together, these findings indicate a role for LHAMCH neurons in modulating time- dependent motivated responding in females, in a manner that depends on estrous cycle stage. In particular, LHAaMCH neurons robustly increased motivated responding during M/D by delaying the “stop” function after the omission of an expected reward. Given that the ventral striatum is important in determining the “stop” function during interval timing tasks, this suggests that these LHAMCH neurons may exert their actions through signaling in the ventral striatum (i.e., in the NAc). In addition, these LHAMCH neurons may interact with circulating gonadal hormones like estrogen to modulate motivated responding. As such, I will next describe the role of MCH in the NAc, as well as evidence indicating that MCH supports temporally mediated food intake. Finally, I will describe the rodent estrous cycle and evidence for interactions between the MCH system and estrogen. 20 1.5 MCH in motivated and time-dependent feeding behavior (i) MCH activity in the NAc The striatum, which I have previously described above in terms of its dorsal and ventral portions and their discriminable roles in interval timing, can alternatively be described by its various nuclei, including those within the basal ganglia (BG). While the dorsal striatum consists of the caudate and putamen, the ventral striatum typically refers to the nucleus accumbens (NAc), which can itself be divided into two distinct subregions: the core and shell (Kelley, 2004; Mogenson et al., 1980, 1983). The NAc has long been associated with motivated behavior and described as an interface between motivation and action (Mogenson et al., 1980). Primarily composed of Median Spiny Neurons (MSNs), the NAc responds to both glutamatergic and GABAergic modulation to alter neural activity and modify motivated behaviors (Mogenson et al., 1980). The NAc both sends and receives projections from the LHA (Haemmerle et al., 2015; O’Connor et al., 2015). Both glutamatergic antagonists and GABAergic agonists infused to the NAc potently increase feeding (Kelley, 2004). MCH1R is expressed on the majority of MSNs in the NAc, including on both DA receptor-1 (DR1) and receptor-2 (DR2)-expressing neurons (Pissios et al., 2008) as well as on those that express enkephalin or dynorphin (Georgescu et al., 2005). Expression of MCH1R is especially dense in the NAc shell, where infusion of the MCH peptide increases feeding (Georgescu et al., 2005). Similarly, infusion of an MCH1R antagonist in this region instead decreases feeding (Georgescu et al., 2005). Although this group does not specify the sex of the animals used in this study, others report that infusion of MCH 21 peptide to the NAc shell increases feeding only in male – but not female – rats (Terrill et al., 2020), suggesting this site as a potential mediator of sex differentiated feeding effects. In line with this, when circulating ovarian hormones were removed via ovariectomy in females, infusion of MCH to the NAc increased food intake in oil, but not estradiol, treated rats (Terrill et al., 2020). Together, these data point to a role for the NAc and MCH in mediating sex- and estradiol-dependent effects on food intake. In addition to its role in MCH-dependent, sex-differentiated feeding behavior, a role for the NAc in “computing coincident events [to] enhance the probability that temporally related actions and [events become] associated” in feeding behavior has also been proposed (Kelley, 2004). Notably, these temporal relationships inherently involve time perception in the seconds to minutes range (i.e., they depend on interval timing). As described previously (section 1.1, above), corticostriatal circuits are integral to interval timing, and coincident activation of these neurons may underlie time perception in this range. Thus, interactions between the LHA and this corticostriatal network may support learning about food-predictive cues by integrating temporal information to inform behavior. Thus, the LHA and NAc may work together to influence food intake through either or both the modulation of time perception and motivation. In particular, given that the ventral striatum (i.e., the NAc) is specifically involved in acquisition of the “stop” function in interval timing paradigms, contacts with the NAc may be responsible for the effects of chemogenetic LHAMCH neuronal excitation on prolonging high rate responding after the criterion duration. (ii) LHAMCH in learned food intake 22 MCH is important for learning about food-predictive cues, which inherently involves understanding temporal relationships between initially neutral stimuli and reinforcing outcomes. MCH increases food intake primarily through alterations in the duration of consumption, which can be analyzed via licking microstructure (Davis & Smith, 1988, 1992; Johnson, 2018; Smith, 2001). Within a meal, individual bursts of licking behavior can be described in terms of their frequency (i.e., burst number) and duration (i.e., burst size) (Davis & Smith, 1988, 1992; Johnson, 2018; Smith, 2001). Intracerebroventricular (ICV) infusions of MCH increase burst size, which is typically interpreted as an increase in the perceived palatability of the food being consumed (Baird et al., 2006). MCH therefore supports consumption by enhancing hedonic taste evaluation (Baird et al., 2006), thereby supporting learning about the rewarding properties of food. While changes in burst size are typically interpreted in terms of hedonic evaluation (i.e., longer burst size implies increased hedonic value), they could alternatively reflect a delay in the “stop” mechanism. As a decision process guided by both time perception and motivation (MacDonald et al., 2012; Matell et al., 2006), either or both of these processes could delay the “stop” function to prolong consumption. Thus, an increased burst size following MCH peptide infusion could alternatively indicate that MCH delays the “stop” function as a result of altered time perception or motivation. Consistent with the idea that MCH modulates the “stop” function to change consumption patterns, MCH1R antagonism decreases lick burst size in response to a food-paired cue (Sherwood et al., 2015). In other words, MCH1R antagonism disrupts 23 the expression of learned overeating, specifically by decreasing the duration of bouts of consumption. In addition, both pharmacological antagonism and genetic deletion of MCH1R disrupt the ability of a reinforcing conditioned cue to support new learning (Sherwood et al., 2012), again indicating a role for MCH neurons in learning about food predictive cues. Importantly, MCH influences food intake primarily through changes the duration of individual bouts of consumption. In addition, the influence of MCH appears to be temporally constrained to the beginning of a meal and/ or within ongoing consumption. For example, optogenetic stimulation of LHAMCH neurons potently increases feeding only when stimulation is applied during consumption, revealing a role for MCH in prolonging – but not initiating – food intake (Dilsiz et al., 2020). Similarly, calcium imaging reveals that LHAMCH neurons are activated in response to food cues and during consumption (Subramanian et al., 2023). However, the activity of these neurons wanes over the course of a meal, again indicating a temporal specificity to their role in feeding behavior. Similarly, the influence of MCH peptide infusion on burst size is also greater at the beginning of a meal (Baird et al., 2006). In each case, the influence of MCH is temporally specific. Effects occur during ongoing consumption to increase the length of bouts of consumption, primarily at the beginning of a meal. In addition, LHAMCH neurons that project to the ventral hippocampus (HPc) have also been shown to increase early responding for food in a task that requires instrumental responding to be inhibited until a criterion duration has elapsed (the differential reinforcement of low rates of responding or DRL task) (Noble et al., 2019). While this effect was interpreted as an increase in behavioral impulsivity, it could 24 alternatively reflect a change in time perception resulting in increased early responding. Notably, in each of these examples, timing matters. Importantly, the PI paradigm (described in 1.1 ii, above) enables the examination of effects of both timing and motivation within the same task. 1.6 Estrogen Modulates the Influence of MCH on Food Intake It is well-established that food intake in females fluctuates with the estrous cycle, an effect largely attributed to circulating levels of estrogen (i.e, 17- β-estradiol benzoate, EB) (Blaustein & Wade, 1976; López & Tena-Sempere, 2015; Morin & Fleming, 1978; ter Haar, 1972; Varma et al., 1999). The rodent estrous cycle is typically divided into four stages: metestrus, diestrus, proestrus, and estrus (M, D, P, and E, respectively) (Goldman et al., 2007). Generally speaking, EB and other circulating gonadal hormones are typically highest during proestrus and behavioral estrus (i.e., the period when rats are sexually receptive) and then fall rapidly during the day of estrus and then remain low throughout metestrus and early diestrus (Goldman et al., 2007). In addition to EB, these hormones include progesterone and luteinizing hormone (LH), which peak in proestrus, as well as follicle stimulating hormone (FSH), which rises and falls rapidly during estrus. Ovulation typically occurs ~10-12 hours after the peak in LH, during the dark phase of estrus (Goldman et al., 2007). Removal of peripheral hormones through adult ovariectomy (OVX) results in a robust increase in food intake and body weight. However, this effect can be normalized to the level of intact, cycling animals through cyclic replacement of estrogen (Asarian & Geary, 2002; Geary & Asarian, 1999). Accordingly, estrogen has been posited to act as an anorexigenic signal, inhibiting food intake (Asarian & Geary, 2002; Eckel, 2011; 25 Geary & Asarian, 1999). The effects of EB on food intake may be mediated at least in part by MCH, as EB inhibits the orexigenic effects normally seen following MCH peptide administration (Santollo & Eckel, 2008, 2013). In addition to influencing food intake behavior, estrogen also influences time perception (Bayer et al., 2020; M. Buhusi et al., 2017; Morita et al., 2005; Sandstrom, 2007; Williams, 2011). Acute EB treatment in OVX rats proportionally shifts the response function to the left, suggesting an increase in clock speed (Pleil et al., 2011; Sandstrom, 2007). However, when administered on subsequent days, the effect of EB treatment wanes after the first session (Pleil et al., 2011). This transient effect further supports the notion that EB increases clock speed, as new learning under EB conditions would allow rats to adjust the reference memory for time. Interestingly, in intact female mice, responding occurs later than in males, although the authors suggest that this effect may be driven by a delayed “start” function and general decrease in incentive motivation rather than an effect on timing, per se (Gür et al., 2019). Although there are only a handful of studies examining estrous cycle or EB in time perception, and the role of gonadal hormones is not yet clear, it is evident that EB is capable of exerting effects on timing and/or motivation (Panfil et al., 2023). Estrogen acts by binding its receptors, estrogen receptors (ERs), in both the CNS and periphery. There are two classic ERs, ER-α and ER-β, which are located in the cytoplasm and nucleus, where they act as transcription factors (Toran-Allerand, 2004). ERs are widely distributed throughout the CNS, and densely expressed throughout the rostrocaudal extent of the hypothalamus. In 1997 Shughrue et al. provided an in-depth overview of the relative distributions of ER-α and ER-β mRNA 26 throughout the rat brain and reported that the relative proportion of ER-α and ER-β differs in various brain regions, including throughout the hypothalamus. However, there is somewhat conflicting evidence regarding the expression of ER-α and ER-β in the LHA and NAc. For example, while Shughrue et al., 1997 reported that only ER-β is expressed in the LHA, ZI, and NAc, others provide evidence of ER-α in the LHA and ZI of both mice (Couse et al., 1997) and rats (Muschamp & Hull, 2007). Similarly, recent evidence has identified ER-α mRNA in the NAc (Muschamp & Hull, 2007; Terrill et al., 2020). The timeframe of effects mediated by nuclear ER-α and ER-β depend on the rate of translocation to the nucleus, transcription, and degradation of various protein products. These effects are therefore typically slow to onset and long lasting. However, the effects of EB application can also occur rapidly, within seconds to minutes (Boulware et al., 2005; McEwen, 2002). These rapid effects of estrogen are typically referred to as non-genomic and are thought to occur via membrane-bound rather than nuclear ERs. Recent evidence has indicated that ER-α and ER-β are also associated with the cellular membrane as homo- and heterodimers, where their effects may occur more rapidly (Almey et al., 2015; Toran-Allerand, 2004). In addition, a third ER membrane-bound receptor, ER-X, has also been putatively identified (Toran-Allerand et al., 2002). Estrogen also interacts with G-protein coupled receptors, of which there are three subtypes: the excitatory mGlurI (mGluR1 and mGluR5, which are Gq receptors), and inhibitory mGluRII and mGluRIII (including mGluR 2 and 3 and mGluR 4-7, respectively, which are Gi/o receptors) (Almey et al., 2015). There has been at least one GPCR ER identified, GPER1, but estrogen has also been shown to act in vitro on 27 mGluR1 and 2 in hippocampal neurons (Boulware et al., 2005) and on mGluR5 in striatal neurons (Grove-Strawser et al., 2010). Thus, estrogen may exert both excitatory and inhibitory effects on a variety of neurons through GPCRs. Rapid effects of estradiol mediated through these non-genomic mechanisms may include changes in membrane permeability and activation of multiple signaling pathways, including cyclic AMP/ protein kinase A (PKA), mitogen-activated protein kinase (MAPK) and phospholipase C (PLC) (Almey et al., 2015; Boulware et al., 2005). The MCH system is robustly modulated by the estrous cycle, and MCH may in turn regulate hormonal fluctuations across the cycle. For example, low levels of MCH peptide, receptor, and MCH-ir fibers have been reported during proestrus and estrus (P/E) (Murray et al., 2000; Santollo & Eckel, 2013). Indeed, the density of MCH-ir fibers rapidly decreases from the morning of proestrus to the evening, as circulating levels of estrogen and progesterone rapidly rise (Gallardo et al., 2004). Additionally, this decrease in MCH signaling has been suggested as a mechanism controlling the release of LH, indicating that MCH may also influence gonadal hormones (Gallardo et al., 2004). Through ovariectomy and estrogen replacement, Santollo & Eckel (2013) have isolated estrogen as the hormone responsible for reducing MCH expression during P/E. However, the mechanism by which estrogen regulates MCH remains unclear, as it is commonly accepted the LHAMCH neurons do not express ER-α (Muschamp & Hull, 2007) and are unlikely to express ER-β based on its distribution (Li et al., 1997). However, the potential action of estrogen on the MCH system via GPCRs has not been ruled out. Given that the MCH1R receptor is a GPCR that couples to both Gi and Gq proteins (Hawes et al., 2000), it is possible that estrogen may compete with 28 circulating MCH to bind MCH1R. Estrogen may also bind directly to LHAMCH neurons through mGlur1 (Huang & van den Pol, 2007). Alternatively, estrogen could mediate the effects of MCH indirectly, as the expression of MCH1R and ER-α are similarly distributed throughout the CNS (Bittencourt et al., 1992; Muschamp & Hull, 2007; Shughrue et al., 1997) and even colocalized in non-MCH neurons, including throughout the LHA and NAc (Muschamp & Hull, 2007; Terrill et al., 2020). In addition, MCH1R is more densely colocalized with ER-α in female rats compared to their males counterparts, providing a potential mechanism for sex- and estrous-cycle effects on MCH mediated behaviors coordinated by the NAc (Terrill et al., 2020). 1.7 Overview of dissertation chapters Given the evidence of sex-dependent effects on MCH-mediated food intake in the NAc, as well as my own findings that LHAMCH neurons modulate time-dependent food-seeking in a sex- and estrous-cycle dependent manner, I hypothesized that LHAMCH neurons that project to the NAc may account for these effects. Support for this hypothesis comes from an extensive literature indicating a role for the NAc in motivated food intake (Kelley, 2004) as well as evidence suggesting that the NAc modulates the “stop” function of time-dependent food-seeking in interval timing tasks (MacDonald et al., 2012), perhaps by communicating the incentive value of reward (Kurti & Matell, 2011) As an integrative peptide, MCH activity in the LHA and NAc may link the temporal context of a changing environment to motivation in order to inform behavior. While a functional role of MCH neurons that project to the NAc from the posterior LHA has been previously described (Terrill et al., 2020), I first sought to confirm that LHAMCH neurons in the anterior LHA also project to the NAc. To do so, I examined 29 projections from the LHAa and LHAp to this region using a retrograde viral tracing technique. Using a non-specific retrograde adeno-associated virus expressing GFP, I identified a dense expression of MCH-ir fibers in the NAc that originate from the LHAa. In line with the reports of others (Pissios et al., 2008), our retrograde tracing from the anterior LHA indicates that LHAa projections primarily contact the NAc shell, rather than core. Figure 1.1 Representative image of eGFP expression in the (a) anterior, dorsolateral LHA injection site and (b) the ACBS following infusion of AAV2-hSyn-eGFP. ac=anterior commissure; ACBc= Nucleus accumbens core; ACBS= Nucleus accumbens shell; ZI = zona incerta. In this dissertation, I separately examined the effects of chemogenetic excitation of NAc-projecting neurons from posterior (LHAp; Chapter 2) and anterior (LHAa; Chapter 3) subregions of the LHA. I hypothesized that while chemogenetic excitation of LHAp-MCH à NAc neurons would fail to influence time-dependent food-seeking, excitation of LHAa-MCH à NAc neurons would prolong highly motivated food-seeking in female rats. In addition, I hypothesized that LHAa-MCH à NAc excitation would produce effects only when rats were tested during diestrus, the period of the rodent 30 estrous cycle when levels of circulating gonadal hormones are generally lower. Finally, in Chapter 4, I hypothesized that removal of peripheral gonadal hormones through ovariectomy would recapitulate the behavioral effect of LHAa-MCH à NAc excitation during diestrus, while estradiol replacement in ovariectomized (OVX) rats would blunt these effects. As expected, chemogenetic excitation of posterior LHAMCH à NAc neurons failed to produce behavioral effects on time-dependent food-seeking (Chapter 2). In addition, excitation of anterior LHAMCH à NAc neurons selectively produced post-criterion behavioral effects during diestrus (Chapter 3). Interestingly, however, these effects were in the opposite direction than hypothesized: chemogenetic excitation of LHAaMCH à NAc neurons reduced food-seeking after the omission of an expected food reward. Finally, removal of peripheral gonadal hormones through ovariectomy (OVX) failed to recapitulate the effects of LHAaMCH à NAc excitation observed during diestrus (Chapter 4). In oil pretreated rats, chemogenetic excitation of LHAaMCH à NAc neurons failed to influence PI responding, whereas this excitation reduced post-peak responding in EB pretreated rats. Thus, excitation of LHAaMCH à NAc neurons in EB pretreated rats produced a behavioral phenotype similar to that observed in diestrus females. These results suggest that estradiol is necessary for the behavioral effects of LHAaMCH à NAc neurons, but raise new questions about the timing of estradiol in its influence on MCH. Implications of these results are discussed in Chapter 5. 31 CHAPTER 2: Accumbens-projecting Melanin Concentrating Hormone neurons that originate in the posterior Lateral Hypothalamic Area do not influence motivation or time perception Abstract Temporal information can be processed across a wide range of timescales, endowing the capacity for an organism to regulate its internal milieu as well as predict and adapt to the external environment. Previously, I demonstrated that neurons that produce the orexigenic peptide Melanin Concentrating Hormone (MCH) may putatively influence time perception. Specifically, chemogenetic excitation of MCH neurons in the posterior lateral hypothalamic area (LHApMCH neurons) altered the timing of responding in a time-dependent food-seeking task, the Peak Interval (PI) paradigm. Excitation of LHApMCH neurons resulted in responding that reached a peak rate at an earlier time (i.e., peak time) and attenuated more quickly compared to when rats were tested under control conditions. These effects were significant in female – but not male – rats, and potentially driven by a subtle decrease in post-peak responding that occurred after the criterion duration. To examine the circuitry underlying these effects in female rats, I examined whether LHApMCH neurons that project to the nucleus accumbens – a ventral striatal region important for motivated behavior, feeding and timing – could influence how female rats perform in the PI paradigm. Using a dual virus approach, I selectively excited LHApMCH neurons that project to the NAc and examined their influence on behavior in the PI task across the estrous cycle. I hypothesized that LHApMCH neurons that project to the NAc would not influence time perception in the PI paradigm, per se, but could potentially modulate motivated responding by influencing the “stop” function. 32 Chemogenetic excitation of LHApMCH neurons failed to influence time-dependent food seeking behavior in the PI task. Moreover, there were no effects of this excitation on either peak time or on motivated responding as revealed by the “stop” function. Thus, the effects of LHApMCH neurons on interval timing do not rely on efferents to NAc and may instead reflect control from alternative downstream dorsal striatal or hippocampal targets. 33 Introduction Core components of metabolic regulation have traditionally been studied within the context of circadian timing, whereby biological rhythms are evoked by brain oscillations coupled to the 24-hr light-dark cycle (Bass & Takahashi, 2010; Turek et al., 2005) via orchestration of the suprachiasmatic nucleus (SCN) (Reppert & Weaver, 2002). Alternatively, interval timing operates in the seconds-to-minutes range, and depends on corticostriatal rather than SCN-dependent modulation (Lewis & Miall, 2003; Mello et al., 2015). The ability to perceive time in the seconds-to-minutes range enables animals to learn predictive relationships and adapt to a changing environment. Importantly, time perception in this range supports temporal contiguity, learning, and decision making (Kacelnik & Brunner, 2002; Marshall, Smith, & Kirkpatrick, 2014; Meck et al., 2012). The Peak Interval (PI) paradigm is an interval timing task in which animals learn to predict when food reinforcement will be available based on the temporal context provided by a discriminative stimulus. Adept time perception in this task allows an animal to coordinate effortful behavior (i.e., instrumental responding) at times when reinforcement is most likely. Because this task inherently involves food intake and learning about food-predictive cues, I hypothesized that neurons that influence feeding behavior may also influence the time-dependent food-seeking observed within this task. Within the lateral hypothalamic area (LHA), neurons that produce the orexigenic peptide Melanin Concentrating Hormone (MCH) are known to influence learned food intake. In addition, these neurons can influence the timing of instrumental behaviors performed in anticipation of food. For example, in male rats, LHAMCH neurons that project to the hippocampus (HPc) increase early responding for sucrose reinforcement, 34 an effect that has been described as increased impulsivity (Noble et al., 2019). However, increased early responding in this task could also indicate an effect of LHApMCH neurons on time perception whereby rats perceive time as having passed more quickly. For example, an increase in internal clock speed could alternatively account for early lever pressing in this task by causing a rat to perceive the 20s criterion as having already passed faster (e.g., at 18s). In order to examine whether LHAMCH neurons could influence time perception in this manner, I used chemogenetics to selectively excite LHApMCH neurons while rats timed a 20s criterion in the PI paradigm. Chemogenetics refers to a class of genetically modified receptors, i.e., Designer Receptors Exclusively Activated by Designer Drugs, or DREADDs, that bind to otherwise inert exogenous chemical ligands like clozapine-N- oxide (CNO) to alter cellular excitability (Roth, 2016). These DREADDs can be packaged into adeno-associated viruses (AAVs) and injected into the brain within regions of interest. Control of DREADD expression by a genetic promoter – in this case, the pMCH promoter – allows the DREADDs to be expressed only in cells of a certain type (e.g., those that contain the pMCH promoter and are capable of producing the MCH peptide). In this case, I used a DREADD virus in which the expression of an excitatory, modified human muscarinic receptor (hM3Dq) was controlled by the pMCH promoter. Previously, I demonstrated that chemogenetic excitation of LHApMCH neurons reduced peak time, suggesting that rats potentially perceived time as passing more quickly following the excitation of these neurons. In addition, when we plotted responding across time as a proportion of peak rate, females – but not males – 35 demonstrated a significant effect of CNO on responding across time. In females, the proportion of peak rate responding under CNO was higher than VEH at 18s, but lower than VEH at 24, 29, and 48s. In other words, CNO-treated females displayed a proportional leftward shift in the timing function, such that they responded at a higher proportion of peak rate before the criterion, and lower proportion of peak rate after. These findings suggest that in female rats, LHApMCH neuronal stimulation accelerated clock speed such that the rats underestimated the criterion duration. In this previous study, I also examined whether the effects of chemogenetic excitation differed when ovarian hormone levels were relatively high (i.e., during proestrus/ estrus or P/E) or low (i.e., during metestrus/ diestrus or M/D). Although there were no significant effects of LHApMCH neuronal excitation on the proportion of peak rate responding within M/D or P/E, it is possible that our retrospective approach – which limited the analysis to a small sample size – may have precluded us from observing an effect of chemogenetic excitation based on estrous stage. While interval timing likely involves a vast network of corticostriatal circuits, a role for the ventral striatum (VS) has been identified in the acquisition of the “stop” function (MacDonald et al., 2012). Thus, effects of chemogenetic excitation of MCH neurons on the “stop” function may occur due to activity of MCH in the VS. Notably, within the VS, the MCH receptor MCH1R is densely expressed in the nucleus accumbens (NAc), which also has an important role in motivated behavior, including MCH-mediated feeding (Berridge, 2004; Floresco, 2015; Georgescu et al., 2005; Kelley, 2004). I thus hypothesized that a portion of these LHAMCH neurons may project to the NAc to influence time-dependent food-seeking. 36 To address these questions, in this chapter I used a dual-viral, chemogenetic approach to express an excitatory DREADD receptor only in MCH neurons within the LHAp that project to NAc. Similar to previous studies, I injected an AAV containing an excitatory DREADD. However, in this case the expression of the modified hM3Dq receptor depended on both cre recombinase and the pMCH promoter. Thus, the hM3Dq receptor could be expressed only in cells that contain both cre recombinase and pMCH. Therefore, in order to selectively transfect only MCH neurons that project to the NAc, I also injected a retrograde virus containing cre recombinase, which will traffick cre from the NAc to neurons that project to this region, including cells in the LHA. In this manner, I selectively excited LHApMCH neurons that project to the NAc (LHApMCHàNAc) during the PI paradigm to investigate their role in time-dependent food seeking across the estrous cycle. Materials & methods Subjects Eight adult female Sprague-Dawley rats (Envigo, Haslett, MI, USA; 12-weeks of age at arrival) were pair housed in groups of 2-3 in standard, plexiglass cages with metal tops. Rats were maintained on a standard 12-hr light-dark cycle (lights on 7:00; lights off 19:00), with ad libitum access to Teklad diet #8940 and reverse osmosis (RO) water. Rats received ³7 days of acclimatization to the vivarium before experimental manipulations began. Following this period of habituation, rats were handled daily for 2- 3 days before undergoing stereotaxic surgery. Post-op, rats were briefly singly housed while they received daily health monitoring. Rats were pair housed with their original cage mate once postoperative bodyweight recovered (≤7 days) and surgical incisions 37 appeared healed. Rats continued to be pair-housed throughout all behavioral experiments. All manipulations were conducted in compliance with the Institutional Animal Care and Use Committee, Michigan State University. Surgical procedures Stereotaxic Viral Infusion and Cannulation Under 2-4% isoflurane anesthesia, subjects received bilateral infusions of the retrograde AAV2(retro)-eSYN-EGFP-T2a-icre-WPRE (Addgene, Watertown, MA) and a cre-dependent, excitatory DREADD AAV2-DIO-rMCHp-hM3D(Gq)-mCherry (gift from Dr. Scott Kanoski) to the NAc and to the LHAp, respectively (Table 2.1). A single, 26- gauge guide cannula (Plastics1, Roanoke, VA) consisting of an 8 mm plastic pedestal and containing a 2.5 mm projection below the base was placed -1.2 A.P. ±2.25 M.L., - 2.5 D.V. This enabled intracerebroventricular (ICV) infusions of the ligand, clozapine-N- oxide (CNO; NIDA Drug Supply Program) to the lateral ventricle. Placement of the guide cannula into the left or right lateral ventricle was counterbalanced between animals. The addition of two surgical screws (Fine Science Tools, Foster City, CA), Loctite superglue (Amazon, Seattle, WA), and dental acrylic (Lang Dental, Wheeling, IL) was used to stabilize the base of the cannula and ensure closure of the surgical space. To prolong stable placement of the guide cannula, skin was sutured over the dental acrylic to prevent new skin growth from displacing the cannula. Guide cannulae were protected with a dummy cannula (Plastics1, Roanoke, VA) cut to fit the 2.5 mm guide without a projection. Rats were treated with 2 mg/ kg meloxicam prior to surgery and as needed during the following week of post-operative monitoring to manage pain. Rats typically 38 did not require more than one additional dose of meloxicam. Unfortunately, one rat had an isoflurane reaction and died during surgery. Table 2.1 Viral approach to selectively target LHApMCH neurons that project to the NAc. Virus Target Infusion coordinates AAV2(retro)-eSYN-EGFP-T2a-icre-WPRE NAc Shell +1.1 A.P., ±0.8 M.L., −7.5 D.V. 0.3 µl / infusion AND -2.6 A.P., ±1.8 M.L., -8.0 D.V. AAV2-DIO-rMCHp-hM3D(Gq)-mCherry MCH -2.6 A.P., ±1.0 M.L., -8.0 D.V. LHAp 0.3 µl / infusion -2.9 A.P., ±1.1 M.L., -8.8 D.V. -2.9. A.P., ±1.6 M.L., -8.8 D.V. Behavioral Paradigm Following recovery from viral infusion and food restriction to 90% baseline weight, rats were trained and tested in the Peak Interval (PI) paradigm. There were three phases of behavioral training and testing, described in detail below. Rats were weighed daily between ~9 – 10 am and when applicable, vaginal lavages were also performed at this time. Behavioral sessions began between 9:30 – 10:30 am (group 1) or 11:30 am – 12:30 pm (group 2) and were run 5-7 days per week throughout training and testing. Days off between behavioral training or testing occurred only when they would provide minimal interruption, i.e., between consecutive days of FI or PI training or between washout days of testing. Due to the length of Peak Interval paradigms, it is typical to run behavioral sessions only 5 days / week and thus these brief interruptions were not expected to influence behavior. Daily chow rations were provided to rats following the completion of the behavioral session. 39 Phase 1: Pre-training Sucrose habituation To reduce neophobia to the 20% w/v sucrose solution used as a reinforcer, rats were given 15 minutes of sucrose pre-exposure in their home cages. Pair-housed rats were separated into clean cages and allowed to rest in their new cages for ≥30 minutes. Water bottles were removed from the cages, and identical bottles instead filled with ≥50 ml of a 20% w/v sucrose solution were placed on the opposite side of the wire food hopper. Consumption was observed to ensure that all rats consumed sucrose freely before the session ended. Bottles were weighed before and after consumption to confirm that rats had consumed ≥ 10g of the sucrose solution. All rats consumed promptly and met the minimum consumption criteria. After the habituation session, rats were returned to pair housing and fed their daily chow ration. Magazine training After sucrose pre-exposure, all behavioral assays were conducted in eight standard operant boxes contained within sound-attenuating cabinets (Med Associates). Boxes were equipped with a recessed food magazine, into which liquid reward solutions could be delivered via automated pumps. Solutions were delivered into a clear acrylic food well located within the recessed food magazine. Infrared (IR) cameras were mounted below the food well to enable consumption to be seen and recorded. In addition, an IR light across the magazine port enabled recording of magazine entries and the overall time spent in the food magazine. Boxes were also equipped with two levers, placed on either side of the food magazine. During magazine training, levers 40 were recessed. In addition, a house light located in the upper corner of the sound- attenuating chamber illuminated the box with red light during magazine training. During magazine training, rats were provided with 16 presentations of the sucrose reinforcer (20% sucrose solution) on a random time 240s reinforcement schedule; sessions lasted approximately 48 – 60 minutes. A brief click produced by the activation of a solenoid located behind the food magazine occurred simultaneously with pump activation and reinforcer delivery to help orient rats toward the food magazine. Rats were required to meet the criteria of having spent ≥10s in the food magazine during reinforcer delivery in order to move onto lever training. Rats typically met criteria within the first magazine training session; however, all rats were given two magazine training sessions to ensure they met criteria. Lever training Rats next received baited lever training in which levers were presented for 25 trials and reinforcement was provided on an FI20 schedule. Levers were baited with a slurry of chow that was made by mashing chow in RO water and then applied to the top and bottom of the extended lever. Baited levers were active and available to rats immediately when rats were placed into the operant box. Lever position relative to the food magazine (left or right) was counterbalanced. The red house light, mounted inside the sound attenuating chamber, was illuminated throughout the lever training session. Rats were required to complete ≥10 of 25 trials during lever training. Rats who failed to meet criteria received additional lever training as needed prior to advancing to the next phase of training. Lever training sessions lasted no longer than 25 minutes; however 41 most rats quickly acquired the instrumental response and finished lever training in 10-15 minutes. Phase 2: Peak Interval training Fixed Interval (FI) Training After meeting lever training criteria, rats moved to fixed interval (FI) training, where they were taught to time a criterion duration of 20s. At this point, the red house light was replaced with a white light. Sessions were dark except during the FI trials, when light illumination and lever presentation indicated the onset of the trial and the to- be-timed period. During FI trials, only the first lever press that occurred at or after the 20s criterion was reinforced with reinforcer delivery. Early lever responses (i.e., before 20s) were neither reinforced nor punished. Trials ended with simultaneous sucrose delivery, lever retraction, and offset of the light. Rats received 10 sessions of FI training; there were 50 FI trials within each session separated by a variable inter-trial interval (ITI) that averaged 60s. Sessions terminated as rats finished the 50th trial or after 120 minutes had elapsed, whichever came first. As rats acquired the criterion duration, they reliably completed the FI sessions within approximately 70 minutes. Peak Interval (PI) Training After FI training, rats received 16 Peak Interval (PI) training sessions. As during FI training sessions, rats also received 50 trials during PI training. However, during PI training 25 probe trials were randomly intermixed with 25 FI trials. Probe trials were identical to FI trials in that they onset with illumination of the house light and presentation of the lever. However, probe trials were unique in that no lever presses were reinforced, regardless of if they occurred at or after the criterion duration. Instead, 42 probe trials lasted at least 3x the length of the criterion time (i.e., 60s). This enabled examination of responding before, during, and after the criterion had elapsed. (a) (b) (c) Figure 2.1 The Peak Interval Paradigm. Conducted in standard operant boxes (a), the peak interval paradigm trains rats to time a 20s criterion duration during fixed interval (FI) trials in which the first lever press at or after 20s results in sucrose reinforcement (b). Following FI training, probe trials are randomly intermixed with FI trials in a 1:1 ration. Probe trials onset in the same manner as FI trials, but last at least 3x the length of the criterion duration and differ in that no lever presses are reinforced. Responses during each trial were time-stamped with centi-second precision, binned within 1s bins, and plotted across time. The total number of responses per bin were normalized as a function of the maximum number of responses that occurred in one bin; this produced the proportion of peak rate response function. The time at which maximum responding occurred was labelled peak time. PI training was grouped into four blocks (PI session 1-4, 5-8, 9-12, and 13-16). Proportion of peak rate response functions were averaged across training blocks to confirm acquisition of the criterion duration. Estrous cycle tracking began during PI training; vaginal lavages were performed prior to behavioral training when rats were weighed, as described in Vaginal Cytology, below. During the final three PI training sessions (i.e., PI session 14 – 16), subjects 43 received mock drug delivery in order to habituate them to ICV infusion and intraperitoneal (i.p.) injections. Drugs administration occurred in a separate room located across the hall from the behavioral boxes; beginning with PI session 14, subjects were placed in this room for approximately 15 minutes prior to behavioral sessions regardless of if drugs were administered (i.e., including during washout days). During this period of drug delivery habituation, cannula placement and patency was verified, and rats were assigned to either intracerebroventricular (ICV) infusion or intraperitoneal (i.p.) injection groups. Preference was given to ICV infusion, however at this point cannula had been implanted for a minimum of 35 days and some had sustained damage from chewing by pair-housed rats. At this stage, n=2 rats were assigned to the i.p. injection group; n=5 rats received ICV infusions. Phase 3: Peak Interval Testing During the final phase of behavior, subjects continued to receive PI sessions identical to PI training. However, during test sessions, rats received administration of either vehicle (VEH; 0.2M PBS) or clozapine-N-oxide (CNO) prior to the behavioral session. Rats were tested twice each under VEH and CNO in proestrus/ estrus (P/E) and metestrus/ diestrus (M/D). Cycle stage was determined via vaginal cytology prior to daily testing, and drug delivery was determined based on previous test conditions. Efforts were made to counterbalance the order of VEH and CNO administration within P/E and M/D when possible. At least 72 hours and two “washout” PI sessions were given between each drug treatment. 44 Histology Vaginal Cytology Estrous cycle tracking was performed daily at 2 ½ - 4 hours after lights on, i.e. between 9:30 – 11:00 am. Thus, cycle samples were taken ~15-45 minutes prior to behavioral training and testing, which accommodated time for drug administration after vaginal lavages were taken and evaluated. Samples were collected via vaginal lavage, a process which involves gently inserting a P1000 pipet tip into the vaginal canal to flush, and immediately withdraw, 0.9% sterile saline solution. Vaginal samples were placed into a 12-well plate for transport and then imaged wet on an Olympus BX53 microscope. Categorization of vaginal cells into estrus cycle stage was performed based on the proportion of leukocytes, cornified, and epithelial cells. For the purpose of data analyses, estrous cycle stages were grouped into periods when gonadal hormones are generally low (e.g., metestrus, diestrus = M/D) and when gonadal hormones are generally high (e.g., proestrus, estrous=P/E). Once tracking began, estrous cycles were tracked continuously throughout behavioral training and testing. Perfusion and tissue collection Euthanasia was performed via fatal overdose of sodium pentobarbital (50mg/ kg), followed by transcardial perfusion with 0.9% saline and 4% paraformaldehyde (PFA). Brain tissue was extracted and post-fixed overnight in 4% PFA with 12% w/v sucrose. Tissue was frozen on dry ice and stored at -80°C until slicing on a standard microtome. Five representative tissue series were sliced at 30 µm; the first of five series was mounted immediately upon slicing. The remaining four series were stored in cryoprotectant at -20°C until later processing. 45 Tissue analysis and immunohistochemistry Tissue was sliced at 30 µm on a standard sliding microtome into 5 serial sets and stored at -20°C in cryoprotectant. The first representative was mounted in full immediately upon slicing and examined for viral expression of the GFP-cre in the NAc and mCherry-hM3Dq in the LHA. A second series was used for confirmation of DREADD expression in MCH+ cells via dual immunofluorescent staining for MCH protein and the mCherry viral tag. Tissue was first rinsed in 6x 8 min washes of 0.1 M phosphate buffered saline (PBS), then incubated in a 0.1 M PBS solution containing 0.03% hydrogen peroxide for 20 min. Next, tissue was washed in 10 min rinses each of 0.1M PBS first containing 0.3% glycine, then containing 0.03% sodium dodecyl sulfide. Tissue was blocked in a blocking solution made of 0.1 M PBS and containing 0.3% normal donkey solution and 0.15% Triton-X. Finally, tissue was incubated overnight in blocking solution with rabbit anti-MCH [1:1000] (Phoenix Pharmaceuticals, Catalog #H- 070-47; Burlingame, CA) and goat anti-RFP [1:1000] (Rockland Antibodies, Catalog #200-101-379; Pottstown, Pennsylvania). On the following day, after a series of six 0.1 M PBS rinses, secondary antibody incubation occurred with [1:200] each of AlexaFluor 488 Donkey anti-Rabbit IgG and AlexaFluor 568 Donkey anti-Goat IgG in blocking solution for a period of 2 hours. Finally, tissue was rinsed twice more in 0.1 M PBS. All staining occurred at room temperature. Following staining, tissue was float-mounted in 0.1M phosphate buffer solution onto gelatin-subbed slides and air dried prior to applying coverslips with Pro-Long Gold Antifade Mountant containing DAPI (Thermofisher). Slices were imaged on an Olympus BX51 epi-fluorescent microscope equipped with DAPI, FITC and CY3 filters and 46 connected to an IBM-compatible Windows 10 computer with Neurolucida imaging software (MBF Bioscience, Williston, VT). Statistics Data transformation Behavioral data was imported into Microsoft Excel using table profiles built in MedPC2XL (Med Associates). This template sorted data to indicate the start of each of 50 trials, as well as the time of discrete lever presses within a session. From here, data was imported to SPSS where syntax was run to sort lever responses by trial in order to examine when each lever press occurred within each individual trial. For each subject, responses per trial were then totaled into 1s bins across all probe trials (n=25) in a session to determine total responding per 1s bin. The first 1s bin in which peak rate occurred indicated the peak time of the session. Peak time indicates the time at which the subject perceives the to-be-timed criterion as having elapsed and reflects the accuracy of temporal perception (Church, 1984; Church & Broadbent, 1991; Meck, 1996; Roberts & Church, 1978). In order to examine responding in the context of peak rate, responding was normalized as a percentage of maximum peak rate, to provide a proportion of peak rate function. To examine the influence of estrous cycle stage, behavioral tests were grouped by periods of the estrous cycle when gonadal hormones are typically low (i.e., metestrus/ diestrus, M/D) or typically high (i.e., proestrus/ estrus, P/E). I first conduced repeated measures analysis of variance (RM ANOVA) with female rats that met the criterion of having been tested under both VEH and CNO during both M/D and P/E (i.e., only rats that were appropriately cycling and had both P/E VEH and CNO as well as 47 M/D VEH and CNO tests were included). During the 20s PI testing, all rats were cycling regularly and included in this analysis (n=7 ♀). Responding/ bin was totaled across the two sessions in each condition (i.e., 50 trials from two P/E VEH tests) prior to normalization. To procure group normalized functions, individual normalized functions were first averaged and then normalized as a function of group peak rate. In addition, predicted proportion of peak rate response functions were modeled using a multivariate, piecewise growth model (PGM). In this case, responding was normalized within session (i.e., over 25 trials) but the model incorporated two sessions under each drug condition (i.e., two P/E VEH, two M/D VEH, two P/E CNO and two M/D CNO tests). Analysis To examine broad changes in behavior that resulted from the estrous cycle or chemogenetic manipulations, I examined the amount of time spent in the food cup (percent time) as well as the average lever response rate (rate/ min) during a session. I also evaluated whether peak time (i.e., the time at which subjects responded maximally during PI trials) differed as a function of estrous cycle stage or chemogenetic excitation of LHApMCHà NAc neurons. Paired t-tests were performed using GraphPad Prism (Graphstat Technologies, Bangalore, India). Effect sizes were calculated as Cohen’s d for paired t-tests. Next, I evaluated how responding differed across time within individual trials by examining whether the proportion of peak rate response function differed as a function of estrous cycle stage (P/E, M/D) or drug treatment (VEH, CNO) within P/E and M/D. The proportion of peak rate response function, which plots response rate across time as a percent of peak rate, was evaluated with repeated measures analysis of variance 48 (ANOVA). I first examined the effects of estrous cycle stage under baseline (vehicle) conditions before separately examining the effects of drug treatment (VEH, CNO) within each estrous cycle stage. The α level for significance was set to .05. Significant interactions were examined using Bonferroni corrected pairwise comparisons to examine when responding differed between treatment conditions in each of sixty 1s time bins. Analyses of variance were performed using Statistica (Statsoft, Tulsa, OK) and SPSS version 28 (IBM, Armonk, NY). Finally, I used multilevel piece-wise growth models (PGM) to model the predicted proportion of peak rate responding that would occur before and after peak time under each treatment condition. The peak time was defined as the first instance of the rat’s maximal response, and all pre- and post-peak effects were examined relative to each subject’s peak time. While the ANOVA examines differences in level or magnitude across time, the PGMs examine differences in magnitude (i.e., intercept) as well as changes in rate (i.e., slope) over time. Each PGM included linear, quadratic, and cubic rates of change across time to predict how the proportion of peak rate response function changed within each treatment condition. These models are particularly sensitive to changes in rate, and could therefore identify subtle differences in the rate of increase or decrease that occurred pre- and post-peak, respectively, due to treatment. The PGMS emphasize the distinction between pre- and post-peak periods of the trial because these periods represent distinct phases of motivated behavior. While the pre-peak period represents an increase in responding in anticipation of reinforcement (i.e., the “start” function), the post-peak period instead represents a decrease in responding following the omission of an expected reward (i.e., the “stop” function). By modeling these pre- 49 and post-peak periods separately in a piecewise manner, I was able to examine how the rate of responding changed in each period accordingly based on each rat’s individual perception of these two qualitatively different aspects of time. As with the ANOVA, I first examined the effects of estrous cycle stage (M/D vs P/E) under vehicle conditions by modeling the predicted proportion of peak rate responding under M/D and P/E following VEH treatment. Next, I separately modeled drug treatment (VEH vs CNO) within P/E, and then within M/D. Overall, pre-peak effects were examined at the mid-point prior to an animal’s peak time, i.e., time was “centered” before the peak. This pre-peak midpoint typically occurred around 10s. Post- peak effects were similarly centered at the midpoint between peak time and 60s, which typically occurred around 40s. The overall PGM thus specifically examines changes in the level and rate of predicted responding at approximately 10s pre-peak and 40s post- peak. A separate interaction model, which predicts estimates of intercepts and slope at these midpoints, was also computed for each PGM analysis. This analysis provided a more refined examination of changes in intercept and slope. I also separately modeled responding in 5s intervals by centering time at 5, 10, 15, 20, 25, 30, 35 and 40s in order to identify when changes to the intercept or slope occurred during pre- and post-peak periods. As in the overall model, responding was predicted in a piecewise manner with pre- and post-peak effects separately predicted based on each subject’s individual peak time. To accomplish this, I coded a new time variable such that time was centered at the interval of interest (e.g., 15s) appropriately before or after the peak for each subject. Thus, for a subject with a peak time of 13s, the 15s timepoint was centered as a post-peak effect, and the pre-peak time was centered 50 relative to the peak as in the overall model (i.e., 7s). On the other hand, a subject who responded maximally at 17s had 15s centered as a pre-peak effect, with post-peak time centered relative to the peak as in the overall model (i.e., at ~40s). Modeling time in this manner allowed level and rate differences to be estimated at each 5s interval across the trial. This provided more insight into when differences in the level or rate of responding specifically occurred during the pre- and post-peak periods. As in the case of the overall model, I first examined estrous effects under vehicle treatment (vehicle: estrous x time), then examined treatment effects within each estrous cycle stage (P/E: drug x time and M/D: drug x time) for each timepoint. For all analyses, the α level for significance was set to .05. Piecewise growth models were conducted using SPSS version 28 (IBM, Armonk, NY). Results Histology To confirm that intact rats were cycling normally, vaginal cells were collected daily via saline lavage. Samples were roughly categorized by the approximate proportion of cell types (e.g., leukocytes and epithelial cells) visible and at least two representative photomicrographs were taken per sample (Figure 2.2). Rats were immediately categorized into either P/E or M/D in order to assign drug conditions prior to behavioral testing. However, representative images from samples were later re- evaluated in context of the preceding and subsequent samples in order to confirm correct assignment of estrous cycle stage on test days. All n=7 rats were cycling appropriately during PI testing and are included in the behavioral results. 51 Figure 2.2 Photomicrographs of vaginal epithelial cells indicate the approximate proportion of cells present during the four stages of the rodent estrous cycle. (a) Proestrus vaginal cell samples consist primarily of larger, rounded epithelial (¨) cells. These cells often clump together in groups or strands. (b) During estrus, the cornification of vaginal epithelial cells produces both flat, cornified (#) and thin, needle- like cells that are highly keratinized (ò). Large round epithelial cells may also be present. (c) During metestrus, the proportion of leukocytes (o) increases relative to the other cell types. However, large epithelial cells in various stages of cornification persist. (d) Diestrus samples are characterized primarily by the presence of leukocytes. Cycle stages were approximated from the relative proportion of each cell type present on a given day, with consideration for preceding and subsequent days. 52 Tissue analysis confirmed bilateral expression of the GPF-cre virus in the NAc. Viral expression of the cre-dependent pMCH-HM3D(Gq)-mCherry virus was limited, but in line with previous reports indicating that this approach transfects only ~10% of MCH neurons (Noble et al., 2018; Terrill et al., 2020). Subjects were included as viral hits so long as they had clear GFP-cre expression in the NAc and evidence of the mCherry- HM3D(Gq) virus in the LHA shown through viral tracts, fibers, and a limited number of fluorescent-labeled cells. Figure 2.3 DREADD Expression in NAc-projecting LHAp neurons. (a) Representative photomicrograph of mCherry labelled DREADD expression in the LHAp. (b) Heat maps indicate representative DREADD expression (red) throughout the LHAp. Coronal sections modified from Paxinos & Watson 6th edition. Time-dependent food-seeking across the estrous cycle Analysis of the overall time spent in the food cup (percent time) or response rate (rate/ min) across the session revealed no baseline effects of estrous cycle stage on 53 behavior (Figure 2.4 a, b; percent time: t=1.237, df=5, p>.05, d=0.5; response rate: t=0.032, df=5, p>.05, d=.01). The time at which animals responded maximally within trials (i.e., peak time) also did not vary based on estrous cycle stage (Figure 2.4 c, t=0.00, df=5, p>.05, d=0.0). Analysis of the proportion of peak rate response function revealed only a main effect of time (Figure 2.4, d; F=138.34, df=(59, 295), p<.001, ηp2=.97), indicating that regardless of estrous cycle stage, rats significantly altered their response rate across individual trials in expectation of receiving reinforcement at the criterion time. Similar to the ANOVA, the overall PGM also revealed only effects of time when rats were tested under VEH across the estrous cycle. Before the peak, response rates increased in a primarily linear manner: i.e., there was a significant linear effect of time (F=65.27, df=1325, p<.001) pre-peak. After the peak, response rates declined at a rate that was determined by both linear and quadratic slopes (linear: F=32.69, df=499 p<.001; quadratic: F=14.10, df=637, p<.001). The overall model failed to reveal any estrous cycle effects pre- or post-peak. Results from the overall PGM and the estimates of intercept and slope predicted by the interaction model are reported in Supplemental Table S2.1 and S2.2, respectively. 54 (a) (b) (c) 15 15 25 P/E 20 M/D Peak Time (s) 10 10 Resp/ min 15 % Time 10 5 5 5 0 0 0 (d) (e) 1.0 1.0 Proportion of peak rate Predicted P/E M/D p<.05 0.5 0.5 proportion of peak rate 0.0 0.0 0 20 40 60 0 20 40 60 Time (s) Time (s) Figure 2.4 Effects of estrous cycle on responding under baseline conditions. Neither the total amount of time spent in the food cup, nor response rate across the session varied based on estrous cycle stage (a, b). Within individual trials, peak time (c) also did not vary between when rats were tested in M/D vs P/E. Finally, estrous cycle stage did not influence the proportion of peak rate responding across time (d), but analyses of predicted proportion of peak rate responding across 5s intervals revealed an effect of estrous cycle stage at 15s (e). Adopting the refined analysis of responding at 5s intervals (Supplementary tables S2.3 and S2.4) revealed an effect of time as in the ANOVA. In addition, this more sensitive analysis also revealed a main effect of estrous (F=3.89, df=170, p=0.05) and an interaction effect of estrous with the quadratic rate of change over time (F=7.95, 55 df=733.57, p=.005) at the 15s interval. Note that this is approximately the point in time when the predicted proportion of peak rate responding that occurs during P/E exceeded that predicted during M/D (i.e., where the lines representing predicted responding in each condition cross in Figure 2.4, e). This effect occurs in the post-peak component of the model, implying that it is driven by rats whose peak responding occurs prior to 15s (i.e., rats for whom 15s falls in the post-peak period). This includes data from n=3 M/D and n=2 P/E tests where rats peaked at or before 15s under vehicle treatment. While subtle, this estrous cycle effect suggests that responding may differ near the peak time based on estrous cycle stage. However, given that peak time itself does not vary as a function of estrous cycle, an estrous cycle effect may be limited to influencing the timing of “start” or “stop” functions without altering peak time, per se. Chemogenetic stimulation of LHApMCH à NAc neurons during P/E Next, I evaluated whether chemogenetic excitation of LHAMCH à NAc neurons influenced responding in the PI task when rats were tested during proestrus/ estrus (P/E). During P/E, there was no difference in the amount of time rats spent in the food cup (t=0.19, df=5, p>.05, d=.08; Figure 2.5 a) nor on the average response rate across the session (t=1.163, df=5, p>.05, d=.23; Figure 2.5 b) under VEH or CNO treatment. Within probe trials, peak time also did not vary following treatment with VEH or CNO (t=0.41, df=5, p>.05, d=.17; Figure 2.5 c). Likewise, the repeated measure ANOVA which examined the effects of drug and time revealed only a significant main effect of time (F=82.80, df=(59, 295), p<.001, ηp2 =.94) on the proportion of peak rate response function. As expected, rats changed their response rate during the trial in order to reach a peak rate around the expected time of 56 reinforcement. However, there was no effect of drug treatment (VEH or CNO) on the proportion of peak rate response function (F=5.48, df=(1, 6), p=.067, ηp2=.52). There was also no significant interaction of drug x time (F=0.99, df=(59, 295), p>.05, ηp2 =.17). (a) (b) (c) 15 15 25 VEH 20 CNO Peak Time (s) 10 10 Resp/ min 15 % Time 10 5 5 5 0 0 0 (d) (e) 1.0 1.0 Proportion of peak rate Predicted VEH CNO p<.05 0.5 0.5 proportion of peak rate 0.0 0.0 0 20 40 60 0 20 40 60 Time (s) Time (s) Figure 2.5 Chemogenetic excitation of LHApMCHàNAc neurons does not influence PI responding in P/E. The amount of time spent in the food cup and response rate across the session did not vary as a function of drug treatment when rats were tested during P/E (a, b). There was also no difference in peak time following CNO-mediated excitation of LHAMCH à NAc neurons (c). There were also no significant differences in the actual or predicted proportion of peak rate response functions (d, e). Piece-wise growth modeling of responding under VEH and CNO treatment during P/E also did not reveal any effects of drug treatment on the predicted proportion of peak rate response function. This model revealed only significant effects of time; there was a 57 significant effect of linear time (F=59.18, df=1365, p<.001) pre-peak and there were significant linear and quadratic effects of time post-peak (F=25.85, df = 536.238, p<.001 and F=14.22, df=688, p<.001, respectively). Results from the piecewise growth model are indicated in supplemental table S2.7; estimates of intercepts and slopes are depicted in S2.8. Results from the PGM examining the effects of drug treatment at 5s intervals are reported in supplemental tables S2.7 and S2.8. The closer inspection provided by examining data in this manner revealed a drug effect at 15s such that VEH-treated rats displayed a significant difference in the intercept (i.e., level) of responding pre-peak relative to CNO (F=3.99, df=260, p=.047). In addition, the post-peak quadratic rate of change at this timepoint was more negative following CNO administration (F=4.59, df=734, p<.05). This may reflect a slowing down of response rate around the time of the peak, which may drive the lower intercept that occurs at 15s in CNO-treated rats relative to VEH. Because this timepoint occurs just prior to the criterion, it may reflect a particularly sensitive period where subtle differences in motivation or expectation of reward may influence decisions to “start” and “stop” responding. Because the PGMs evaluate changes in linear, quadratic, and cubic slope, they may better capture these subtle changes in response rate that occur around the time of expected reward. Chemogenetic stimulation of LHApMCH à NAc neurons during M/D Rats tested during M/D also did not differ in the amount of time they spent in the food cup (t=.76, df=5, p>.05, d=.15), their rate of responding (t=1.49, df=5, p>.05, d=.20) or peak time (t=0.37, df=5, p>.05, d=.17) (Figure 2.6). 58 Analysis of the proportion of peak rate response function under each drug condition (VEH, CNO) also failed to reveal any significant effects of CNO on responding across time (Drug: F=0.026, df=(1, 5), p=.88, ηp2 =.00; Drug x time: F=0.73, df=(59, 295), p=.93, ηp2 =.13). (a) (b) (c) 15 15 25 VEH 20 CNO Peak Time (s) 10 10 Resp/ min 15 % Time 10 5 5 5 0 0 0 (d) (e) 1.0 1.0 Proportion of peak rate Predicted VEH CNO 0.5 0.5 proportion of peak rate 0.0 0.0 0 20 40 60 0 20 40 60 Time (s) Time (s) Figure 2.6 Chemogenetic excitation of LHApMCHàNAc neurons during M/D. There were no significant differences in response rate, time spent in the food cup, or peak time (a-c) when rats were treated with VEH or CNO. The actual and predicted proportion of peak rate response functions (d, e) also did not differ following CNO-mediated excitation of LHApMCHàNAc neurons. Modeling the predicted proportion of peak rate revealed no differences in the level of responding following CNO treatment in rats tested during M/D (see 59 supplemental table S2.9 and S2.10). Moreover, there were no differences in response rate or magnitude as revealed by the overall PGM and analyses at 5s intervals (Table S2.11 and S2.12). Interestingly, the lack of an effect at 15s is in contrast to subtle baseline (vehicle) estrous effects and to effects of chemogenetic excitation of LHApMCHàNAc neurons in rats tested during P/E. Discussion Summary of results There were no significant effects of chemogenetic excitation of LHApMCHàNAc neuronal excitation on interval timing, per se, in female rats. Excitation of LHApMCHàNAc neurons did not influence peak time or the proportion of peak rate response function in a manner indicative of a change in clock speed. Thus, in line with our expectations, LHApMCH àNAc neurons did not influence time perception in the peak interval task, suggesting that if LHApMCH neurons are indeed capable of influencing interval timing, they do so through an alternative target. This LHApMCHàNAc projection is not capable of influencing interval timing. Although there were no effects of LHApMCH àNAc neuronal excitation on interval timing, per se, there was a subtle effect of estrous cycle stage on responding during the PI task under baseline conditions. Specifically, when the predicted proportion of peak rate response function was modeled in a piecewise manner to examine the influence of estrous cycle on behavior at 5s intervals, estrous cycle selectively influenced responding in the 5s interval just prior to the criterion time (i.e., at 15s). Under baseline conditions, rats accelerated responding more quickly at 15s during P/E relative to M/D (i.e., there was a significant effect of estrous cycle on the quadratic slope). This resulted 60 in a higher level of responding under P/E than M/D at 15s, although the magnitude of this difference was not significant. This effect was also transient, as there were no significant differences in magnitude or slope at the next 5s interval at 20s. Rats thus differed in their pre-peak responding immediately prior to the criterion time depending on estrous cycle stage but exhibited no differences in the magnitude or rate of responding at the criterion itself. In line with this, there was also no difference in peak time when rats were tested in M/D compared to P/E. Altogether, these findings suggest that there may be subtle differences in the rate of PI responding based on estrous cycle stage, but that these effects are not driven by a change in time perception. To my knowledge, only one other study to date has examined estrous cycle effects on timing performance in intact female rats (Panfil et al., 2023), and has reported mixed effects of cycle stage on timed performance. However, accumulating evidence from studies examining the effects of exogenous estradiol (EB) replacement in gonadectomized rats suggests that there may be multiple mechanisms of EB on timing (Pleil et al., 2011; Ross & Santi, 2000; Sandstrom, 2007), including both acute and phasic effects. Evidence indicating that long-term estradiol treatment does not influence time perception (Ross & Santi, 2000), suggests that any baseline effect of fluctuating gonadal hormones on PI performance may be subtle, especially in well-trained rats who have repetitively experienced the task in all phases of the estrous cycle. Thus, while acute estradiol replacement in OVX rats accelerates clock speed (Sandstrom, 2007), these effects are transient and may not be representative of the effects of endogenous fluctuations of estrogen in intact rats. For example, Ross and Santi (2000) reported that two weeks of EB replacement decreased discrimination accuracy but did not influence 61 time perception in a duration discrimination task. Likewise, Pleil et al. (2011) found that cyclic EB replacement in OVX rats produced a clock speed effect only during the first hormone cycle. These authors attribute this effect to previous experience of the task following EB pretreatment and/or the extensive training rats received between testing cycles (Pleil et al., 2011). Given that rats in the present experiment had ample experience performing the task during both M/D and P/E, they may have learned to rapidly and flexibly adjust responding to compensate for any hormone-induced changes in clock speed. In fact, that an effect of estrous cycle on PI responding at 15s was captured by the refined PGM analysis examining responding in 5s intervals speaks to the sensitivity of this analysis. This analysis was able to identify transient behavioral effects not revealed by the ANOVA or overall PGM because it separately examines pre- and post- peak effects at 5s intervals, rather than across the entire trial, or entire pre- vs post- peak period. This estrous effect was constrained to the period immediately preceding the peak, and occurred just prior to the criterion time, which is a behaviorally distinct period during the trial that coincides with the highest probability of reinforcement delivery. Thus, during this period the expectation of reward is highest and rats may be most motivated to respond for reinforcement. Thus, this refined analysis may capture transient changes in motivation that occur within a trial. In addition, the more sensitive, PGM analysis across 5s intervals also revealed an effect of chemogenetic excitation of LHApMCH àNAc neurons at 15s when rats were tested during P/E. In this case, CNO-mediated excitation of LHApMCH àNAc neurons altered both the level and rate of responding at 15s. Responding at 15s was significantly 62 greater in VEH-treated than in CNO-treated P/E rats pre-peak. Post-peak (i.e., for rats that had already peaked responding prior to 15s), the quadratic rate of decrease was steeper in CNO-treated rats. Thus, CNO-treated rats that had already reached a peak rate prior to 15s attenuated responding more quickly than VEH-treated rats at 15s. This suggests that chemogenetic excitation of LHApMCH àNAc neurons may subtly accelerate the “stop” function, as CNO-treated P/E rats “stop” responding more quickly at this time. As was the case with the baseline estrous cycle effect, chemogenetic excitation of LHApMCH àNAc neurons also only influenced responding in a manner that was both subtle and transient: neither the ANOVA nor the overall PGM revealed effects of drug treatment during P/E, and effects were only observed at 15s. Thus, the effect once again coincided with the period when rats perceived reinforcement delivery as being most likely and thus had high expectations of reward delivery. As motivated responding reaches a peak at this time, effects of LHApMCH àNAc neuronal excitation on motivation may be more easily revealed. Differences in responding that occur immediately before and after the peak indicate finite changes in motivated behavior coordinated around time of expectation – and subsequent omission – of reward delivery. Interestingly, in this case the effect of LHApMCH àNAc neuronal excitation on the magnitude of responding occurred pre-peak, whereas effects on the rate of responding occurred post- peak. This is possible because some rats had already reached a peak time at 15s, whereas other rats had not (i.e., for some rats 15s is pre-peak, whereas for others it is post-peak). Thus, In this case, chemogenetic excitation of LHApMCHàNAc neuronal excitation appears to modestly increase the level of pre-peak responding, while also 63 reducing the rate of response decay post-peak, resulting in a higher level of responding that rapidly decays. Although these effects are centered around 15s, they are subtle, and occur without influencing the peak time. Thus, they are not clearly indicative of an effect on time perception, but may reveal a subtle decrease in motivation. That these LHApMCH àNAc neurons may be capable of influencing motivated responding during P/E is unexpected given the inhibitory effect of estrogen on MCH. However, the subtle nature of the effect is in line with the idea that the effects of the neurons may be attenuated during P/E. Surprisingly, however, there were no effects of LHApMCHàNAc neuronal excitation on motivated responding in rats tested during M/D. This is unexpected not only because effects – albeit modest and transient – were observed in P/E, but also because the inhibitory influence of estradiol is absent during M/D. In other words, although there was a subtle effect of LHApMCHàNAc neuronal excitation on post-peak response rate during P/E, there was no effect of the same excitation during M/D. This suggests provides additional evidence for estrous cycle effects on PI performance in intact female rats and further indicates that LHAMCH neurons that project to the NAc interact with circulating ovarian hormones. Limitations Despite the subtle effects captured by the refined PGM analysis in this study, which indicated potential estrous cycle and chemogenetic effects of LHApMCHàNAc neuronal excitation at 15s, there was little evidence of a role for these neurons in influencing PI responding. Regardless, this general lack of effect should be interpreted with caution, as only a small proportion of LHAp neurons expressed the mCherry 64 fluorophore indicative of successful DREADD expression in the LHA. While a limited number of mCherry-labelled cells were present in the expected LHAp regions (see Figure 2.3), mCherry-labelled fibers were more apparent than cell bodies. Poor expression of the mCherry fluorophore persisted even after amplification using immunohistochemistry, which limited the extent of histological analyses included. These issues also made it difficult to examine MCH protein expression in mCherry, DREADD- expressing neurons. Regardless, that there is no overall effect of CNO administration in these rats, particularly during M/D, indicates that even if an insufficient number of cells expressed the DREADD for a behavioral effect, there were also no off-target effects of CNO. Conclusion In conclusion, chemogenetic excitation of LHApMCHàNAc neurons failed to influence timing in the PI paradigm, indicating that this projection is not capable of altering time perception. In addition, there were no effects of LHApMCHàNAc neuronal excitation on motivated responding during M/D. There were subtle effects of chemogenetic excitation of LHApMCHàNAc neurons on the rate of responding around 15s when rats were tested during P/E, potentially indicating an acceleration of the “stop” function during P/E. There was also a baseline effect of estrous cycle stage at 15s such that P/E rats attenuated high rate responding more quickly (i.e., the “stop” occurred more abruptly). In both cases, these subtle effects are not indicative of a change in time perception, but may reveal a slight attenuation in motivation to food seek during P/E. 65 CHAPTER 3: MCH Neurons in the anterior LHA interact with estrous cycle stage to influence motivation in a time-dependent manner Abstract In order to appropriately coordinate motivated behavior, an individual must decide when to start or stop behavior using information from the local environment. Temporal information, which allows an individual to understand predictive relationships between stimuli and outcomes, is particularly important for learning and decision making. Previously, I demonstrated that chemogenetic excitation of cells that produce the appetite-stimulating neuropeptide Melanin Concentrating Hormone (MCH) could influence time-dependent responding in female rats tested in the Peak Interval (PI) paradigm. Specifically, chemogenetic excitation of MCH neurons delayed when female rats stopped responding after the omission of an expected reward during probe trials. In other words, MCH neurons prolonged high rate responding in female rats, indicating an increase in motivation to continue working for reinforcement. This influence of MCH neurons on motivated behavior may reflect a role for MCH in the nucleus accumbens (NAc), a region important for both motivated behavior and the decision to “stop” responding in interval timing tasks (Floresco, 2015; Kelley, 2004; MacDonald et al., 2012). Thus, in this chapter I examined whether MCH neurons that project to the NAc would likewise prolong motivated responding in the PI paradigm. To examine the influence of these cells on motivation, I used chemogenetics to selectively excite NAc- projecting MCH neurons while female rats were tested in both the PI paradigm, as well as a more typical task used to study broad features of motivation: the progressive ratio (PR) task. Rats were tested during periods of both low (i.e., metestrus/ diestrus, M/D) 66 and high (i.e., proestrus/ estrus, P/E) circulating gonadal hormones. I hypothesized that chemogenetic excitation of NAc-projecting MCH neurons would prolong high rate responding in the PI paradigm, reflecting an increase in motivation. Furthermore, because estradiol inhibits the activity of MCH and because chemogenetic excitation of MCH neurons previously prolonged high rate responding predominantly during M/D, I hypothesized that projection-specific excitation of these neurons would also selectively produce effects when rats were tested during M/D. In line with my hypotheses, excitation of NAc-projecting MCH neurons influenced motivation in the PI task by altering post-peak responding, and this effect occurred in M/D. However, contrary to my initial hypothesis, excitation of NAc-projecting MCH neurons decreased post-peak responding, reflecting a decrease in motivation to work for an omitted food reward. Interestingly, these effects were limited to the post-peak period (i.e., they were temporally selective) and did not influence responding in the PI task overall. In line with this, there was no effect of LHAaàNAc neuronal excitation on PR responding, indicating that these effects are selective to the timing – rather than overall rate – of motivated responding. The activation of LHAaMCH neurons appears to alter the decision of when to “stop” responding for a sucrose reward. In contrast to my previous studies, in which activation of all LHAaMCH neurons prolonged high rate responding, activation of only those LHAaMCH neurons that project to the NAc instead decreased high rate responding. In both cases, these effects were limited to the post-criterion period, and occurred selectively when cells were activated during M/D. Altogether, these results indicate that LHAaMCHàNAc neurons interact with the estrous cycle to influence 67 decisions about how to respond for food within a temporal context that predicts food availability. 68 Introduction Previously, I demonstrated that LHAaMCH neurons could delay the “stop” function in female – but not male - rats. Using a chemogenetic approach, I selectively excited MCH neurons in the anterior LHA (LHAa) in male and female rats tested in the Peak Interval (PI) paradigm. The excitation of these neurons had no influence on peak time, indicating that these neurons did not influence time perception, per se. However, chemogenetic excitation of LHAaMCH neurons selectively prolonged high rate responding after the criterion in female rats. Moreover, this effect occurred only during metestrus/ diestrus (M/D), when circulating levels of ovarian hormones are typically lower than during proestrus/ estrus (P/E). In addition, neither peak time nor responding prior to the criterion time were affected, indicating an intact “start” function and accurate time perception. These findings not only indicate that LHAaMCH neurons interact with the estrous cycle, but also suggested a role for the nucleus accumbens (NAc), a ventral striatal region important in both motivated behavior and the “stop” function of interval timing (Floresco, 2015; MacDonald et al., 2012). Within interval timing procedures, the “start” and “stop” functions are described as dissociable decision processes wherein a rat begins to respond at a high rate prior to the expectation of reward and then stops high rate responding after the reward has been omitted, as occurs in PI probe trials (Balcı, 2014; Church et al., 1994; Church & Broadbent, 1991; Gallistel & Gibbon, 2000; Gibbon, 1977; MacDonald et al., 2012). The “stop” function relies on an intact ventral striatum (MacDonald et al., 2012), which is also an area instrumental in motivated behavior and food intake (Floresco, 2015; Kelley, 2004; Kelley et al., 1996, 2005; Stratford & Kelley, 1997). Indeed, the a delay in the 69 “stop” function can be interpreted as an increase in motivation to continue responding for an omitted food reward, and changes in reward magnitude can separately influence the “start” and “stop” functions without influencing time perception (Galtress & Kirkpatrick, 2009; Roberts, 1981). For instance, decreasing reward value by devaluing the food reinforcer used in timing tasks generally delays the “start” function (Galtress & Kirkpatrick, 2009; Roberts, 1981). This delay would be interpreted as a decreased motivation to respond for the devalued reinforcer and occurs without influencing time perception itself. On the other hand, continued high rate responding after the omission of an expected reward, as observed following chemogenetic excitation of LHAaMCH neurons during M/D, would be interpreted as an increase in motivation to continue responding for the omitted food reinforcer. This type of perseverative, unproductive reward seeking has also been associated with activity in the NAc (Ambroggi et al., 2011; Floresco, 2015; Lafferty et al., 2020). Furthermore, the inhibitory tone produced by median spiny neurons (MSNs) in the NAc is thought to be differentially modulated via GABA- and glutamatergic inputs (Lafferty et al., 2020), which may include afferents from LHAMCH neurons. Indeed, LHAMCH neurons densely innervate the NAc, and the MCH receptor (MCH1R) is particularly strongly expressed within the NAc shell (Georgescu et al., 2005; Haemmerle et al., 2015; O’Connor et al., 2015). Thus, LHAaMCH neurons may project to the NAc to modulate motivated behavior, including the ”stop” function of PI responding. Thus, I hypothesized that LHAaMCH neurons that project to the NAc (LHAaMCHà NAc neurons) could increase motivation and delay the PI “stop” function. In this chapter, I used a dual virus approach to selectively transfect LHAaMCHà NAc 70 neurons with an excitatory DREADD. To examine the influence of these cells on motivation, rats were tested in both the progressive ratio (PR) and peak interval (PI) paradigm. While the PI paradigm specifically examines the “stop” decision within the temporal context of probe trials, the PR task more generally examines motivation by requiring progressively more instrumental responding for reinforcement. Previously, chemogenetic excitation of LHAMCH neurons delayed the “stop” function during PI responding, without influencing overall response rates in the PI task. This suggests that excitation of LHAMCH à NAc neurons may similarly influence the “stop” function in PI responding without affecting PR performance. In addition, given that LHAMCH neurons influenced motivated food-seeking only during M/D, I expect that LHAMCH à NAc neurons will also influence responding selectively during this period of the estrous cycle. The lack of effect of chemogenetic excitation of LHAMCH neurons on responding during P/E suggests that high levels of circulating ovarian hormones may block the action of these neurons. Given that the MCH system is generally inhibited by estradiol (Messina et al., 2006; Santollo & Eckel, 2008, 2013; Terrill et al., 2020), it is possible that high levels of estradiol during P/E block the effects of LHAMCH neuronal excitation. Thus, the lack of effect of chemogenetic excitation of LHAMCH neurons during P/E may represent a protective role of estradiol. In contrast, relatively low levels of estradiol circulating during M/D may create a vulnerability for LHAMCH neuronal excitation to influence behavior beyond the level observed in P/E when high levels of estradiol act as a “brake” on the MCH system. This “brake” could prevent chemogenetic excitation of 71 LHAMCH neurons from influencing the “stop” function during P/E, whereas the release of the brake in M/D enables these cells to influence behavior. Given that the striatum is highly sexually dimorphic (Becker, 1990a, 1990b; Becker & Ramirez, 1981) and contains a dense colocalization of MCH1R and the estrogen receptor-a (ER-a) (Terrill et al., 2020), estrogen may mediate MCH activity via the NAc. In line with this notion, Terrill et al. (2020) report sex differences in feeding behavior following infusion of MCH to the NAc, and that MCH peptide infusion to the NAc increased feeding in oil-treated – but not estrogen-treated – ovariectomized (OVX) females. Thus, the projections from LHAaMCH neurons to the NAc may both influence the ”stop” function and underlie the estrous cycle effects observed in the PI paradigm. Thus, in this chapter I will examine the influence of chemogenetic excitation of LHAaMCHà NAc neurons on PI responding across the estrous cycle (i.e., in both P/E and M/D) to determine whether this projection can recapitulate the motivated phenotype observed following LHAaMCH neuronal excitation in M/D females. I will also separately examine the effects of this projection more broadly on motivated behavior in a PR task. I hypothesize that LHAaMCHà NAc neuronal excitation will selectively prolong high rate responding after the criterion duration during M/D, but not when rats are tested in P/E. Furthermore I expect that this excitation will not influence overall response rates in the PR task. Materials & methods Subjects Eight adult female Sprague-Dawley rats (Envigo, Haslett, MI, USA; 12-weeks of age at arrival) were housed as described in Chapter 2. 72 Surgical procedures Stereotaxic Viral Infusion and Cannulation Rats underwent stereotaxic viral infusion and cannulation as described in Chapter 2, except that infusions of the cre-dependent excitatory DREADD was selectively targeted to the anterior LHA (Table 3.1). Table 3.1 Viral approach to selectively target LHAaMCH neurons that project to the NAc. Dual Virus Approach Virus Target Infusion coordinates AAV2(retro)-eSYN-EGFP-T2a-icre-WPRE NAc Shell +1.1 A.P., ±0.8 M.L., −7.5 D.V. 0.3 µl / infusion AND AAV2-DIO-rMCHp-hM3D(Gq)-mCherry LHAa-MCH -2.12 A.P., ±2.1 M.L., -8.4 D.V. 0.5 µl / infusion Behavioral Paradigms Rats were first trained and tested in the Peak Interval paradigm (Experiment 3a), as described in Chapter 2. After completing PI training and testing, rats were also tested in a Progressive Ratio task (Experiment 3b). Peak Interval Paradigm All training and testing procedures during Phase 1 and Phase 2 were identical to those described in Chapter 2, as behavioral assays in these chapters were performed simultaneously. Progressive Ratio Task Following completion of the PI paradigm, rats were next tested in a Progressive Ratio (PR) task. The PR task provides a classic assessment of how hard an animal is willing to work in order to earn a reward. In this task, rats must make progressively more 73 instrumental responses in order to receive reinforcement. The task ends when a rat fails to make an instrumental response within the timeout period (i.e., 15 min) or after 5 hours, whichever comes first. Notably, unlike the PI paradigm, there is no temporal component involved in the PR paradigm except that animals must respond at least once every 15 min. Autoshaping Given that rats had been extensively trained to lever press in the PI paradigm, rats were shifted to a nosepoke instrumental response in order to help distinguish this task from prior experience in the PI paradigm. Nosepoke ports, which consist of small, recessed openings containing an IR beam across the opening, were placed into the operant box on either side of the food magazine, thus taking the former position of the levers. Rats first underwent two sessions of FI20 nosepoke training to acquire the instrumental response. Then, rats were shifted to a progressive ratio (PR) schedule of reinforcement in which they needed to make progressively more nosepokes to obtain the same amount of sucrose reinforcement. The number of responses required to obtain a single reward increased following a variable schedule (i.e., PR1, 2, 4, 9, 12, 15, 20, 25, 32, 40, 50, 62, 77, 95, 118, 145, 178, 219, 268, 328, 402, 492, 603, 737, 901, 1102, 1347, 1647, 2012). PR sessions ended when rats failed to make at least one nosepoke over a 900 s (15 min) period or after five hours, whichever came first. Rats received 12 consecutive days of PR training before moving onto testing with VEH and CNO. 74 Progressive Ratio Testing PR tests sessions were identical to PR training sessions except that rats received an i.p. dose of either vehicle (0.2 M PBS) or clozapine-N-oxide (CNO; 0.3mg/ kg) 15 min prior to beginning the behavioral session. Rats were tested under each drug once in P/E and once in M/D; sessions in which drugs were administered were separated by at least a 72-hour washout period. Between tests sessions, animals performed ordinary PR sessions. Efforts were made to counterbalance the order of VEH and CNO administration between subjects, with consideration of the estrous cycle stage. In order to capture all phases of the estrous cycle, some rats received additional washout days between test sessions; all rats completed PR testing in 10-14 sessions. Histology Vaginal cytology, perfusion, tissue collection, and analysis were performed as described in Chapter 2. Statistics Data transformation Behavioral data from both the PI and PR tasks were imported into Microsoft Excel using table profiles built in MedPC2XL (Med Associates). Data from the Peak Interval task was sorted and normalized as described in Chapter 2. Statistics Data from the Peak Interval paradigm were analyzed as described in Chapter 2. Data from the Progressive Ratio task were first examined for differences in PR responding across the estrous cycle, and then as a result of chemogenetic excitation of LHAaMCHàNAc neurons separately in P/E and M/D. Using paired t-tests, I first 75 examined whether there were differences in session time or response rate between P/E and M/D rats tested under vehicle conditions. Next, I examined whether these measures differed based on drug treatment (VEH, CNO) when rats were tested in P/E or M/D. Finally, I used a Mantel-Cox, log-rank survival analysis to determine whether there were any differences in the probability of survival of (1) P/E vs M/D rats tested under VEH, (2) VEH vs CNO treatment in P/E, and (3) VEH vs CNO treatment in M/D. Results Tissue Analysis DREADD expression was confirmed by examining LHA sections from approximately - 1.20 mm to -3.90 mm posterior to bregma. In these subjects, DREADD expression extended from approximately -2.04 mm to -3.48 mm posterior to bregma. Expression was sparse, with few mCherry-labelled cells identified within each subject. Fibers and processes were slightly more apparent, suggesting successful DREADD targeting in spite of the low number of mCherry-labelled cell bodies. Expression was primarily concentrated in the more anterior and dorsolateral aspects of the LHA and entirely absent in slices beyond -3.48 mm posterior to bregma (Figure 3.1). 76 Figure 3.1 DREADD Expression in NAc-projecting LHAa neurons. (a) Representative photomicrograph of mCherry labelled DREADD expression in the LHAa. (b) Heat maps indicate representative DREADD expression (red) throughout the LHAa. Coronal sections modified from Paxinos & Watson 6th edition. Estrous cycle influences time-dependent food seeking To determine whether baseline differences in responding occurred due to fluctuating levels of circulating gonadal hormones, we first examined whether responding differed between rats tested in P/E and M/D under VEH treatment. There were no differences in the amount of time rats spent in the food cup (Figure 3.2 a; t=0.090, df=7, p>.05, d=.03) or the overall response rate during the behavioral session (Figure 3.2 b; t=1.08, df=7, p>.05, d=.38). Although rats tested in P/E may appear to respond at a peak rate earlier, there was also no significant difference in peak time between rats tested in P/E and M/D under VEH (Figure 3.2 c; t=1.93, df=7, p=.095, d=.68). The repeated measures ANOVA evaluating proportion of peak rate responding 77 across time revealed both a main effect of time (F=128.51, df=(59, 413), p<.001, ηp2=0.95) and a significant interaction of estrous cycle stage and time (F=2.12, df=(59, 413) p<.001, ηp2=0.23). Planned comparisons evaluating whether responding differed in M/D and P/E in 1s bins revealed that responding was higher under M/D at 20, 27, 32, 35 and 50s. This suggests that M/D rats continued to respond at a higher level longer than P/E rats after the omission of an expected reward. Next, I used piecewise growth modeling (PGM) to model the predicted proportion of peak rate responding across time during P/E and M/D. The effects of estrous cycle stage were examined at the mid-point of the pre- and post-peak periods in the overall model. This analysis revealed only significant effects of time on the predicted proportion of peak rate response functions (see supplemental tables S3.1 and S3.2). The overall model, which estimates rate and intercept at the mid-point pre- and post-peak, did not reveal any overall effects of estrous cycle stage. Although the overall model did not capture estrous cycle effects, modeling the data separately in 5s intervals revealed effects of estrous cycle on the predicted proportion of peak rate responding at 25s, 30s, and 35s. Specifically, after the peak, there was a significant difference in the quadratic rate of change over time (i.e., quadratic slope) based on estrous cycle stage. At each time, the quadratic slope was more negative during M/D compared to P/E. This indicates a significant difference in the rate at which the slope changes over time. The overall slope is also influenced by the linear and cubic slopes, but a difference in the quadratic rate of change can influence how steeply responding decreases post-peak. In this case, M/D animals initially decay high rate responding less quickly than P/E animals, resulting in higher level of 78 responding lasting longer than in P/E animals. However, as time passes the rate of decrease in M/D accelerates, resulting in a steeper negative slope than in P/E animals (Figure 3.2 e). Because M/D rats are predicted to initially delay decreasing post-peak responding, the level of responding predicted under M/D continued to exceed the level predicted under P/E during this period. This suggests that the predicted response rate in M/D animals may be decreasing more quickly later in the trial in order to “catch up” or normalize back down to the level of responding observed in P/E animals. Altogether, these differences in the rate of change indicate that P/E animals attenuate high rate responding more quickly after the peak than M/D animals, who delay attenuating high rate responding post-peak. These effects agree with the results of the ANOVA, which revealed differences in the magnitude of responding under P/E and M/D selectively after the criterion duration. Together, the piecewise growth model and ANOVA suggest that post-criterion responding is higher under M/D as rats attenuated high level responding post-peak more slowly. In other words, the “stop” function is delayed in M/D rats. Notably, this finding is in line with the general idea of estrogen as an anorexigenic feeding signal (Eckel, 2011). Higher levels of circulating estrogen during P/E may attenuate the motivation of an animal to continue food-seeking after the criterion duration has elapsed. 79 (a) (b) (c) 15 15 25 0.0950 P/E 20 M/D Peak Time (s) 10 10 Resp/ min % Time 15 5 10 5 5 0 0 0 P/E M/D P/E M/D P/E M/D (d) (e) 1.0 1.0 Proportion of peak rate Predicted M/D 0.5 0.5 P/E proportion of peak rate p<.05 0.0 0.0 0 20 40 60 0 20 40 60 Time (s) Time (s) Figure 3.2 Responding in the Peak Interval paradigm across the estrous cycle. (a) The amount of time rats spent in the food cup did not differ based on estrous cycle stage. (b) Rats responded at comparable rates in P/E and M/D. (c) P/E rats respond slightly earlier than M/D rats, but this effect did not reach significance. (d) M/D rats respond at a higher proportion of peak rate than P/E rats after the criterion duration. (e) Piecewise growth modeling indicates that the predicted proportion of peak rate responding is higher post-peak in M/D rats than in P/E rats. Proestrus/ estrus rats are not sensitive to LHAaMCHàNAc excitation Due to the higher levels of estrogen present during proestrus and estrus (P/E), I hypothesized that LHAaMCHàNAc neuronal excitation would fail to produce effects on PI responding when rats were tested during P/E. Indeed, there was no effect of CNO- 80 mediated excitation of LHAaMCHàNAc neurons on the amount of time rats spent in the food cup (Figure 3.3; t=.27, df=7, p>.05, d= .10) nor on their response rate across the session (t=.50, df=7, p=.63, d=.18). While rats tended to reach their peak response rate at a later time (i.e., peak time) under CNO, this effect did not reach significance (t=2.11, df=7, p=.073, d=.74). A repeated measures ANOVA was used to evaluate whether the proportion of peak rate response function differed following VEH or CNO treatment revealed an interaction of drug X time (F=1.65, df=(59, 413, p<.05, ηp2=.19), indicating that CNO- mediated excitation of LHAaMCHàNAc neurons influenced the timing of motivated responding in the PI task. However, pairwise comparisons revealed that the effect was modest given that the level of responding that occurred was greater under CNO than VEH only at 39s. 81 (a) (b) (c) 15 15 25 .073 VEH 20 CNO Peak Time (s) 10 10 Resp/ min 15 % Time 10 5 5 5 0 0 0 VEH CNO VEH CNO VEH CNO (d) (e) 1.0 1.0 Proportion of peak rate Predicted VEH CNO 0.5 0.5 proportion of peak rate p<.05 0.0 0.0 0 20 40 60 0 20 40 60 Time (s) Time (s) Figure 3.3 Peak interval responding following chemogenetic excitation of LHAaMCHàNAc neurons during P/E. (a) Chemogenetic excitation of LHAaMCH àNAc neurons did not influence the amount of time P/E rats spent in the food cup or (b) response rate during the session. (c) While there was a trend toward a later peak time following chemogenetic excitation of LHAaMCH àNAc neurons in P/E rats, this effect did not reach significance. (d) CNO-treated rats responded at a higher proportion of peak rate after the peak, an effect that was significant only at 39s. (e) Piecewise growth modeling revealed an effect of drug treatment at 15s, where predicted proportion of peak rate responding continued to increase in CNO-treated rats while leveling off in VEH-treated rats. 82 The predicted proportion of peak rate responding indicated by the overall PGM did not reveal any treatment effects at the pre- and post-peak midpoints (supplemental tables S3.5 and S3.6). However, the 5s analyses revealed a post-peak treatment effect at 15s. At this time, there was a significant interaction of drug treatment (VEH, CNO) and the quadratic rate of change over time. As seen in Figure 3.3 e, the predicted proportion of peak rate responding in CNO-treated continues to increase at this time while the predicted proportion of peak rate responding in VEH-treated rats begins to decay. However, there is no significant effect on the intercept (i.e., level of magnitude) of predicted proportion of peak rate responding, indicating that although the slopes differ, the overall level of responding is not significantly changed by CNO-treatment. Indeed, any effects of CNO treatment are brief, as treatment effects are not revealed at any other time points. Chemogenetic excitation of LHAaàNAc MCH neurons reduced motivated food-seeking post-peak during M/D Chemogenetic excitation of LHAMCHàNAc neurons did not influence the amount of time rats spent in the food cup (t=.26, df=7, p>.05, d=.09), response rate (t=.55, df=7, p>.05, d=.20), or peak time (t=1.136, df=7, p>.05, d=.40). The ANOVA evaluating differences in the proportion of peak rate response function following drug treatment revealed a significant interaction of drug x time (F=1.40, df=(59, 413), p<.05, ηp2=.17), indicating that the effect of drug treatment varied as a function of time within the trial. Pairwise comparisons revealed significant effects of drug treatment in six 1s bins, at 17, 20, 25, 28, 29, and 32s (p’s<.05). While responding was significantly higher under CNO than VEH at 17s, responding was significantly lower under CNO than VEH at 20, 25, 28, 83 29, and 32s. Notably, these latter times all occurred after the mean peak time in both VEH and CNO-treated rats (18 and 19.125s, respectively). In other words, post-peak responding was significantly reduced following CNO-mediated excitation of LHAaMCH à NAc neurons when rats were tested during M/D. Notably, the overall PGM analyzing the predicted proportion of peak rate response function also revealed a post-peak effect of treatment in M/D rats. Specifically, drug treatment interacted with the post-peak cubic rate of change over time at the post- peak mid-point (~40s). In addition, the 5s analyses also revealed post-peak treatment effects in M/D animals. There was a significant interaction of treatment x the quadratic rate of change at 30 and 35s (p’s <.05). In addition, there was a significant interaction of treatment x the cubic rate of change at 35 and 40s (p’s <.05). Initially, post-peak predicted responding decreases at a faster rate in CNO than in VEH-treated rats. However, the rate of decrease in VEH-treated rats accelerates as the trial continues, resulting in comparable responding predicted under VEH and CNO at the end of the trial. 84 (a) (b) (c) 15 15 25 .293 VEH 20 CNO Peak Time (s) 10 10 Resp/ min 15 % Time 10 5 5 5 0 0 0 VEH CNO VEH CNO VEH CNO (d) (e) 1.0 1.0 Proportion of peak rate Predicted VEH CNO 0.5 0.5 proportion of peak rate p<.05 0.0 0.0 0 20 40 60 0 20 40 60 Time (s) Time (s) Figure 3.4 Peak interval responding following chemogenetic excitation of LHAaMCHàNAc neurons during M/D. Chemogenetic excitation of LHAaMCH àNAc neurons did not influence the amount of time M/D rats spent in the food cup or (b) response rate during the session. (c) There was no effect of chemogenetic excitation of LHAaMCH àNAc neurons on peak time in M/D rats. (d) CNO-treated rats responded at a higher proportion of peak rate pre-peak at 17s, and a lower proportion of peak rate after the peak, at 20, 25, 28, 29 and 32s. (e) Piecewise growth modeling revealed an effect of drug treatment at 30, 35, and 40s. Progressive ratio responding is unaffected by LHAaMCH à NAc neuronal excitation To further evaluate the effects of LHAaMCH à NAc neuronal excitation on motivation, I also evaluated the influence of these cells on motivated responding in a 85 progressive ratio task. Unlike the PI paradigm, in which time-dependent effects of LHAaMCH à NAc neuronal excitation were revealed in the post-peak period, there were no effects of LHAaMCH à NAc neuronal excitation on PR responding. PR sessions continued for 5 hours or until a rat failed to make a response for 15 minutes, whichever came first. First, I examined whether PR responding differed based on estrous cycle under baseline, vehicle conditions. Estrous cycle did not affect how long vehicle-treated rats spent in a session before timing out (t=.06, df=4, p>.05, d=.03) or how many nosepokes they made during the session ( t=.23, df=4, p>.05, d=.11). As in the PI task, there were also no effects of chemogenetic excitation of LHAaMCHà NAc neurons on PR responding. P/E rats did not differ in the length of sessions (t=.16, df=4, p>.05, d=.07) nor in the number of responses they made (t=.32, df=4, p>.05, d=.14) during a session. There was also no effect of drug on PR responding in M/D rats: these rats also did not differ in session time (t=.37, df=4, p>.05, d=.16) or number of nosepokes (t=.40, df=4, p>.05, d= .18). There were also no effects of estrous cycle or drug treatment on the probability of survival in the task (c2=.13, df=3, p>.05). Thus, there were no effects of chemogenetic excitation of LHAaMCHàNAc neurons on progressive ratio responding. 86 (a) (b) VEH 4 1500 CNO Session Time (hrs) 3 1000 Nosepokes 2 500 1 0 0 P/E M/D P/E M/D (c) 100 CNO MD Probability of Survival CNO PE VEH MD VEH PE 50 0 0 100 200 300 Time Figure 3.5 Progressive ratio responding following CNO-mediated excitation of NAc- projecting LHAaMCH neurons. (a) There were no differences in session time (hours) following VEH or CNO treatment in M/D and P/E. (b) There was also no effect of estrous cycle stage or drug treatment on the number of nosepokes rats made during sessions. (c) The probability of survival in the task was also unaffected by estrous cycle stage or drug treatment. 87 Discussion Summary of results In this chapter, I examined whether projections from LHAaMCH neurons to the NAc could influence motivated responding in the peak interval (PI) task, specifically through alterations of the “stop” function. Given that little is known about estrous cycle effects on interval timing, I also examined baseline effects of the estrous cycle on PI responding. Indeed, even at baseline, estrous cycle stage influenced peak interval responding: rats attenuated high rate responding more abruptly after the criterion duration when tested under vehicle (VEH) during P/E compared to M/D. However, there was no significant effect of estrous cycle stage on peak time, indicating that post-criterion responding was altered without affecting time perception, per se. This decrease in post-criterion responding indicates a decrease in motivation to continue seeking an omitted food reward during P/E, when circulating levels of estrogen are highest. This effect is thus in line with the general anorexigenic effect of estradiol, which is marked by decreased food consumption driven by a decrease in meal size (Eckel, 2011). Despite this, rats reached a peak response rate at comparable times during P/E and M/D, and their overall amount of responding did not differ based on estrous cycle stage. This suggests that estrous cycle influences the timing of motivated behavior within a trial, without necessarily affecting time perception or the overall response rate. Likewise, estrous cycle stage did not influence responding in the progressive ratio (PR) task. Taken together, these findings suggest that estrous cycle may subtly modulates the timing of motivated behavior without influencing willingness to respond. 88 While estrogen has been reported to increase clock speed in OVX rats, these effects are acute and transient, indicating that any clock speed effects are quickly compensated for by an updating reference memory system (Pleil et al., 2011; Sandstrom, 2007). Indeed, prolonged estrogen administration in OVX rats failed to produce a change in time perception (Ross & Santi, 2000), although it did reduce discrimination accuracy in OVX females. In contrast to the effects of estrogen on time perception, effects of EB on motivation may be able to more robustly and persistently influence the timing of responding. Thus, estrous cycle effects captured across multiple tests – as in the case of the present study – may more readily reveal motivational rather than timing effects. That the “stop” function is influenced by circulating gonadal hormones in the present study may thus reflect an estrogen-mediated influence on motivation to continue responding after the omission of an expected sucrose reward. In line with this, Gur et al. (2019) suggest that sex differences in interval timing may arise from differences in incentive motivation between males and females rather than time perception, per se (Gür et al., 2019). Although this group observed no differences in time perception in mice tested in the PI paradigm, they report that female mice began high rate responding later than male mice, indicating a delayed “start” function. Thus, the “start” function was selectively modified by sex, with females displaying lower motivation to initiate high rate responding than males (Gür et al., 2019). Similarly, Pleil et al. (2011) also report that male rats respond earlier than females under baseline conditions. However, in this case the effects of sex are interpreted as true timing effects, with the delay in females attributed to a prolonged reference memory for time (Pleil et al., 2011). This is consistent with reports from the human literature, in 89 which women tend to overestimate durations (Block et al., 2000; Morita et al., 2005; Williams, 2011). Altogether these findings indicate that ovarian hormones may influence the timing of motivated behavior through multiple mechanisms (Pleil et al., 2011). Interestingly, when LHAaMCHàNAc neurons were excited during P/E, the opposite effect was observed: rats tended to respond later under CNO, resulting in a subtle increase in responding post-criterion. Although this effect was limited, it perhaps suggests an ability of MCH neurons to increase motivated food intake when motivation for food is otherwise low (i.e., during P/E). However, given that estradiol inhibits the actions of MCH, any effect of MCH during P/E is likely blunted by high levels of this ovarian hormone. On the other hand, when LHAaMCHàNAc neurons were excited during M/D, post- criterion responding was consistently decreased, an effect that is similar to our previous findings which revealed a role for LHAMCH neurons in determining post-criterion responding during M/D. Recall that non-projection-specific excitation of LHAaMCH neurons prolonged high rate responding after the omission of an expected sucrose reward, suggesting an increase in motivation and delay in the “stop” function. I thus hypothesized that excitation of LHAaMCHàNAc neurons during M/D would replicate this finding and prolong high rate responding after the omission of sucrose. However, in contrast to my hypothesis, projection-specific chemogenetic excitation of LHAaMCHàNAc neurons decreased post-criterion responding by accelerating the “stop” function. This indicates a decrease in motivation to continue working for an omitted food reward, and suggests an inhibitory action of these neurons on food-seeking. 90 Although I initially hypothesized that this projection-specific manipulation would increase motivation and result in increased responding post-peak, the observed decrease in responding is not altogether surprising. For example, the NAc is also important for behavioral inhibition, which includes the inhibition or attenuation of ongoing motivated behaviors (Ambroggi et al., 2011; Lafferty et al., 2020; Zamorano et al., 2014). Although the PI paradigm is not typically described in terms of behavioral inhibition, it inherently involves the inhibition of high rate responding based on the temporal context of the task. Animals typically inhibit high rate responding prior to the period of time when reinforcement is most likely, respond at a high rate around the criterion time when reinforcement might occur, and then inhibit responding again after they perceive that too much time has elapsed for reinforcement to occur. Thus, the “break-run-break” pattern of responding observed during probe trials could instead be described as an “inhibit- allow- inhibit” pattern of behavioral control informed by the temporal constraints of the task. From this perspective, the “start” and “stop” function of interval timing would reflect a release from inhibition (the “start”) and return to inhibition (the “stop”). In the present study, LHAaMCHàNAc neuronal excitation resulted in an altered “stop” function by rapidly decreasing post-peak responding, perhaps by increasing inhibition of post-peak responding. These results suggest that NAc-projecting LHAaMCH neurons may guide behavioral state transitions, an effect which is in line with the role of the LHA as an integrative relay station and the NAc as an important region involved in action selection (Berthoud & Münzberg, 2011; Floresco, 2015; Stuber & Wise, 2016). LHAaMCH neurons may contact the NAc to coordinate food-related behaviors. Indeed, the “stop” function 91 may represent an inhibition of high rate responding mediated by the action of GABAergic median spiny neurons (MSNs) in the NAc (Ambroggi et al., 2011; Floresco, 2015; Lafferty et al., 2020). A delay in the “stop” function could thus be mediated by inputs onto these MSNs, which could originate from MCH neurons in the LHAa neurons. Given that more general excitation of LHAaMCH neurons delayed the “stop” function consistent with a decrease in behavioral inhibition, it is possible that additional LHAaMCH neurons not captured in this projection-specific approach instead promote prolonged responding. This could be accomplished either by indirectly modulating the inhibitory control of the NAc or through actions in another target region. Regardless, these data indicate that LHAaMCH neurons modulate the “stop” function to prolong or attenuate motivated behavior. While this phenotype was expressed in a time-dependent food- seeking paradigm, the general effect of altering the duration of feeding related behaviors – like burst or meal size – is in line with the typical mechanism through which MCH promotes feeding (Baird et al., 2006; Messina et al., 2006; Sherwood et al., 2015). Consistent with our previous findings, excitation of NAc-projecting LHAaMCH neurons also influenced motivated responding predominantly during M/D. While the direction of this effect was opposite to that observed following non-projection-specific excitation of LHAaMCH neurons, the estrous cycle phase is consistent: neuronal excitation of LHAaMCH neurons produces motivational effects only when rats are tested during M/D. As described previously, this likely reflects a vulnerability produced by comparably low levels of estrogen than what are present during P/E, when the effects of these neurons appear to be attenuated. The influence of estradiol on LHAaMCHàNAc neuronal excitation during PI responding is examined in Chapter 4. 92 Limitations As in chapter 2, mCherry labelling of DREADD expression in the LHAa was limited to small number of neurons and primarily visible through evidence of mCherry- labelled fibers of passage. Again, it was difficult to perform detailed histological analyses on tissue in which the mCherry fluorophore was difficult to visualize, even after amplification. However, unlike the previous study, behavioral effects were clearly apparent following CNO treatment in these animals. This suggests that the CNO- mediated excitation of this limited number of LHAa neurons was sufficient to influence the ”stop” function. In addition, the effect of this neuronal excitation was consistent with previous findings indicating that LHAaMCH neurons influence that the “stop” function only during M/D. Conclusion CNO-mediated excitation of LHAaMCHàNAc neurons accelerates the “stop” function during M/D. This finding is consistent with previous work in our lab, which indicated that non-projection specific excitation of LHAaMCH neurons also influence the “stop” function during M/D. However, excitation of LHAaMCH neurons prolonged high rate responding during M/D by delaying the “stop” function, excitation of only the NAc- projecting neurons within this population instead accelerates the “stop” function. This indicates that LHAaMCH neurons may bidirectionally influence motivational processes influencing the decision when to “stop” motivated behaviors. While a small subset of these neurons that project to the NAc decrease motivated responding after the omission of an expected reward, LHAaMCH that do not project directly to the NAc may instead increase motivated responding. Altogether, this data supports a role for LHAaMCH 93 neurons in gating motivated behavior, particularly after the omission of an expected reward after a criterion duration. Thus, these neurons may be incorporating temporal cues to guide motivated behavior. 94 CHAPTER 4: Estrogen is necessary for LHAaMCHàNAc neuronal effects on post- criterion responding Abstract Previously, I demonstrated that Nucleus Accumbens (NAc) projecting Melanin Concentrating Hormone (MCH) neurons in the anterior Lateral Hypothalamic Area (LHAa) decrease time-dependent motivated responding in the PI task during periods of the estrous cycle when estrogen levels are typically lower. Specifically, chemogenetic excitation of LHAaMCH à NAc neurons attenuated responding during the “stop” function in rats tested during metestrus/ diestrus (M/D). In contrast, when rats were tested when ovarian hormone levels peak during proestrus and estrus (P/E), chemogenetic excitation of LHAaMCH à NAc neurons did not influence the “stop” function. This suggests that circulating ovarian hormones interfere with the excitation of LHAaMCH à NAc neurons to blunt their effects. Although multiple hormones fluctuate across the estrous cycle, estrogen is known to influence both interval timing and MCH-mediated feeding behaviors (Eckel, 2011; Panfil et al., 2023; Ross & Santi, 2000; Sandstrom, 2007; Santollo & Eckel, 2008, 2013). Estradiol attenuates the orexigenic effects of MCH, which suggests that higher concentrations of circulating estrogen during P/E might interfere with the effects of LHAaMCH à NAc neuronal excitation. In the present study, I ovariectomized adult female rats to remove the primary source of endogenous estrogen. I then tested OVX rats in the PI paradigm with and without estrogen (17-B- estradiol benzoate; EB) replacement before chemogenetically exciting LHAaMCH à NAc neurons. In OVX rats, estradiol replacement delayed the “start” of high rate responding in the PI task, without influencing the “stop” function or peak time. This suggests that EB 95 alone reduced motivation to respond for a sucrose reward and delayed when rats initiated high rate responding. Interestingly, in OVX rats that were pretreated with oil, chemogenetic excitation of LHAaMCH à NAc neurons did not influence PI responding. Instead, in contrast to my hypothesis, chemogenetic excitation of LHAaMCH à NAc neurons decreased motivated food-seeking only in EB pretreated rats. This suggests that EB is necessary for LHAaMCH à NAc neurons to influence post-criterion responding, but that the timing or source of estradiol (i.e., high levels of endogenous estrogens in P/E or following administration of exogenous EB) can influence how these systems interact. Additionally, EB may act in concert with other ovarian hormones to modulate the influence of LHAMCH neurons in intact, cycling rats. 96 Introduction LHAaMCH neurons alter the “stop” function when rats are tested in the peak interval (PI) task during metestrus/ diestrus (M/D). In the previous chapter, I revealed that projection-specific excitation of LHAaMCH à NAc neurons accelerated the “stop” function selectively when rats were tested during M/D. While circulating levels of ovarian hormones – including luteinizing hormone (LH), follicle stimulating hormone (FSH), progesterone (P) and estrogen (EB) – are typically highest during proestrus and early estrus (P/E), M/D is characterized by relatively low levels of these fluctuating hormones (Goldman et al., 2007). Estrogen, in particular, is of interest because of its inhibitory influence over MCH (Messina et al., 2006; Mystkowski et al., 2000; Santollo & Eckel, 2008, 2013). Thus, the absence of high levels of estrogen during M/D may create a vulnerability for LHAaMCH à NAc neurons to influence motivated behavior beyond the level typically observed during P/E. Thus, in order to isolate the effects of estrogen, in this chapter I ovariectomized (OVX) adult female rats prior to training and testing in the PI paradigm. Rats were then tested with and without estrogen replacement to examine both baseline effects of estrogen on PI responding as well as its influence on the chemogenetic excitation of LHAaMCH à NAc neurons. Although sex differences in interval timing procedures are apparent (M. Buhusi et al., 2017; Gür et al., 2019; Williams, 2011), few studies have directly examined the influence of ovarian hormones (Morita et al., 2005; Morofushi et al., 2001; Panfil et al., 2023), or estradiol on time perception (Pleil et al., 2011; Ross & Santi, 2000; Sandstrom, 2007). In interval timing tasks, acute estrogen replacement via 17-b- estradiol (EB) in OVX rats shifts the proportion of peak rate response function to the left, 97 consistent with an increase in clock speed (Pleil et al., 2011; Sandstrom, 2007). This effect occurs rapidly and acutely. As such, it suggests an abrupt increase in the accumulation rate of striatal DA which in turn increases clock speed (Meck, 1996; Sandstrom, 2007). Indeed, like DA agonists, EB treatment appears to produce transient effects on clock speed, prior to the formation of an updated internal reference memory for time under this altered clock speed (Pleil et al., 2011). Similarly, when EB is administered only once prior to PI testing, effects dissipate by 72 hours later (Sandstrom, 2007). These rapid effects of EB suggest that they are mediated by non- genomic mechanisms of EB in the striatum (Becker, 1990b; Grove-Strawser et al., 2010; Micevych & Mermelstein, 2008). In contrast to the rapid effects of estrogen on interval timing, effects of EB on food intake can occur over multiple timeframes, including in both a tonic and phasic manner (Eckel, 2004, 2011; Varma et al., 1999). The phasic effects of estrogen can be observed in cycling rats who display a decrease in food intake during estrus, after the periovulatory release of estrogen has peaked and fallen following proestrus (Eckel, 2004, 2011). Thus, this phasic decrease in food intake actually occurs during a period of the estrous cycle when circulating estrogens are lower (Eckel, 2004, 2011). Both phasic and tonic effects of EB can be observed in OVX rats treated with exogenous EB as part of a hormone replacement regimen (Asarian & Geary, 2002; Geary & Asarian, 1999). However, the anorexigenic effects of estradiol on food intake primarily occur after a delay of 36 – 40 hours, indicating a phasic effect (Eckel, 2011). Estrogen exerts its influence on food intake indirectly, primarily by influencing signals that control meal size (Butera, 2010; Eckel, 2004, 2011). Thus, as an orexigenic 98 neuropeptide that promotes consumption by increasing meal size, MCH is a candidate peptide to mediate the effects estrogen on food intake. Indeed, others have examined interactions between MCH and estrogen in both intact and OVX rats (Murray et al., 2000; Santollo & Eckel, 2008, 2013) and reported that circulating estrogen modulates the MCH system. While mRNA of the pMCH promoter is detected in comparable levels during proestrus and diestrus in intact rats (Murray et al., 2000), both MCH and MCH1R protein are decreased during proestrus (Santollo & Eckel, 2013). In OVX rats, exogenous EB reduces both pMCH mRNA as well as the MCH and MCH1R protein (Murray et al., 2000; Santollo & Eckel, 2013), indicating that estrogen typically inhibits the MCH system. Notably, the behavioral effects of EB on MCH occur through phasic effects: when EB is replaced on a four day, cyclic regimen of two days of EB injection followed by two days of washout, the orexigenic effects of MCH are attenuated on the fourth day of this cycle (Santollo & Eckel, 2008). Likewise, the orexigenic effects of intra-NAc infusion of MCH are attenuated on the fourth day of EB replacement in this cycle (Terrill et al., 2020). Although this does not exclude rapid, nongenomic effects of EB on MCH, these findings indicate that at least some of the effects of EB on MCH occur in a phasic manner, likely due to genomic effects. In keeping with cyclic hormone regimens that replace estrogen rhythmically across four to five days, I administered EB 30 minutes prior to the behavioral session on two consecutive days, followed by a two day washout period. In order to increase the probability that EB replacement would influence MCH signaling in the PI task, I chose to time the chemogenetic excitation of LHAaMCHàNAc neurons with the second day of EB administration. This would enable the test to capture both slower-acting genomic effects 99 from the previous day of EB priming, as well as any acute, nongenomic effects of EB administration. Furthermore, to enable within subjects testing but ensure that no phasic or lingering effects of estrogen were present during oil tests, all rats received oil pretreatment tests prior to EB testing. Given that EB has been reported to increase clock speed and shift the proportion of peak rate response function to the left, I hypothesized that EB treatment would produce effects on PI responding under baseline, vehicle conditions, perhaps by initially shifting the response function leftward before renormalizing following repeated testing. In addition, because the effects of LHAaMCH àNAc neuronal excitation were pronounced during M/D but not P/E, I hypothesized that LHAaMCH àNAc neuronal excitation would influence post-criterion food-seeking in oil-treated OVX – but not EB-treated OVX – rats. In line with the inhibitory effects of EB on MCH, I expected that EB pretreatment would attenuate or block the effects of LHAaMCH à NAc neuronal excitation. Materials & methods Subjects Eight adult female Sprague-Dawley rats (Envigo, Haslett, MI, USA; 12-weeks of age at arrival) were pair housed in groups of 2-3 in standard, plexiglass cages with metal tops. Rats were maintained on a standard 12-hr light-dark cycle (lights on 7:00; lights off 19:00), with ad libitum access to Teklad diet #8912. Rats received ³7 days of acclimatization to the vivarium before experimental manipulations began. Following the period of habituation, rats were handled daily for 2-3 days before undergoing surgical procedures. Post-op, rats were briefly singly housed while they received daily health monitoring. Rats were pair housed with their original cage mate once postoperative 100 bodyweight recovered and surgical incisions appeared healed (≤7 days). Rats continued to be pair-housed throughout all behavioral experiments. All manipulations were conducted in compliance with the Institutional Animal Care and Use Committee, Michigan State University. Surgical procedures Stereotaxic Viral Infusion and Cannulation Under isoflurane anesthesia, subjects received bilateral infusions of the retrograde AAV2(retro)-eSYN-EGFP-T2a-icre-WPRE and a cre-dependent, excitatory DREADD AAV2-DIO-rMCHp-hM3D(Gq)-mCherry to the NAc and to the LHAa, respectively as described in Table 4.1. In contrast to surgical procedures described in Chapter 3, rats received an additional infusion of 0.25 µl of the cre-dependent, excitatory DREADD virus into the LHAa to increase the probability of successful DREADD expression. Table 4.1 Viral approach to selectively target LHAaMCH neurons that project to the NAc in OVX rats. Dual Virus Approach Virus Target Infusion coordinates AAV2(retro)-eSYN-EGFP-T2a-icre-WPRE NAc Shell +1.1 A.P., ±0.8 M.L., −7.5 D.V. 0.3 µl / infusion AND AAV2-DIO-rMCHp-hM3D(Gq)-mCherry #1: -2.12 A.P., ±2.1 M.L., -8.4 D.V. 0.5 µl / infusion #1 LHAa-MCH And 0.25 µl / infusion #2 #2: -2.40 A.P., ±2.1 M.L., -8.4 D.V. Ovariectomy & Hormone Replacement Immediately following completion of the viral infusion surgery, rats were moved from the stereotaxic set-up to a standard nosecone to maintain anesthesia under 101 isoflurane gas. Flanks were shaved and sterilized. Bilateral incisions were made, and the fat pad and ovary were identified and moved out of the body cavity. Fallopian tubes were clamped, and the ovary was removed via cauterization. Muscle incisions were closed with interrupted absorbable sutures; skin was closed with surgical staples and covered with triple antibiotic cream. Rats were treated with 2 mg/ kg meloxicam to reduce post-operative pain. To confirm complete ovariectomy, estrous cycle tracking was performed as previously described, beginning during PI sessions. Unfortunately, one rat continued to display evidence of cyclicity and was excluded. OVX rats were trained in the absence of hormone replacement. During behavioral testing, rats received hormone (17-b-estradiol benzoate, EB; 5 µg/ 0.1ml sesame oil) or control (sesame oil, 0.1 ml) treatment via subcutaneous (s.c.) injections 30 minutes prior to the behavioral session. Oil or EB was administered in four-day cycles: two consecutive days of oil or EB pretreatment, followed by a two-day washout period. Tests sessions, in which animals also received either vehicle (0.2M PBS) or clozapine-N-oxide (CNO; 0.3 mg/ kg) prior to the behavioral session, occurred during the second day of hormone/ oil treatment during each four-day cycle. Oil tests (2x VEH, 2x CNO) were performed first. Behavioral Paradigm Following recovery from viral infusion and food restriction to 90% baseline weight, subjects received training in the Peak Interval (PI) paradigm, as described in Chapter 2. All training and testing procedures during Phase 1 and Phase 2 were identical to those described in Chapter 2. Training occurred in the absence of hormone replacement. Beginning on the sixteenth session of PI training, rats began receiving the 102 first of four oil treatment cycles, in which rats were primed with oil (0.1ml sesame oil) 30 min prior to the behavioral session for two consecutive days, then allowed two days of washout. The same cyclic regimen was followed for 17-b-estradiol (EB; 5 µg EB/ 0.1ml sesame oil) replacement. Chemogenetic excitation of LHAaMCH neurons occurred during the second day of oil or hormone replacement (i.e., 24 hours after the first dose and 30 minutes after second does). Although the order of drug treatment (VEH or CNO) was counterbalanced, all rats received oil tests prior to EB tests to avoid unintended effects of EB on subsequent oil tests. Histology Vaginal cytology was performed in ovariectomized rats to confirm cessation of the estrous cycle. Cytology was performed as described in Chapter 2. Although rats were not cycling, sample collection continued throughout behavior to ensure that rats were handled in a consistent manner regardless of whether they were OVX or intact. Perfusion, tissue collection, and analysis were performed as described in Chapter 2. Results Histology Complete ovariectomy was confirmed by performing vaginal lavages in OVX rats and verifying a lack of round or cornified epithelial cells present within the sample over a period of at least four consecutive days. While the lack of estrous cycles was quickly confirmed, daily lavages continued throughout testing to ensure that rats were handled in a manner consistent with previous cohorts of intact, cycling animals. Interestingly, cyclic replacement of EB produced changes in vaginal epithelial cells even after rats had lacked changes in these cells for a period of several weeks. While EB replacement 103 alone is not sufficient to fully recapitulate the profile of fluctuating ovarian hormones in intact animals, this treatment has been reported to alter vaginal epithelial cells in OVX rats (Montes & Luque, 1988) and in the present study led to a qualitative increase in the proportion of round epithelial cells, particularly after multiple cycles of EB replacement. Bilateral DREADD expression was confirmed in each of n=7 adult OVX females. In this group, the addition of a second LHA infusion of the cre-dependent pMCH driven DREADD resulted in robust expression from approximately -1.92 to -3.72 mm posterior to Bregma. Figure 4.1 DREADD Expression in NAc-projecting LHAa neurons of OVX animals. (a) Representative photomicrograph of mCherry labelled DREADD expression in the LHAa (red) and cells expressing the MCH protein (green); colocalization is indicated in yellow/ orange. (b) Heat maps indicate representative DREADD expression (red) throughout the LHAa. Coronal sections modified from Paxinos & Watson 6th edition. 104 Peak Interval responding is influenced by estradiol To determine whether estradiol influenced PI responding in ovariectomized (OVX) rats, I first examined whether responding differed when OVX rats were tested under control vehicle (0.2M PBS) conditions following oil (0.1 ml sesame oil) or estradiol (5 µg/ 0.1 ml oil) pretreatment. There were no significant effects of hormone replacement on the amount of time rats spent in the food cup (t=2.2, df=6, p=.07, d=.84), response rate (t=.60, df=6, p>.05, d=.23) or peak time (t=.81, df=6, p>.05, d=0.31). Despite not affecting these overall measures of responding or the accuracy of time perception, hormone replacement in adult OVX rats delayed the “start” function by significantly reducing the proportion of peak rate responding that occurred pre-peak following EB pretreatment. In addition to a main effect of time (F=156, df=(59, 354), p<.001, ηp2=.96), there was also a trend toward a main effect of estrous in these rats (F=5.45, df=(1, 6), p=.058, ηp2=.48). In addition, there was a significant interaction of estrous cycle stage and time (F=2.74, df=(59, 354), p<.001, ηp2=.31), indicating that time and estrous cycle stage interact to guide responding. Pairwise comparisons using a Bonferroni correction were computed at each of sixty 1s time bins to identify when responding significantly differed between oil and EB-treated rats. Responding significantly differed at: 4-12, 15, 18, 22 and 59s (p’s <.05). Notably, responding was significantly lower following EB pretreatment before the criterion (i.e., at 4-12, 15, and 18s) but significantly higher following EB pretreatment after the criterion at 22s. 105 (a) (b) (c) 20 20 25 Oil 20 EB 15 15 Resp/ min Peak Time (s) 15 % Time 10 10 10 5 5 5 0 0 0 (d) (e) 1.0 Proportion of peak rate OIL EB p<.05 0.5 0.0 0 20 40 60 0 20 40 60 Time (s) Time (s) Figure 4.2 Effects of hormone on PI responding. There were no effects of hormone (oil control or estradiol, EB, replacement) in adult OVX rats on overall measures of responding during the behavioral session, including (a) amount of time spent in the food cup or (b) response rate. (c) EB replacement in adult OVX rats also did not significantly change peak time. (d) The “start” of the proportion of peak rate response function was shifted rightward in EB-treated rats relative to oil-treated controls. (e) Piecewise growth models of the predicted proportion of peak rate response function following oil or EB pretreatment confirm that the “start” function is delayed in EB treated rats. In contrast to this pattern, responding was higher following oil pretreatment at 59s. However, because response rates were so low and the proportion of peak rate response was near zero at 59s, this effect may be an artifact. These results suggest 106 that EB-treated rats delay the “start” function relative to oil-treated rats. However, this delay in start function does not significantly affect peak time (no significant change). In addition, the effects of EB on the “start” function do not coincide with an effect of EB on the “stop” function. While the overall multilevel PGM centered at the pre- and post-peak mid-points did not reveal any effects of hormone pretreatment (supplemental tables S4.1 and S4.2), the 5s analysis revealed pre-peak effects of hormone at 5 and 10s (p’s<.05; supplemental tables S4.3 and S4.4). There were significant interaction effects of hormone with the quadratic and cubic rate of change at both 5 and 10s (p’s<.05; see supplemental tale S4.3 and S4.4). While the difference in magnitude was not significant, oil pretreated rats initially responded at a higher level and increase their response rate more abruptly than EB-treated rats. The predicted proportion of peak rate increased in both groups pre-peak as the criterion neared, but the oil-treated rats accelerated their response rate more quickly than EB-treated rats at 5 and 10s, indicating that EB-treated rats delay the “start” of high rate responding relative to oil controls. As the criterion nears and both groups approach peak rate, these differences in rate of change dissipate (i.e., there are no significant differences in slope after 10s). Notably, all EB pretreatment exclusively influences the “start” function, leaving peak time and the “stop” function unaffected. LHAaMCHà NAc excitation in oil pretreated rats does not influence the “stop” function As in previous studies, chemogenetic excitation of LHAaMCHà NAc neurons did not influence overall measures of responding during the PI task in oil pretreated rats. There was no effect of LHAaMCHà NAc neuronal excitation on the amount of time rats 107 spent in the food cup (t=.67, df=6, p>.05, d=.25) or their overall response rate during the session (t=.57, df=6, p>.05, d=.22). As was the case in intact animals, there was also no effect of LHAaMCHà NAc neuronal excitation on peak time in oil pretreated rats (F=1.27, df=6, p>.05, d=.48). Contrary to my hypothesis, there was also no effect of LHAaMCHà NAc neuronal excitation on the proportion of peak rate response function in oil pretreated rats. The ANOVA revealed a main effect of time (F=161.35, df=(59, 354), p<.001, ηp2=.31) but no effect of drug (F=.05, df=(1, 6), p>.05, ηp2=.01) or interaction effect of drug x time (F=.70, df=(59, 354), p>.05, ηp2=.10). The overall PGM also failed to reveal any effect of chemogenetic excitation of LHAaMCHà NAc neurons in oil pretreated rats at the pre- or post-peak mid-points (see supplemental tables S4.5 and S4.6). This is not surprising given that the predicted proportion of peak rate response functions under VEH and CNO are nearly superimposed in oil pretreated rats. However, the refined PGM analysis that examined responding in 5s intervals did capture an effect of CNO on responding at 15s. At this time, there was a significant interaction of drug treatment (VEH, CNO) and the pre-peak quadratic rate of change in oil-treated rats. Consistent with previous transient effects observed around 15s in Chapters 2 and 3, this effect also dissipated by the 20s interval. Thus, CNO briefly accelerated the rate at which responding decreased at 15s in oil- treated rats. Notably, this subtle effect is in contrast to the robust post-peak reduction observed in intact, M/D rats. 108 (a) (b) (c) 20 20 25 VEH 20 CNO 15 15 Resp/ min Peak Time (s) 15 % Time 10 10 10 5 5 5 0 0 0 (d) (e) 1.0 Proportion of peak rate CNO VEH 0.5 p<.05 0.0 0 20 40 60 0 20 40 60 Time (s) Time (s) Figure 4.3 Effects of LHAaMCH à NAc excitation on PI responding in oil-treated rats. (a) There were no differences in the amount of time rats spent in the food magazine during PI sessions following treatment with VEH or CNO. (b) There was also no influence on the overall response rate during the interval timing task. (c) While peak time was reduced following CNO mediated excitation of LHAaMCHà NAc neurons in oil pretreated rats, this effect did not reach significance. 109 Estradiol pretreatment enables post-peak effects of LHAaMCHà NAc neuronal excitation on the “stop” function As in previous studies, there were no effects of chemogenetic excitation of LHAaMCHà NAc neurons on overall measures of behavior from the session, including the time spent in the food magazine (t=.70, df=6, p>.05, d=.26) or response rate (t=1.07, df=6, p>.05, d=.4). In contrast to previous studies, however, chemogenetic excitation of LHAaMCHà NAc neurons significantly reduced peak time in EB pretreated rats (t=2.45, df=6, p<.05, d=.93). This reduction in peak time indicates that CNO-mediated excitation of LHAaMCHà NAc neurons in EB-treated rats results in earlier responding, an effect that was not evident in oil-treated rats. In addition to the effects on peak time, chemogenetic excitation of LHAaMCHà NAc neurons also influenced the proportion of peak rate response function. The ANOVA evaluating the proportion of peak rate responding (Figure 4.4c) revealed a main effect of time (F=101.22, df=(59, 354), p<.001, ηp2=.94). While there was no main effect of drug (F=2.26, df=(1,6), p>.05, ηp2=.27), there was a significant interaction effect of drug x time (F=2.56, df=59, 354, p>.05, ηp2=.30). Pairwise comparisons evaluating where responding significantly differed under VEH and CNO in 1s bins revealed significant differences at 25, 26, 30, 31 and 35s – all times that occur after the criterion duration, and typically after peak time. Importantly, at each of these time points, responding was greater under VEH than CNO, indicating that CNO-treated rats respond at a lower level post-peak than their VEH-treated counterparts. 110 (a) (b) (c) ✱ 20 20 25 VEH 20 CNO 15 15 Resp/ min Peak Time (s) 15 % Time 10 10 10 5 5 5 0 0 0 (d) (e) 1.0 Proportion of peak rate CNO VEH 0.5 p<.05 0.0 0 20 40 60 0 20 40 60 Time (s) Time (s) Figure 4.4 Effects of LHAaMCH à NAc excitation on PI responding in EB-treated rats. There were no effects on overall measures of responding during the behavioral session, including (a) amount of time in the food cup or (b) response rate. (c) Chemogenetic excitation of LHAaMCH à NAc neurons reduced peak time in rats pretreated with EB. (d) Proportion of peak rate responding was reduced in EB pretreated rats following CNO- mediated excitation of LHAaMCH à NAc neurons. (e) Modeling the predicted proportion of peak rate response functions did not reveal any differences in the predicted responding under VEH and CNO in EB pretreated rats. Interestingly, neither the overall PGM nor the refined analyses at 5s intervals revealed effects of CNO in EB pretreated rats. This suggests that while there are differences in the actual magnitude of responding – as revealed by the ANOVA – the 111 predicted magnitude and the rate of change is similar regardless of CNO treatment. These differences may be attributable to methodological approaches: while the ANOVA collapses data from two VEH tests and two CNO tests together, the PGM is capable of treating each test independently due to its multivariate structure. Discussion Summary of results In this chapter, I examined the influence of estrogen on NAc-projecting MCH neurons from the anterior LHA. In the previous chapter, I demonstrated that excitation of these LHAaMCHà NAc neurons decreased post-peak responding in the peak interval (PI) task in a manner that suggests a change in the “stop” function. In intact, cycling rats excitation of these neurons produced behavioral effects only when rats were tested during metestrus/ diestrus, when circulating levels of estrogen are typically lower. In the present study, I directly modulated the level of plasma estradiol (EB) by removing the primary source of endogenous estrogen through ovariectomy (OVX). Adult female rats were OVX prior to undergoing behavioral training in the PI paradigm and testing with and without estrogen replacement. Chemogenetic excitation of LHAaMCHàNAc neurons first occurred in rats pretreated with oil; rats were tested under both oil + vehicle (VEH) and oil + clozapine-N-oxide (CNO) conditions. Next, rats were pretreated with EB before repeating PI testing with VEH and CNO. To separately examine the influence of estrogen on this task, I first examined the influence of EB on PI responding under baseline (i.e., VEH) conditions. Following EB pretreatment, rats delayed high rate responding pre-peak compared to when they received oil pretreatment. This indicated a delay in the “start” function; EB delays the 112 start of high rate motivated responding. This effect occurred without significantly altering peak time or producing any effects on the stop function. That EB shifts the response function to the right is consistent with reports suggesting that females respond later than males in interval timing tasks (Gür et al., 2019; Pleil et al., 2011). The delayed “start” suggests that EB reduces motivation to begin high rate responding (i.e., begin exerting high levels of effort) for potential sucrose reinforcement. In addition, given that the peak time and the “stop” function were unaffected, it appears that this motivational effect of EB occurs without influencing the perception of time, per se. This delay in the “start” function is consistent with reports of delays in the “start” of motivated responding following reinforcer devaluation (Galtress & Kirkpatrick, 2010; Roberts, 1981). Thus, EB pretreatment decreased motivation to engage in effortful lever pressing until the time when reinforcement was more likely (i.e., around the criterion duration). Interestingly, this effect is in contrast to reports from the literature suggesting that EB increases clock speed, perhaps by increasing striatal dopamine (DA) release (Sandstrom, 2007). However, our paradigm differs from Sandstrom et al. in that rats were administered EB on two consecutive days and received two separate VEH tests intermixed with CNO testing. Thus, rats in our study had received PI training under EB at least once prior to their first VEH test, and across at least three cycles prior to their second VEH test. Thus, in our case clock speed effects of EB may be masked by new learning that has occurred during the previous days of PI training under EB. In other words, the reference memory for time may have updated to reflect the new clock speed induced by EB pretreatment. However, while clock speed effects are transient and dependent on an outdated reference memory for time, the motivational effects of EB 113 occur independently of these clock and memory effects. Thus, a persistent delay in high rate responding following EB pretreatment suggests that EB decreases motivation to respond for potential sucrose reinforcement in the PI task, regardless of the capacity of rats to accurately time the criterion. In contrast to my hypothesis, chemogenetic excitation of LHAaMCHà NAc neurons in OVX rats failed to influence the “stop” function in the absence of EB pretreatment. While CNO treatment briefly influenced the quadratic rate of change at 15s by accelerating the rate of response attenuation, there was no other influence of this treatment on responding in the PI task. Given that the accelerated “stop” function was observed following LHAaMCHà NAc neuronal excitation during M/D, but not P/E, I expected that removal of circulating gonadal hormones would facilitate these effects. Instead, chemogenetic excitation of LHAaMCHà NAc neurons following oil pretreatment failed to influence post-peak responding. This suggests that the removal of ovarian hormones in adulthood reduced the ability of these neurons to influence motivated responding via alterations in the “stop” decision. In other words, removal of ovarian hormones reduced the susceptibility of female rats to the action of these neurons. Thus, without fluctuating ovarian hormones, adult female rats may be protected against the influence of LHAaMCHà NAc neurons on time-dependent motivated responding. Although this finding is surprising, it is perhaps in line with evidence from my initial studies including both males and females. In these studies, females – but not males – were vulnerable to the influence of LHAaMCH neurons on motivated responding post-peak. Thus, males are somehow protected against the effects of LHAaMCH neuronal excitation. Although OVX females are not male-like, per se, the absence of 114 fluctuating ovarian hormones may be similarly protective against the influence of LHAaMCHà NAc neurons on PI responding. In line with the idea that estrogen may create a vulnerability to the effects of LHAaMCHà NAc neurons on time-dependent motivated responding, LHAaMCHà NAc neuronal excitation unexpectedly accelerated the “stop” function in EB pretreated rats. EB-pretreated rats attenuate high rate responding more quickly under CNO than VEH. They also reached a peak significantly earlier than VEH-treated rats (i.e., peak time is reduced), an effect which may be driven by an abrupt “stop” function rather than change in time perception, per se. While these effects are in contrast to my hypothesis that EB pretreatment would attenuate the effects of LHAaMCHà NAc neuronal excitation, the post-peak effect closely resembles the phenotype observed in M/D females that I thought would occur in oil pretreated rats. Thus, the presence – rather than absences – of EB may be necessary to enable LHAaMCHà NAc neurons to influence motivated responding. However, that this effect occurs following acute administration of EB 30 minutes prior to testing is unexpected, given that EB generally inhibits the actions of MCH. Importantly, however, the reported inhibitory effects of EB on MCH occur through phasic rather than tonic mechanisms. When EB is replaced in a four day, cyclic regimen, the effects of EB on MCH occur on the fourth day – i.e.,, 36-72 hours after administration, during the washout period (Messina et al., 2006; Santollo & Eckel, 2013; Terrill et al., 2020). Thus, rather than reflect an interaction of EB with the chemogenetic excitation of LHAaMCHà NAc neurons, I may instead be capturing a phasic effect. The inability to separate acute and phasic effects of estradiol is one limitation of this study, discussed below. 115 Limitations While I also replaced EB in a four day regiment, I intentionally tested rats on the second day of EB replacement in order to capture both slower, genomic and rapid, acute effects of EB on PI performance. I expected that the level of estradiol present on the second day of EB replacement would resemble that observed in early estrus, when estrogen levels still remain high after rising steadily during the preceding day of proestrus (Asarian & Geary, 2002; Geary & Asarian, 1999; Goldman et al., 2007; Hu et al., 2004). After testing with VEH or CNO, rats received two days of PI sessions without hormone or drug delivery during a 48 h washout period before the cycle of hormone administration and chemogenetic testing was repeated. However, this procedure limits the ability to identify whether effects of EB are driven by rapid or genomic effects, and also complicates the interpretation of these results in context of data from intact, cycling animals presented in Chapter 3. It would be interesting to examine whether the influence of EB on LHAaMCHà NAc neuronal excitation is an effect of the initial dose of EB 24 hours prior to testing, the acute EB administered 30 min prior, or both. If the behavioral effects of EB on LHAaMCHà NAc neuronal excitation are reproduced after only administering one dose of EB 24 hours prior to testing, this phenotype may be comparable to that observed during metestrus, when hormone levels are relatively low after peaking ~24 hours previously. It would also be beneficial to examine whether EB modulates behavior when chemogenetic excitation of LHAaMCHà NAc neurons is applied 36-72 hours after EB pretreatment. This is a timeframe more in line with when effects of EB on MCH are typically reported (Messina et al., 2006; Santollo & Eckel, 2008; Terrill et al., 2020), and would indicate that the rise and then subsequent fall – 116 rather than current plasma EB level – is important for modulation of the effects of LHAaMCHà NAc neuronal excitation. In contrast to previous studies, DREADD viral expression in these animals was robust and easily visualized. Given that expression patterns were poor in the previous cohort, I had adapted the DREADD infusion protocol to include an additional 0.25 µl infusion, bilaterally, to the LHAa. These animals thus received an additional 0.5 µl of virus compared to the LHAa infusion performed in intact, cycling animals in Chapter 3. Animals were also sacrificed much earlier than in previous cohorts, within three months after viral infusion rather than after 8-9 months. Tissue collected from these subjects generally appeared healthier, with DREADD expression clearly visualized by the mCherry label in neuronal cell bodies as well as fibers. While this modified infusion protocol resulted in ample DREADD expression in LHAaMCH neurons, DREADD expression also extended into more posterior aspects of the LHA than in previous LHAa groups. Thus, the cells targeted may have overlapped to a greater extent with LHAp neurons targeted in Chapter 2. However, this infusion was still less than the total volume of DREADD infused to LHAp targets (1.5 µl vs 2.4 µl / rat). It is thus especially interesting that such robust DREADD expression was observed in these animals, but not in the LHAp group. While the LHAp animals were euthanized much later after DREADD expression (i.e., after 8-9 months), expression of the DREADD receptor and mCherry label should remain intact across this timeframe. However, DREADD expression depended on the presence of the pMCH promoter in LHAa neurons. Therefore, the amount of DREADD expression observed in OVX animals may arise from differences in the relative expression of pMCH in the LHA. 117 Given that estrogen inhibits pMCH expression (Messina et al., 2006; Murray et al., 2000; Santollo & Eckel, 2013), simultaneous OVX in these subjects may have permitted more robust pMCH expression by facilitating a rapid drop in plasma estrogens. This release from inhibition by estrogen would facilitate greater pMCH expression, in turn facilitating more robust DREADD expression in these OVX subjects. This potential effect could be examined by altering the timing of DREADD infusion and OVX, or by comparing DREADD expression in OVX animals receiving immediate post-operative EB vs oil treatment. Conclusion Despite these challenges, this study adds to a limited body of work examining the influence of estradiol on timing (Panfil et al., 2023; Pleil et al., 2011; Ross & Santi, 2000; Sandstrom, 2007; Williams, 2011) and provides the first direct evidence that estrogen interacts with LHAaMCH neurons to influence the timing of motivated behavior. Consistent with work examining the influence of LHAaMCH à NAc neurons on PI responding in intact rats, effects of LHAaMCH à NAc neuronal excitation also produced effects primarily on the “stop” function during the interval timing task. In addition, although a greater number of LHAaMCH à NAc DREADD-expressing neurons were identified, the behavioral phenotype observed – an accelerated “stop” function post criterion – was similar. Thus, these neurons are robustly capable of altering the “stop” function, even when only a few neuronal cell bodies are recruited. Shockingly, the effects of LHAaMCH à NAc neuronal excitation were only observed in EB-pretreated rats, suggesting that estrogen is necessary for the effects of this manipulation. It remains unclear whether effects of EB pretreatment are driven by 118 one or both doses of EB administered 24 h and 30 min prior to testing, respectively. Thus, while effects of LHAaMCH à NAc neuronal excitation appear to require EB, the mechanism of this effect (i.e., a tonic or phasic effect driven by membrane-bound or genomic ERs, respectively) remains unclear. The accelerated “stop” function observed in EB-pretreated rats resembles the behavioral phenotype observed following chemogenetic excitation of LHAaMCH à NAc neurons in M/D females. This suggests that effects of EB on LHAaMCH neurons may differ based on the source of estrogen (exogenous vs endogenous) and/ or occur through multiple mechanisms. Regardless, the present study confirms that estrogen interacts with LHAa-MCH à NAc neurons to mediate their effects on motivated behavior. In fact, the presence of estrogen in adult females appears to be necessary to observe any effects of this neuronal manipulation. In addition, this study also indicated an influence of cyclic EB replacement on the “start” function of adult OVX rats. This is consistent with sex differences reported in female mice, who generally delay the “start” function relative to males (Gür et al., 2019) and provide more insight into a potential role for endogenous estrogen on time perception than studies in which estrogen is replaced only acutely. In contrast to the baseline effects of EB, which selectively influenced the “start” function, chemogenetic excitation of LHAaMCH à NAc neurons selectively influenced the “stop” function, which is consistent effects reported in intact, cycling animals. This indicates that these neurons are capable of affecting the decision to “stop” engaging in motivated behavior, and suggests these neurons can guide food-related decision making within a temporal context that predicts food availability. 119 CHAPTER 5: Overall Discussion Summary of key findings Previously, I demonstrated that chemogenetic excitation of MCH neurons is capable of producing distinct phenotypes in the Peak Interval (PI) paradigm in female depending on the location of MCH neurons within the LHA. Excitation of MCH neurons in the LHAp reduced peak time in female rats, indicating a potential change in time perception. In contrast, LHAaMCH neuronal excitation prolonged high rate responding without affecting time perception, indicating a motivational effect on the “stop” decision process. Importantly, this effect was observed only when rats were tested during metestrus/ diestrus (M/D), when levels of circulating ovarian hormones are relatively low. This suggests that LHAMCH neurons may be capable of modulating time perception and/ or motivation in the PI task, and do so in a manner that depends on estrous cycle stage. The Nucleus Accumbens (NAc) is implicated in both timing and motivated behavior (Floresco, 2015; Kelley et al., 2005; MacDonald et al., 2012; Meck, 1996; Meck et al., 2008), is modulated by estrogen in females (Becker, 1990a; Becker & Ramirez, 1981; Robinson et al., 1980), and is a site of MCH action (Georgescu et al., 2005; Haemmerle et al., 2015; Karlsson et al., 2016; Terrill et al., 2020). Therefore, in this dissertation, I examined whether projections to the NAc from MCH neurons in the LHAp (Chapter 2) and LHAa (Chapters 3 & 4) could account for these effects. In Chapters 2 and 3, I first examined whether excitation of NAc-projecting LHAp and LHAa MCH neurons, respectively, could influence PI responding in intact, cycling female rats. Given that little is known about the effects of estrous cycle on PI responding (Gür et al., 2019; Panfil et al., 2023; Pleil et al., 2011; Ross & Santi, 2000; 120 Sandstrom, 2007; Williams, 2011), I first examined whether or not the estrous cycle influenced task performance at baseline. I then examined the influence of LHAMCH à NAc neuronal excitation separately while rats were in proestrus/ estrus (P/E) and metestrus/ diestrus (M/D). In Chapter 2, chemogenetic excitation of with LHApMCH à NAc neurons did not influence time perception in the PI paradigm, indicating that this projection is not capable of accelerating clock speed to increase early responding. Although there was potentially a subtle effect of CNO on the rate at which responding changed at 15s when rats were tested during P/E, this effect occurred without any other changes in responding (i.e., a change in peak time, proportional changes to the “start” and “stop” function) and likely does not indicate a change in time perception itself. In addition, there was also a baseline effect of estrous cycle stage on responding at 15s, indicating that multiple factors may contribute to subtle effects at this timepoint. As such, the potential effect of LHApMCH à NAc neuronal excitation on responding at 15s during P/E should be interpreted with caution. Thus, altogether, this chapter revealed that LHApMCH à NAc neurons have little to no influence on responding in the PI task, regardless of estrous cycle stage. In contrast, in Chapter 3, chemogenetic excitation of LHAaMCH à NAc neurons robustly influenced post-peak responding when rats were tested during M/D. This effect occurred with influencing the “start” function or peak time, indicating that it is a selective modulation of the decision to “stop” motivated responding after the omission of an expected reward. Thus, these neurons integrate temporal information to guide decision making and attenuate effortful responding during periods when reinforcement is not 121 likely (i.e., after the criterion). That this effect occurred only when rats were tested during M/D is in line with our previous findings in which LHAaMCH neuronal excitation also influenced post peak responding. However, previously, I demonstrated that LHAaMCH neuronal excitation prolonged – rather than attenuated! – high rate responding after the omission of an expected reward. Thus, while LHAaMCH neurons overall delay the ”stop” function, a subset of these neurons that project to the NAc instead accelerate the ”stop” function. Interestingly, this is in line with the role of the NAc in behavioral inhibition, which posits that motivated behaviors are controlled via an inhibitory influence of the NAc (Ambroggi et al., 2011; Floresco, 2015; Lafferty et al., 2020). Thus, NAc- projecting LHAaMCH neurons may modulate activity within the NAc to inhibit high rate responding after the criterion duration has elapsed and reinforcement is perceived as being unlikely. Changes to the “stop” function in interval timing procedures are often interpreted in terms of motivation because they represent a form of perseverative responding that occurs in the absence of reinforcer delivery, even as the animal correctly perceives the criterion time as having elapsed. Thus, given that LHAaMCH à NAc neurons influenced a form of motivated responding in the PI task, I also examined whether they could more broadly modulate motivated behavior in task that measures motivated responding more directly: the progressive ratio (PR) task. Despite having altered how quickly rats give up high rate responding after the criterion duration in the PI task, LHAaMCH à NAc neuronal excitation failed to influence when rats gave up in the PR task. That is, rats responded comparably regardless of LHAaMCH à NAc neuronal excitation. In addition, there were no estrous cycle effects on responding in the PR task, which is in line with previous 122 findings suggesting estrous cycle influences PR responding only during early training (Quigley et al., 2021). That there were no effects of DREADD manipulation on PR responding coincide with the lack of effect observed on overall response rates in the PI task. That is, chemogenetic excitation of LHAaMCH à NAc neurons also did not influence overall response rate or magnitude during the PI task, but rather selectively modulated the profile of responding across time within trials. Thus, the influence of LHAaMCH à NAc neurons on behavior is time-dependent, indicating an ability of these neurons to incorporate information from food predictive cues to determine when reinforcer is likely to become available. Rather than modulate the amount of effort extended over a behavioral session, these neurons modulate how that effort is extended (i.e., by coordinating responding around the time of expected reinforcer delivery). Because these effects of LHAaMCH à NAc neuronal excitation on the “stop” function were observed only when rats were tested during M/D, in Chapter 4 I ovariectomized rats to isolate the effect of estradiol (EB) on DREADD-mediated PI responding. Shockingly, chemogenetic excitation of LHAaMCH à NAc neurons in adult OVX rats failed to influence PI performance! Because LHAaMCH à NAc neurons influence motivated behavior selectively during M/D, when plasma estrogen is relatively low, I hypothesized that chemogenetic excitation of these neurons in oil pretreated rats would produce robust effects on the “stop” function. In addition, evidence suggests that sex differences in striatal circuits underlying interval timing are regulated by genetic or prenatal organization and these differences cannot be reversed by hormones in adulthood (Pleil et al., 2011). Thus, I expected that an absence of EB during adulthood 123 would simply enable LHAaMCH neurons to influence behavior without being inhibited by endogenous estrogens. Perhaps even more shockingly, when adult OVX rats were treated with EB, chemogenetic excitation of LHAaMCH à NAc neurons attenuated post peak responding, consistent with an effect on the “stop” function. Although the absence of high levels of circulating estrogen was associated with DREADD effects in intact animals, EB priming was necessary for DREADD effects in OVX animals! Thus, estrogen is necessary in some capacity to sensitize the system to the effects of LHAaMCH à NAc neuronal excitation. In addition, although the hormonal condition at test (relatively low vs high [EB]) differed, the direction of the effect was consistent: chemogenetic excitation of LHAaMCH à NAc neurons reduced high rate after the peak, indicating an acceleration of the “stop” function. In addition, in OVX rats this change in “stop” function was also sufficient to influence peak time, as CNO treatment resulted in a decreased peak time in EB pretreated rats, without proportionally altering the “start” function. Altogether, results from this dissertation indicate that MCH neurons in the LHAa – but not the LHAp – control motivated behavior through the “stop” function. In addition, these effects are modulated by estrous cycle stage and EB, perhaps requiring fluctuations in EB for the expression of a behavioral phenotype. Given that both MCH and estrogen influence feeding behavior through meal size, they may interact to determine the “stop” function to determine when bouts of feeding are terminated. Limitations While these studies clearly indicate a role for NAc-projecting MCH neurons in the LHAa, but not LHAp, clarity regarding how these populations of neurons differ is limited. 124 Furthermore, limited DREADD expression in both the LHAa and LHAp of intact, cycling animals relative to OVX animals suggests that DREADD expression itself – which is controlled by the pMCH promoter – may be influenced by OVX and circulating estrogen. This is not surprising, given that the pMCH promoter is downregulated by EB replacement in OVX rats (Messina et al., 2006), but suggests a need for examining this interaction more directly. A second limitation of the present studies is that while DREADD expression was selectively targeted to neurons expressing the pMCH promoter, these cells are capable of producing other peptides and neurotransmitters, which may be released in addition to, or in lieu of, the MCH peptide (Bonnavion et al., 2016; Mickelsen et al., 2017, 2019). Without the coadministration of an MCH1R antagonist, it is not possible to isolate behavioral effects of this DREADD manipulation to the MCH peptide. For example, it is also possible that these LHAMCH neurons corelease glutamate, which could also modulate the activity of MSNs in the NAc to alter motivated responding. Future studies could address this shortcoming by administering an MCH1R antagonist (e.g., H6408) with CNO. Although I intended to complete these studies via ICV infusion of H6408 prior to CNO infusion, issues with cannula patency prevented me from reaching a sufficient sample size to make any meaningful conclusions. There is limited data to suggest that H6408 is capable of crossing the blood brain barrier in rats , but preliminary data from a subset of rats (n=2) suggests that i.p. administration of this ligand may be capable of MCH1R binding in the CNS. Regardless, with a limited sample size and knowledge of the pharmacokinetics, this data is also not included. Future work should include pharmacological antagonism of the MCH1R receptor in conjunction with chemogenetic 125 manipulations, or directly examine the influence of intra-NAc MCH on PI performance to isolate a role for the MCH peptide. Conclusions In conclusion, this body of work provides evidence for a role of LHAMCH neurons on the timing of motivated behavior, particularly within the context of when to “stop” effortful responding. In particular, MCH neurons in the anterior LHA are important for modulating the “stop” function. Previously, I demonstrated that non-projection-specific excitation of LHAaMCH neurons prolonged high rate responding by delaying the "stop” function. Here, I demonstrated that projections from these LHAaMCH neurons to the NAc do not underlie this effect. Instead, LHAaMCH à NAc neurons attenuate post peak responding by accelerating the “stop” function. This indicates a bidirectional control of the “stop” by LHAaMCH neurons, depending in part on their afferents. Critically, these effects were observed only in females tested during M/D, indicating that these neurons are modulated by the estrous cycle in intact, cycling females. Furthermore, an influence of estradiol was isolated through OVX and selective EB replacement. This manipulation indicated that EB is necessary to observe behavioral effects of LHAaMCH à NAc neuronal excitation, as there was no effect of DREADD manipulations in oil-treated rats. Given that both MCH and estrogen modulate food intake via changes in meal size (Baird et al., 2006; Eckel, 2011; Messina et al., 2006), their effects may occur through changes to the “stop” decision that determines when to stop consuming. Although the present studies do not directly examine consumption patterns, and the “stop” function is specifically modulated in the absence of an expected food reinforcer, they still provide evidence for the behavioral control of inhibition by MCH and estrogen. 126 Specifically, the PI paradigm provides unique insight into the control of the “stop” decision by examining how motivated responding changes within a trial based on the expectation of reinforcer delivery, which is informed by the passage of time. Thus, these neurons integrate temporal information to guide effortful behavior. Changes to the decision of when to “stop” high rate responding reflect control of these neurons over motivated responding in real time. That these effects are time-dependent, and occur only in the post-peak period after the omission of an expected reward, further indicates the subtle control of these LHAaMCH on decision processes affecting motivated behavior. Altogether, these data suggest that MCH and estrogen interact to influence the “stop” decision in motivated food-seeking. Typically, estrogen decreases food intake by reducing meal size, while MCH promotes food intake by increasing meal size – in both cases, these effects may also occur through changes in the decision processes that guide when to “stop” feeding. The present study demonstrates that MCH neurons in the anterior LHA interact with estrogen to modulate the “stop” decision in an interval timing task. This provides a potential mechanism through which estrogen and MCH may interact to influence the “stop” decision process in feeding-related behaviors, especially those that include a temporal component. Thus, these neurons could influence a variety of learned and motivated feeding behaviors, providing a potential mechanism to explore in maladaptive feeding behaviors. 127 BIBLIOGRAPHY Almey, A., Milner, T. A., & Brake, W. G. (2015). Estrogen receptors in the central nervous system and their Implication for Dopamine-Dependent Cognition in Females. Hormones and Behavior, 74, 125–138. https://doi.org/10.1016/j.yhbeh.2015.06.010.Estrogen Ambroggi, F., Ghazizadeh, A., Nicola, S. M., & Fields, H. L. (2011). Roles of Nucleus Accumbens Core and Shell in Incentive-Cue Responding and Behavioral Inhibition. Journal of Neuroscience, 31(18), 6820–6830. https://doi.org/10.1523/JNEUROSCI.6491-10.2011 Arrigoni, E., Chee, M. J. S., & Fuller, P. M. (2019). To eat or to sleep: That is a lateral hypothalamic question. Neuropharmacology, 154(August 2018), 34–49. https://doi.org/10.1016/j.neuropharm.2018.11.017 Asarian, L., & Geary, N. (2002). Cyclic estradiol treatment normalizes body weight and restores physiological patterns of spontaneous feeding and sexual receptivity in ovariectomized rats. Hormones and Behavior, 42(4), 461–471. https://doi.org/10.1006/hbeh.2002.1835 Bäckberg, M., Ultenius, C., Fritschy, J. M., & Meister, B. (2004). Cellular localization of GABAA receptor α subunit immunoreactivity in the rat hypothalamus: Relationship with neurones containing orexigenic or anorexigenic peptides. Journal of Neuroendocrinology, 16(7), 589–604. https://doi.org/10.1111/j.1365- 2826.2004.01207.x Baird, J., Rios, C., Gray, N. E., Walsh, C. E., Fischer, S. G., Pecora, A. L., Rios, C., Gray, N. E., Walsh, E., & Fischer, S. G. (2006). Effects of melanin-concentrating hormone on licking microstructure and brief-access taste responses. American Journal of Physiology - Regulatory Integrative and Comparative Physiology, 291(5), 1265–1274. https://doi.org/10.1152/ajpregu.00143.2006. Balcı, F. (2014). Interval Timing, Dopamine, and Motivation. Timing & Time Perception, 2(3), 379–410. https://doi.org/10.1163/22134468-00002035 Balcı, F., & Freestone, D. (2020). The Peak Interval Procedure in Rodents: A Tool for Studying the Neurobiological Basis of Interval Timing and Its Alterations in Models of Human Disease. Bio-Protocol, 10(17), 1–20. https://doi.org/10.21769/bioprotoc.3735 Balsam, P., Drew, M., & Gallistel, C. (2010). Time and Associative Learning. Comparative Cognition & Behavior Reviews, 5, 1–22. https://doi.org/10.3819/ccbr.2010.50001 Balsam, P., Sanchez-Castillo, H., Taylor, K., Van Volkinburg, H., & Ward, R. D. (2009). 128 Timing and anticipation: Conceptual and methodological approaches. European Journal of Neuroscience, 30(9), 1749–1755. https://doi.org/10.1111/j.1460- 9568.2009.06967.x Bass, J., & Takahashi, J. S. (2010). Circadian Integration of Metabolism and Energetics. Science, 330(6009), 1349–1354. Bayer, J., Rusch, T., Zhang, L., Gläscher, J., & Sommer, T. (2020). Dose-dependent effects of estrogen on prediction error related neural activity in the nucleus accumbens of healthy young women. Psychopharmacology, 237(3), 745–755. https://doi.org/10.1007/s00213-019-05409-7 Becker, J. B. (1990a). Direct effect of 17β-estradiol on striatum: Sex differences in dopamine release. Synapse, 5(2), 157–164. https://doi.org/10.1002/syn.890050211 Becker, J. B. (1990b). Estrogen rapidly potentiates amphetamine-induced striatal dopamine release and rotational behavior during microdialysis. Neuroscience Letters, 118(2), 169–171. https://doi.org/10.1016/0304-3940(90)90618-J Becker, J. B., & Ramirez, V. D. (1981). (Accepted June 26th, 1980). Water, 204, 361– 372. Berridge, K. C. (2004). Motivation concepts in behavioral neuroscience. Physiology and Behavior, 81(2), 179–209. https://doi.org/10.1016/j.physbeh.2004.02.004 Berridge, K. C., & Valenstein, E. S. (1991). What Psychological Process Mediates Feeding Evoked by Electrical Stimulation of the Lateral Hypothalamus? Behavioral Neuroscience, 105(1), 3–14. https://doi.org/10.1037/0735-7044.105.1.3 Berthoud, H. R., & Münzberg, H. (2011). The lateral hypothalamus as integrator of metabolic and environmental needs: From electrical self-stimulation to opto- genetics. Physiology and Behavior, 104(1), 29–39. https://doi.org/10.1016/j.physbeh.2011.04.051 Bittencourt, J., Presse, F., Arias, C., Peto, C., Vaughan, J., Nahon, J., Vale, W., & Sawchenko, P. (1992). The melanin concentrating hormone system of the rat brain An immuno and hybridisation histochemical characterisation.pdf. The Journal of Comparative Neurology, 319, 218–245. https://doi.org/10.1002/cne.903190204 Blaustein, J. D., & Wade, G. N. (1976). Ovarian influences on the meal patterns of female rats. Physiology and Behavior, 17(2), 201–208. https://doi.org/10.1016/0031-9384(76)90064-0 Block, R. A., Hancock, P. A., & Zakay, D. (2000). Sex differences in duration judgments: A meta-analytic review. Memory and Cognition, 28(8), 1333–1346. https://doi.org/10.3758/BF03211834 129 Blouin, A. M., Fried, I., Wilson, C. L., Staba, R. J., Behnke, E. J., Lam, H. A., Maidment, N. T., Karlsson, K. Æ., Lapierre, J. L., & Siegel, J. M. (2013). Human hypocretin and melanin-concentrating hormone levels are linked to emotion and social interaction. Nature Communications, May 2012. https://doi.org/10.1038/ncomms2461 Bonnavion, P., Mickelsen, L. E., Fujita, A., de Lecea, L., & Jackson, A. C. (2016). Hubs and spokes of the lateral hypothalamus: cell types, circuits and behaviour. Journal of Physiology, 594(22), 6443–6462. https://doi.org/10.1113/JP271946 Boulware, M. I., Weick, J. P., Becklund, B. R., Kuo, S. P., Groth, R. D., & Mermelstein, P. G. (2005). Estradiol Activates Group I and II Metabotropic Glutamate Receptor Signaling , Leading to Opposing Influences on cAMP Response Element-Binding Protein. The Journal of Neuroscience, 25(20), 5066–5078. https://doi.org/10.1523/JNEUROSCI.1427-05.2005 Broberger, C., De Lecea, L., Sutcliffe, J. G., & Hökfelt, T. (1998). Hypocretin/orexin- and melanin-concentrating hormone-expressing cells form distinct populations in the rodent lateral hypothalamus: relationship to the neuropeptide y and agouti gene- related protein systems. Journal of Comparative Neurology, 402(4), 460–474. https://doi.org/10.1002/(sici)1096-9861(19981228)402:4<460::aid-cne3>3.0.co;2-s Brown, J. A., Woodworth, H. L., & Leinninger, G. M. (2015). To ingest or rest? Specialized roles of lateral hypothalamic area neurons in coordinating energy balance. Frontiers in Systems Neuroscience, 9(FEB), 1–25. https://doi.org/10.3389/fnsys.2015.00009 Buhusi, M., Bartlett, M. J., & Buhusi, C. V. (2017). Sex differences in interval timing and attention to time in C57Bl/6J mice. Behavioural Brain Research, 324, 96–99. https://doi.org/10.1016/j.bbr.2017.02.023 Buhusi, C. V., & Meck, W. H. (2002). Differential effects of methamphetamine and haloperidol on the control of an internal clock. Behavioral Neuroscience, 116(2), 291–297. https://doi.org/10.1037/0735-7044.116.2.291 Buhusi, C. V., & Meck, W. H. (2005). What makes us tick? Functional and neural mechanisms of interval timing. Nature Reviews Neuroscience, 6(10), 755–765. https://doi.org/10.1038/nrn1764 Buhusi, C. V., & Meck, W. H. (2006). Time sharing in rats: A peak-interval procedure with gaps and distracters. Behavioural Processes, 71(2–3), 107–115. https://doi.org/10.1016/j.beproc.2005.11.017 Buhusi, C. V., & Meck, W. H. (2009). Relative time sharing: New findings and an extension of the resource allocation model of temporal processing. Philosophical Transactions of the Royal Society B: Biological Sciences, 364(1525), 1875–1885. 130 https://doi.org/10.1098/rstb.2009.0022 Butera, P. C. (2010). Estradiol and the control of food intake. Physiology & Behavior, 99(2), 175–180. https://doi.org/10.1016/j.physbeh.2009.06.010 Caggiula, A. R., & Hoebel, B. (1966). “Copulation-Reward Site” in the Posterior Hypothalamus Author. Science, 153(3741), 1284–1285. Challet, E. (2019). The circadian regulation of food intake. Nature Reviews Endocrinology, 15(7), 393–405. https://doi.org/10.1038/s41574-019-0210-x Chambers, J., Ames, R. S., Bergsma, D., Muir, A., Fitzgerald, L. R., Hervieu, G., Dytko, G. M., Foley, J. J., Martin, J., Liu, W. S., Park, J., Ellis, C., Ganguly, S., Konchar, S., Cluderay, J., Leslie, R., Wilson, S., & Sarau, H. M. (1999). Melanin- concentrating hormone is the cognate ligand for the orphan G- protein-coupled receptor SLC-1. Nature, 400(6741), 261–265. https://doi.org/10.1038/22313 Chee, M. J. S., Arrigoni, E., & Maratos-Flier, E. (2015). Melanin-concentrating hormone neurons release glutamate for feedforward inhibition of the lateral septum. Journal of Neuroscience, 35(8), 3644–3651. https://doi.org/10.1523/JNEUROSCI.4187- 14.2015 Church, R. M. (1984). Properties of the Internal Clock. In J. Gibbon & A. Lorraine (Eds.), Time and Time Perception (pp. 566–582). Annals of the New York Academy of Sciences. Church, R. M., & Broadbent, H. A. (1991). A connectionist model of timing. In M. L. Commons, S. Grossberg, & J. E. R. Staddon (Eds.), Quantitative Models of Behavior: Neural Networks and Conditioning. Lawrence Erlbaum Associates, Inc. https://books.google.com/books?hl=en&lr=&id=oeYcDQAAQBAJ&oi=fnd&pg=PT24 0&dq=A+connectionist+model+of+timing&ots=eWUT5DUVv- &sig=e4J03AdA5H4oWtcE_qLChnlyxRY Church, R. M., Meck, W. H., & Gibbon, J. (1994). Application of Scalar Timing Theory to Individual Trials. Journal of Experimental Psychology: Animal Behavior Processes, 20(2), 135–155. https://doi.org/10.1037/0097-7403.20.2.135 Couse, J. F., Lindzey, J., Grandien, K. A. J., Gustafsson, J.-åke, & Korach, K. S. (1997). Tissue Distribution and Quantitative Analysis of Estrogen Receptor- a (ER a ) and Estrogen Receptor-b (ERb) Messenger Ribonucleic Acid in the Wild-Type and ER a -Knockout Mouse. Endocrinology, 138(11). Davis, J. D., & Smith, G. P. (1988). Analysis of lick rate measures the positive and negative feedback effects of carbohydrates on eating. Appetite, 11(3), 229–238. Davis, J. D., & Smith, G. P. (1992). Analysis of the Microstructure of the Rhythmic 131 Tongue Movements of Rats Ingesting Maltose and Sucrose Solutions. Behavioral Neuroscience, 106(1), 217–228. https://doi.org/10.1037/0735-7044.106.1.217 Della-Zuana, O., Presse, F., Ortola, C., Duhault, J., Nahon, J. L., & Levens, N. (2002). Acute and chronic administration of melanin- concentrating hormone enhances food intake and body weight in Wistar and Sprague – Dawley rats. International Journal of Obesity, 26, 1289–1295. Dilsiz, P., Aklan, I., Sayar Atasoy, N., Yavuz, Y., Filiz, G., Koksalar, F., Ates, T., Oncul, M., Coban, I., Ates Oz, E., Cebecioglu, U., Alp, M. I., Yilmaz, B., & Atasoy, D. (2020). MCH Neuron Activity Is Sufficient for Reward and Reinforces Feeding. Neuroendocrinology, 110(3–4), 258–270. https://doi.org/10.1159/000501234 Diniz, G. B., & Bittencourt, J. C. (2017). The melanin-concentrating hormone as an integrative peptide driving motivated behaviors. Frontiers in Systems Neuroscience, 11(May), 1–26. https://doi.org/10.3389/fnsys.2017.00032 Eckel, L. A. (2004). Estradiol: A rhythmic, inhibitory, indirect control of meal size. Physiology and Behavior, 82(1), 35–41. https://doi.org/10.1016/j.physbeh.2004.04.023 Eckel, L. A. (2011). The ovarian hormone estradiol plays a crucial role in the control of food intake in females. Physiology and Behavior, 104(4), 517–524. https://doi.org/10.1016/j.physbeh.2011.04.014 Fiorillo, C. D., Tobler, P. N., & Schultz, W. (2003). Discrete coding of reward probability and uncertainty by dopamine neurons. Science, 299(5614), 1898–1902. https://doi.org/10.1126/science.1077349 Floresco, S. B. (2015). The nucleus accumbens: An interface between cognition, emotion, and action. Annual Review of Psychology, 66, 25–32. https://doi.org/10.1146/annurev-psych-010213-115159 Gallardo, M. G. P., Chiocchio, S. R., & Tramezzani, J. H. (2004). Changes of melanin- concentrating hormone related to LHRH release in the median eminence of rats. Brain Research, 1030(1), 152–158. https://doi.org/10.1016/j.brainres.2004.10.005 Gallistel, C. R., & Gibbon, J. (2000). Time, Rate, and Conditioning. Psychological Review, 107(2), 289–344. Galtress, T., & Kirkpatrick, K. (2009). Reward value effects on timing in the peak procedure. Learning and Motivation, 40(2), 109–131. https://doi.org/10.1016/j.lmot.2008.05.004 Galtress, T., & Kirkpatrick, K. (2010). Reward magnitude effects on temporal discrimination. Learning and Motivation, 41(2), 108–124. 132 https://doi.org/10.1016/j.lmot.2010.01.002 Geary, N., & Asarian, L. (1999). Cyclic estradiol treatment normalizes body weight and test meal size in ovariectomized rats. Physiology and Behavior, 67(1), 141–147. https://doi.org/10.1016/S0031-9384(99)00060-8 Georgescu, D., Sears, R. M., Hommel, J. D., Barrot, M., Bolaños, C. A., Marsh, D. J., Bednarek, M. A., Bibb, J. A., Maratos-Flier, E., Nestler, E. J., & DiLeone, R. J. (2005). The hypothalamic neuropeptide melanin-concentrating hormone acts in the nucleus accumbens to modulate feeding behavior and forced-swim performance. Journal of Neuroscience, 25(11), 2933–2940. https://doi.org/10.1523/JNEUROSCI.1714-04.2005 Gibbon, J. (1977). Scalar expectancy theory and Weber’s law in animal timing. Psychological Review, 84(3), 279–325. https://doi.org/10.1037/0033-295X.84.3.279 Goldman, J. M., Murr, A. S., & Cooper, R. L. (2007). The Rodent Estrous Cycle. Birth Defects Research, 80(2), 83–97. Gomori, A., Ishihara, A., Ito, M., Mashiko, S., Ishihara, A., Ito, M., Matsushita, H., Yumoto, M., Ito, M., Tanaka, T., & Tokita, S. (2023). Chronic intracerebroventricular infusion of MCH causes obesity in mice. 583–588. Goodless-Sanchez, N., Moore, R. Y., & Morin, L. P. (1991). Lateral hypothalamic regulation of circadian rhythm phase. Physiology and Behavior, 49(3), 533–537. https://doi.org/10.1016/0031-9384(91)90276-T Grove-Strawser, D., Boulware, M. I., & Mermelstein, P. G. (2010). Membrane estrogen receptors activate the metabotropic glutamate receptors mGluR5 and mGluR3 to bidirectionally regulate CREB phosphorylation in female rat striatal neurons. Neuroscience, 170(4), 1045–1055. https://doi.org/10.1016/j.neuroscience.2010.08.012 Gür, E., Fertan, E., Kosel, F., Wong, A. A., Balcı, F., & Brown, R. E. (2019). Sex differences in the timing behavior performance of 3xTg-AD and wild-type mice in the peak interval procedure. Behavioural Brain Research, 360(September 2018), 235–243. https://doi.org/10.1016/j.bbr.2018.11.047 Haemmerle, C. A. S., Campos, A. M. P., & Bittencourt, J. C. (2015). Melanin- concentrating hormone inputs to the nucleus accumbens originate from distinct hypothalamic sources and are apposed to GABAergic and cholinergic cells in the Long-Evans rat brain. Neuroscience, 289, 392–405. https://doi.org/10.1016/j.neuroscience.2015.01.014 Hahn, J. D., & Swanson, L. W. (2010). Distinct patterns of neuronal inputs and outputs of the juxtaparaventricular and suprafornical regions of the lateral hypothalamic 133 area in the male rat. Brain Research Reviews, 64(1), 14–103. https://doi.org/10.1016/j.brainresrev.2010.02.002 Hahn, J. D., & Swanson, L. W. (2015). Connections of the juxtaventromedial region of the lateral hypothalamic area in the male rat. Frontiers in Systems Neuroscience, 9(MAY), 1–53. https://doi.org/10.3389/fnsys.2015.00066 Harthoorn, L. F., Sañé, A., Nethe, M., & Van Heerikhuize, J. J. (2005). Multi- transcriptional profiling of melanin-concentrating hormone and orexin-containing neurons. Cellular and Molecular Neurobiology, 25(8), 1209–1223. https://doi.org/10.1007/s10571-005-8184-8 Hassani, O. K., Lee, M. G., & Jones, B. E. (2009). Melanin-concentrating hormone neurons discharge in a reciprocal manner to orexin neurons across the sleep-wake cycle. Proceedings of the National Academy of Sciences, 106(7), 2418–2422. https://doi.org/10.1073/pnas.0811400106 Hawes, B. E., Erin, K. I. L., Green, B., O’Neill, K. I. M., Fried, S., & Graziano, M. P. (2000). The melanin-concentrating hormone receptor couples to multiple G proteins to activate diverse intracellular signaling pathways. Endocrinology, 141(12), 4524– 4532. https://doi.org/10.1210/endo.141.12.7833 Hervieu, G. J., Cluderay, J. E., Harrison, D., Meakin, J., Maycox, P., Nasir, S., & Leslie, R. A. (2000). The distribution of the mRNA and protein products of the melanin- concentrating hormone (MCH) receptor gene, slc-1, in the central nervous system of the ratt. European Journal of Neuroscience, 12(4), 1194–1216. https://doi.org/10.1046/j.1460-9568.2000.00008.x Hinton, S. C., & Meck, W. H. (2004). Frontal-striatal circuitry activated by human peak- interval timing in the supra-seconds range. Cognitive Brain Research, 21(2), 171– 182. https://doi.org/10.1016/j.cogbrainres.2004.08.005 Hoebel, B., & Teitelbaum, P. (1962). Hypothalamic Control of Feeding and Self- Stimulation Author ( s ): Bartley G . Published by : American Association for the Advancement of Science Stable URL : http://www.jstor.org/stable/1709352 Accessed : 15-07-2016 22 : 49 U. American Association for the Advancement of Science, 135(3501), 375–377. Hollerman, J. R., & Schultz, W. (1998). Dopamine neurons report an error in the temporal prediction of reward during learning. Nature Neuroscience, 1(4), 304–309. https://doi.org/10.1038/1124 Hu, M., Crombag, H. S., Robinson, T. E., & Becker, J. B. (2004). Biological Basis of Sex Differences in the Propensity to Self-administer Cocaine. Neuropsychopharmacology, 29(1), 81–85. https://doi.org/10.1038/sj.npp.1300301 134 Huang, H., & van den Pol, A. N. (2007). Rapid Direct Excitation and Long-Lasting Enhancement of NMDA Response by Group I Metabotropic Glutamate Receptor Activation of Hypothalamic Melanin- Concentrating Hormone Neurons. The Jounral of Neuroscience, 27(43), 11560–11572. https://doi.org/10.1523/JNEUROSCI.2147- 07.2007 Huang, W., Ramsey, K. M., Marcheva, B., & Bass, J. (2011). Huang 2011 circadian rhythms sleep food.pdf (pp. 2133–2141). Hutchinson, R. R., & Renfrew, J. W. (1966). STALKING ATTACK AND EATING BEHAVIORS ELICITED FROM THE SAME SITES IN THE HYPOTHALAMUS 1. Journal of Comparative and Physiological Psychology, 61(3), 360–367. Jego, S., Glasgow, S. D., Herrera, C. G., Ekstrand, M., Reed, S. J., Boyce, R., Friedman, J., Burdakov, D., & Adamantidis, A. R. (2013). Optogenetic identification of a rapid eye movement sleep modulatory circuit in the hypothalamus. Nature Neuroscience, 16(11), 1637–1643. https://doi.org/10.1038/nn.3522 Johnson, A. W. (2018). Characterizing ingestive behavior through licking microstructure: Underlying neurobiology and its use in the study of obesity in animal models. International Journal of Developmental Neuroscience, 64(March 2017), 38–47. https://doi.org/10.1016/j.ijdevneu.2017.06.012 Kacelnik, A., & Brunner, D. (2002). Timing and foraging: Gibbon’s scalar expectancy theory and optimal patch exploitation. Learning and Motivation, 33(1), 177–195. https://doi.org/10.1006/lmot.2001.1110 Karlsson, C., Aziz, A. M. A., Rehman, F., Pitcairn, C., Barchiesi, R., Barbier, E., Wendel Hansen, M., Gehlert, D., Steensland, P., Heilig, M., & Thorsell, A. (2016). Melanin- Concentrating Hormone and Its MCH-1 Receptor: Relationship Between Effects on Alcohol and Caloric Intake. Alcoholism: Clinical and Experimental Research, 40(10), 2199–2207. https://doi.org/10.1111/acer.13181 Kelley, A. E. (2004). Ventral striatal control of appetitive motivation: Role in ingestive behavior and reward-related learning. Neuroscience and Biobehavioral Reviews, 27(8), 765–776. https://doi.org/10.1016/j.neubiorev.2003.11.015 Kelley, A. E., Baldo, B. A., Pratt, W. E., & Will, M. J. (2005). Corticostriatal-hypothalamic circuitry and food motivation: Integration of energy, action and reward. Physiology and Behavior, 86(5), 773–795. https://doi.org/10.1016/j.physbeh.2005.08.066 Kelley, A. E., Bless, E. P., & Swanson, C. J. (1996). Investigation of the effects of opiate antagonists infused into the nucleus accumbens on feeding and sucrose drinking in rats. Journal of Pharmacology and Experimental Therapeutics, 278(3), 1499–1507. 135 Kurti, A. N., & Matell, M. S. (2011). Nucleus accumbens dopamine modulates response rate but not response timing in an interval timing task. Behavioral Neuroscience, 125(2), 215–225. https://psycnet.apa.org/record/2011-06370-006 Lafferty, C. K., Yang, A. K., Mendoza, J. A., & Britt, J. P. (2020). Nucleus Accumbens Cell Type- and Input-Specific Suppression of Unproductive Reward Seeking. Cell Reports, 30(11), 3729-3742.e3. https://doi.org/10.1016/j.celrep.2020.02.095 Lee, J., Raycraft, L., & Johnson, A. W. (2021). The dynamic regulation of appetitive behavior through lateral hypothalamic orexin and melanin concentrating hormone expressing cells. Physiology and Behavior, 229(October 2020), 113234. https://doi.org/10.1016/j.physbeh.2020.113234 Lewis, P. A., & Miall, R. C. (2003). Brain activation patterns during measurement of sub- and supra-second intervals. Neuropsychologia, 41(12), 1583–1592. https://doi.org/10.1016/S0028-3932(03)00118-0 Li, X., Schwartz, P. E., & Rissman, E. F. (1997). Distribution of estrogen receptor-β-like immunoreactivity in rat forebrain. Neuroendocrinology, 66(2), 63–67. https://doi.org/10.1159/000127221 López, M., & Tena-Sempere, M. (2015). Estrogens and the control of energy homeostasis: A brain perspective. Trends in Endocrinology and Metabolism, 26(8), 411–421. https://doi.org/10.1016/j.tem.2015.06.003 Ludwig, D. S., Flier, J. S., Maratos-flier, E., Ludwig, D. S., Tritos, N. A., Mastaitis, J. W., Kulkarni, R., Kokkotou, E., Elmquist, J., Lowell, B., Flier, J. S., & Maratos-flier, E. (2001). Melanin-concentrating hormone overexpression in transgenic mice leads to obesity and insulin resistance Find the latest version : Melanin-concentrating hormone overexpression in transgenic mice leads to obesity and insulin resistance. 107(3), 379–386. MacDonald, C. J., Cheng, R.-K., & Meck, W. H. (2012). Acquisition of “Start” and “Stop” response thresholds in peak-interval timing is differentially sensitive to protein synthesis inhibition in the dorsal and ventral striatum. Frontiers in Integrative Neuroscience, 6(March), 1–16. https://doi.org/10.3389/fnint.2012.00010 Malapani, C., & Fairhurst, S. (2002). Scalar timing in animals and humans. Learning and Motivation, 33(1), 156–176. https://doi.org/10.1006/lmot.2001.1105 Marsh, D. J., Weingarth, D. T., Novi, D. E., Chen, H. Y., Trumbauer, M. E., Chen, A. S., Guan, X. M., Jiang, M. M., Feng, Y., Camacho, R. E., Shen, Z., Frazier, E. G., Yu, H., Metzger, J. M., Kuca, S. J., Shearman, L. P., Gopal-Truter, S., MacNeil, D. J., Strack, A. M., … Qian, S. (2002). Melanin-concentrating hormone 1 receptor- deficient mice are lean, hyperactive, and hyperphagic and have altered metabolism. Proceedings of the National Academy of Sciences of the United States 136 of America, 99(5), 3240–3245. https://doi.org/10.1073/pnas.052706899 Marshall, A. T., Smith, A. P., & Kirkpa. (2014). Mechanisms of Impulsive Choice: I. Individual Differences in Interval Timing and Reward Processing. Journal of the Experimental Analysis of Behavior, 102, 86–101. https://doi.org/10.1007/s10071- 020-01456-2 Matell, M. S., Bateson, M., & Meck, W. H. (2006). Single-trials analyses demonstrate that increases in clock speed contribute to the methamphetamine-induced horizontal shifts in peak-interval timing functions. Psychopharmacology, 188(2), 201–212. https://doi.org/10.1007/s00213-006-0489-x Matell, M. S., & Meck, W. H. (2004). Cortico-striatal circuits and interval timing: Coincidence detection of oscillatory processes. Cognitive Brain Research, 21(2), 139–170. https://doi.org/10.1016/j.cogbrainres.2004.06.012 Matell, M. S., Meck, W. H., & Nicolelis, M. A. L. (2003). Interval timing and the encoding of signal duration by ensembles of cortical and striatal neurons. Behavioral Neuroscience, 117(4), 760–773. https://doi.org/10.1037/0735-7044.117.4.760 McEwen, B. S. (2002). Recent Progress in Hormone Research. In Endocrine Reviews (Vol. 57, Issue 1). https://doi.org/10.2307/1292058 Meck, W. H. (1983). Selective adjustment of the speed of internal clock and memory processes. Journal of Experimental Psychology: Animal Behavior Processes, 9(2), 171–201. https://doi.org/10.1037/0097-7403.9.2.171 Meck, W. H. (1996). Neuropharmacology of timing and time perception. Cognitive Brain Research, 3(3–4), 227–242. https://doi.org/10.1016/0926-6410(96)00009-2 Meck, W. H. (2005). Neuropsychology of timing and time perception. Brain and Cognition, 58(1), 1–8. https://doi.org/10.1016/j.bandc.2004.09.004 Meck, W. H. (2006). Frontal cortex lesions eliminate the clock speed effect of dopaminergic drugs on interval timing. Brain Research, 1108(1), 157–167. https://doi.org/10.1016/j.brainres.2006.06.046 Meck, W. H., Cheng, R., Macdonald, C. J., Gainetdinov, R. R., Caron, M. G., & Özlem, M. (2012). Gene-dose dependent effects of methamphetamine on interval timing in dopamine-transporter knockout mice. Neuropharmacology, 62(3), 1221–1229. https://doi.org/10.1016/j.neuropharm.2011.01.042 Meck, W. H., & Church, R. M. (1987). Cholinergic Modulation of the Content of Temporal Memory. Behavioral Neuroscience, 101(4), 457–464. https://doi.org/10.1037/0735-7044.101.4.457 137 Meck, W. H., Penney, T. B., & Pouthas, V. (2008). Cortico-striatal representation of time in animals and humans. Current Opinion in Neurobiology, 18(2), 145–152. https://doi.org/10.1016/j.conb.2008.08.002 Mello, G. B. M., Soares, S., & Paton, J. J. (2015). A Scalable Population Code for Time in the Striatum. Current Biology, 25(9), 1113–1122. https://doi.org/10.1016/j.cub.2015.02.036 Merchant, H., Harrington, D. L., & Meck, W. H. (2013). Neural basis of the perception and estimation of time. Annual Review of Neuroscience, 36, 313–336. https://doi.org/10.1146/annurev-neuro-062012-170349 Messina, M. M., Boersma, G., Overton, J. M., & Eckel, L. A. (2006). Estradiol decreases the orexigenic effect of melanin-concentrating hormone in ovariectomized rats. Physiology and Behavior, 88(4–5), 523–528. https://doi.org/10.1016/j.physbeh.2006.05.002 Micevych, P. E., & Mermelstein, P. G. (2008). Membrane estrogen receptors acting through metabotropic glutamate receptors: An emerging mechanism of estrogen action in brain. Molecular Neurobiology, 38(1), 66–77. https://doi.org/10.1007/s12035-008-8034-z Mickelsen, L. E., Bolisetty, M., Chimileski, B. R., Fujita, A., Beltrami, E. J., Costanzo, J. T., Naparstek, J. R., Robson, P., & Jackson, A. C. (2019). Single-cell transcriptomic analysis of the lateral hypothalamic area reveals molecularly distinct populations of inhibitory and excitatory neurons. Nature Neuroscience, 22(4), 642–656. https://doi.org/10.1038/s41593-019-0349-8 Mickelsen, L. E., Kolling, F. W., Chimileski, B. R., Fujita, A., Norris, C., Chen, K., Nelson, C. E., & Jackson, A. C. (2017). Neurochemical heterogeneity among lateral hypothalamic hypocretin/orexin and melanin-concentrating hormone neurons identified through single-cell gene expression analysis. ENeuro, 4(5). https://doi.org/10.1523/ENEURO.0013-17.2017 Mieda, M., & Yanagisawa, M. (2002). Sleep, feeding, and neuropeptides: Roles of orexins and orexin receptors. Current Opinion in Neurobiology, 12(3), 339–345. https://doi.org/10.1016/S0959-4388(02)00331-8 Mistlberger, R. E., Antle, M. C., Kilduff, T. S., & Jones, M. (2003). Food- and light- entrained circadian rhythms in rats with hypocretin-2-saporin ablations of the lateral hypothalamus. Brain Research, 980(2), 161–168. https://doi.org/10.1016/S0006- 8993(03)02755-0 Mogenson, G. J., Jones, D. L., & Yim, C. Y. (1980). From motivation to action: Functional interface between the limbic system and the motor system. Progress in Neurobiology, 14(2–3), 69–97. https://doi.org/10.1016/0301-0082(80)90018-0 138 Mogenson, G. J., Swanson, L. W., & Wu, M. (1983). Neural projections from nucleus accumbens to globus pallidus, substantia innominata, and lateral preoptic-lateral hypothalamic area: An anatomical and electrophysiological investigation in the rat. Journal of Neuroscience, 3(1), 189–202. https://doi.org/10.1523/jneurosci.03-01- 00189.1983 Montes, G. S., & Luque, E. H. (1988). Effects of Ovarian Steroids on vaginal smears in Rat. Acta Anat, 133, 192–199. Monti, J. M., Torterolo, P., & Lagos, P. (2013). Melanin-concentrating hormone control of sleep-wake behavior. Sleep Medicine Reviews, 17(4), 293–298. https://doi.org/10.1016/j.smrv.2012.10.002 Morin, L. P., & Fleming, A. S. (1978). Variation of food intake and body weight with estrous cycle, ovariectomy, and estradiol benzoate treatment in hamsters (Mesocricetus auratus). Journal of Comparative and Physiological Psychology, 92(1), 1–6. https://doi.org/10.1037/h0077435 Morita, T., Nishijima, T., & Tokura, H. (2005). Time sense for short intervals during the follicular and luteal phases of the menstrual cycle in humans. Physiology and Behavior, 85(2), 93–98. https://doi.org/10.1016/j.physbeh.2005.02.024 Morofushi, M., Shinohara, K., & Kimura, F. (2001). Menstrual and circadian variations in time perception in healthy women and women with premenstrual syndrome. Neuroscience Research, 41(4), 339–344. https://doi.org/10.1016/S0168- 0102(01)00290-5 Murray, J. F., Baker, B. I., Levy, A., & Wilson, C. A. (2000). The influence of gonadal steroids on pre-pro melanin-concentrating hormone mRNA in female rats. Journal of Neuroendocrinology, 12(1), 53–59. https://doi.org/10.1046/j.1365- 2826.2000.00425.x Muschamp, J. W., & Hull, E. M. (2007). Melanin concentrating hormone and estrogen receptor-α are coexstensive but not coexpressed in cells of male rat hypothalamus. Neuroscience Letters, 427(3), 123–126. https://doi.org/10.1016/j.neulet.2007.09.031 Mystkowski, P., Seeley, R. J., Hahn, T. M., Baskin, D. G., Havel, P. J., Matsumoto, A. M., Wilkinson, C. W., Peacock-kinzig, K., Blake, K. A., & Schwartz, M. W. (2000). Hypothalamic Melanin-Concentrating Hormone and Estrogen- Induced Weight Loss. 20(22), 8637–8642. Nieuwenhuys, R., Geeraedts, L. M. G., & Veening, J. G. (1982). The medial forebrain bundle of the rat. I. General introduction. In Journal of Comparative Neurology (Vol. 206, Issue 1). https://doi.org/10.1002/cne.902060106 139 Noble, E. E., Hahn, J. D., Konanur, V. R., Hsu, T. M., Page, S. J., Cortella, A. M., Liu, C. M., Song, M. Y., Suarez, A. N., Szujewski, C. C., Rider, D., Clarke, J. E., Darvas, M., Appleyard, S. M., & Kanoski, S. E. (2018). Control of Feeding Behavior by Cerebral Ventricular Volume Transmission of Melanin-Concentrating Hormone. Cell Metabolism, 28(1), 55-68.e7. https://doi.org/10.1016/j.cmet.2018.05.001 Noble, E. E., Wang, Z., Liu, C. M., Davis, E. A., Suarez, A. N., Stein, L. M., Tsan, L., Terrill, S. J., Hsu, T. M., Jung, A. H., Raycraft, L. M., Hahn, J. D., Darvas, M., Cortella, A. M., Schier, L. A., Johnson, A. W., Hayes, M. R., Holschneider, D. P., & Kanoski, S. E. (2019). Hypothalamus-hippocampus circuitry regulates impulsivity via melanin-concentrating hormone. Nature Communications, 10(1), 1–16. https://doi.org/10.1038/s41467-019-12895-y O’Connor, E. C., Kremer, Y., Lefort, S., Harada, M., Pascoli, V., Rohner, C., & Lüscher, C. (2015). Accumbal D1R Neurons Projecting to Lateral Hypothalamus Authorize Feeding. Neuron, 88(3), 553–564. https://doi.org/10.1016/j.neuron.2015.09.038 Olds, J., & Milner, P. (1954). Positive Reinforcement Produced By Electrical Stimulation of Septal Area and Other Regions of Rat Brain. Journal of Comparative and Physiological Psychology, 47(6), 419–427. https://doi.org/10.1037/h0058775 Panfil, K., Deavours, A., & Kirkpatrick, K. (2023). Effects of the estrous cycle on impulsive choice and interval timing in female rats. Hormones and Behavior, 149(January), 105315. https://doi.org/10.1016/j.yhbeh.2023.105315 Paxinos, G., & Watson, C. (1998). The rat brain in stereotaxic coordinates (4th edition) (Fourth). Academic Press. Petrovich, G. D. (2018). Lateral hypothalamus as a motivation-cognition interface in the control of feeding behavior. Frontiers in Systems Neuroscience, 12(April), 1–7. https://doi.org/10.3389/fnsys.2018.00014 Pfeffer, M., Plenzig, S., Gispert, S., Wada, K., Korf, H. W., & von Gall, C. (2012). Disturbed sleep/wake rhythms and neuronal cell loss in lateral hypothalamus and retina of mice with a spontaneous deletion in the ubiquitin carboxyl-terminal hydrolase L1 gene. Neurobiology of Aging, 33(2), 393–403. https://doi.org/10.1016/j.neurobiolaging.2010.02.019 Pissios, P., Bradley, R. L., & Maratos-Flier, E. (2006). Expanding the scales: The multiple roles of MCH in regulating energy balance and other biological functions. Endocrine Reviews, 27(6), 606–620. https://doi.org/10.1210/er.2006-0021 Pissios, P., Frank, L., Kennedy, A. R., Porter, D. R., Marino, F. E., Liu, F. F., Pothos, E. N., & Maratos-Flier, E. (2008). Dysregulation of the Mesolimbic Dopamine System and Reward in MCH-/- Mice. Biological Psychiatry, 64(3), 184–191. https://doi.org/10.1016/j.biopsych.2007.12.011 140 Pleil, K. E., Cordes, S., Meck, W. H., & Williams, L. C. (2011). Rapid and acute effects of estrogen on time perception in male and female rats. Frontiers in Integrative Neuroscience, 5(October), 1–15. https://doi.org/10.3389/fnint.2011.00063 Qu, D., Ludwig, D. S., Gammeltoft, S., Piper, M., Pelleymounter, M. a, Cullen, M. J., Mathes, W. F., Przypek, R., Kanarek, R., & Maratos-Flier, E. (1996). A role for melanin-concentrating hormone in the central regulation of feeding behaviour. Nature, 380(6571), 243–247. https://doi.org/10.1038/380243a0 Qualls-Creekmore, E., & Münzberg, H. (2018). Modulation of feeding and associated behaviors by lateral hypothalamic circuits. Endocrinology, 159(11), 3631–3642. https://doi.org/10.1210/en.2018-00449 Quigley, J. A., Logsdon, M. K., Graham, B. C., Beaudoin, K. G., & Becker, J. B. (2021). Activation of G protein-coupled estradiol receptor 1 in the dorsolateral striatum enhances motivation for cocaine and drug-induced reinstatement in female but not male rats. Biology of Sex Differences, 12(1), 1–11. https://doi.org/10.1186/s13293- 021-00389-w Rakitin, B. C., Penney, T. B., Gibbon, J., Malapani, C., Hinton, S. C., & Meck, W. H. (1998). Scalar expectancy theory and peak-interval timing in humans. Journal of Experimental Psychology: Animal Behavior Processes, 24(1), 15–33. https://doi.org/10.1037/0097-7403.24.1.15 Reppert, S. M., & Weaver, D. R. (2002). Coordination of circadian timing in mammals. Nature, 418(August), 935–941. Roberts, S. (1981). Isolation of an internal clock. Journal of Experimental Psychology: Animal Behavior Processes, 7(3), 242–268. https://doi.org/10.1037/0097- 7403.7.3.242 Roberts, S., & Church, R. M. (1978). Control of an internal clock. Journal of Experimental Psychology: Animal Behavior Processes, 4(4), 318–337. https://doi.org/10.1037//0097-7403.4.4.318 Robinson, T. E., Becker, J. B., & Ramirez, V. D. (1980). Sex differences in amphetamine-elicited rotational behavior and the lateralization of striatal dopamine in rats. Brain Research Bulletin, 5(5), 539–545. https://doi.org/10.1016/0361- 9230(80)90260-9 Ross, L., & Santi, A. (2000). The effects of estrogen on temporal and numerical processing in ovariectomized female rats. Psychobiology, 28(3), 394–405. https://doi.org/10.3758/BF03331997 Roth, B. L. (2016). DREADDs for Neuroscientists. Neuron, 89(4), 683–694. 141 https://doi.org/10.1016/j.neuron.2016.01.040 Saito, Y., Nothacker, H., Wang, Z., Lin, S. H. S., Leslie, F., & Civelli, O. (1999). Molecular characterization of the melanin-concentrating- hormone receptor. Letters to Nature, 400(July), 265–269. Sandstrom, N. J. (2007). Estradiol Modulation of the Speed of an Internal Clock. Behavioral Neuroscience, 121(2), 422–432. https://doi.org/10.1037/0735- 7044.121.2.422 Santollo, J., & Eckel, L. A. (2008). The orexigenic effect of melanin-concentrating hormone (MCH) is influenced by sex and stage of the estrous cycle. Physiology and Behavior, 93(4–5), 842–850. https://doi.org/10.1016/j.physbeh.2007.11.050 Santollo, J., & Eckel, L. A. (2013). Oestradiol Decreases Melanin-Concentrating Hormone (MCH) and MCH Receptor Expression in the Hypothalamus of Female Rats. Journal of Neuroendocrinology, 25(6), 570–579. https://doi.org/10.1111/jne.12032 Saper, C.B., Swanson, L. W., & Cowan, W. M. (1979). An Autoradiographic Study of the Efferent Connections of the Lateral Hypothalamic Area in the Rat. Journal of Comparative Neurology, 183(4), 689–706. Saper, Clifford B, Chou, T. C., & Elmquist, J. K. (2002). Review The Need to Feed: Homeostatic and Hedonic Control of Eating mechanisms that underlie regulation of feeding and make it such a rewarding experience (Figure 1). Neuron, 36, 199–211. https://ac-els-cdn-com.ep.fjernadgang.kb.dk/S0896627302009698/1-s2.0- S0896627302009698-main.pdf?_tid=fdb2abcc-804f-469e-a02d- 22863bcb04cd&acdnat=1536577440_61c910faaf6419f9c0a00c7a54d2ea35 Schneider, B. A. (1969). A Two-State Analysis of Fixed-Interval Responding in the Pigeion. Journal of Experimental Analysis of Behavior, 12, 677–687. Sherwood, A., Holland, P. C., Adamantidis, A., & Johnson, A. W. (2015). Deletion of Melanin Concentrating Hormone Receptor-1 disrupts overeating in the presence of food cues. Physiology and Behavior, 152, 402–407. https://doi.org/10.1016/j.physbeh.2015.05.037 Sherwood, A., Wosiski-Kuhn, M., Nguyen, T., Holland, P. C., Lakaye, B., Adamantidis, A., & Johnson, A. W. (2012). The role of melanin-concentrating hormone in conditioned reward learning. European Journal of Neuroscience, 36(8), 3126–3133. https://doi.org/10.1111/j.1460-9568.2012.08207.x Shimada, M., Tritos, N. A., Lowell, B. B., Flier, J. S., & Maratos-Flier, E. (1998). Mice lacking melanin-concentrating hormone receptor are hypophagic and lean. Nature, 396(December), 670–674. 142 Shughrue, P. J., Lane, M. V, & Merchenthaler, I. (1997). Comparative Distribution of Estrogen Receptor- a and - B mRNA in Rat CNS. The Journal of Comparative Neurology, 388, 507–525. Skinner, B. F. (1938). The Behavior of Organisms. https://doi.org/10.1037/13145-003 Smith, G. P. (2001). John Davis and the meanings of licking. Appetite, 36(1), 84–92. https://doi.org/10.1006/appe.2000.0371 Stelzer, G., Rosen, N., Plaschkes, I., Zimmerman, S., Twik, M., Fishilevich, S., Stein, T. I., Nudel, R., Lieder, I., Mazor, Y., Kaplan, S., Dahary, D., Warshawsky, D., Guan- golan, Y., Kohn, A., Rappaport, N., & Safran, M. (2016). The GeneCards Suite : From Gene Data Mining to Disease Genome Sequence Analyses. June, 1–33. https://doi.org/10.1002/cpbi.5 Stratford, T. R., & Kelley, A. E. (1997). GABA in the nucleus accumbens shell participates in the central regulation of feeding behavior. Journal of Neuroscience, 17(11), 4434–4440. https://doi.org/10.1523/jneurosci.17-11-04434.1997 Stuber, G. D., & Wise, R. A. (2016). Lateral hypothalamic circuits for feeding and reward. Nature Neuroscience, 19(2), 198–205. https://doi.org/10.1038/nn.4220 Subramanian, K. S., Lauer, L. T., Hayes, A. M. R., Décarie-spain, L., Mcburnett, K., Nourbash, A. C., Donohue, K. N., Kao, A. E., Bashaw, A. G., Burdakov, D., Noble, E. E., Schier, L. A., & Kanoski, S. E. (2023). Hypothalamic melanin-concentrating hormone neurons integrate food-motivated appetitive and consummatory processes in rats. 1–14. https://doi.org/10.1038/s41467-023-37344-9 Swanson, L. W., Sanchez-Watts, G., & Watts, A. G. (2005). Comparison of melanin- concentrating hormone and hypocretin/orexin mRNA expression patterns in a new parceling scheme of the lateral hypothalamic zone. Neuroscience Letters, 387(2), 80–84. https://doi.org/10.1016/j.neulet.2005.06.066 Tan, C. P., Sano, H., Iwaasa, H., Pan, J., Sailer, A. W., Hreniuk, D. L., Feighner, S. D., Palyha, O. C., Pong, S. S., Figueroa, D. J., Austin, C. P., Jiang, M. M., Yu, H., Ito, J., Ito, M., Ito, M., Guan, X. M., MacNeil, D. J., Kanatani, A., … Howard, A. D. (2002). Melanin-concentrating hormone receptor subtypes 1 and 2: Species- specific gene expression. Genomics, 79(6), 785–792. https://doi.org/10.1006/geno.2002.6771 ter Haar, M. B. (1972). Circadian and estrual rhythms in food intake in the rat. Hormones and Behavior, 3(3), 213–219. https://doi.org/10.1016/0018- 506X(72)90034-7 Terrill, S. J., Subramanian, K. S., Lan, R., Liu, C. M., Cortella, A. M., Noble, E. E., & 143 Kanoski, S. E. (2020). Nucleus accumbens melanin-concentrating hormone signaling promotes feeding in a sex-specific manner. Neuropharmacology, 178, 108270. https://doi.org/10.1016/j.neuropharm.2020.108270 Toran-Allerand, C. D. (2004). Minireview: A Plethora of Estrogen Receptors in the Brain: Where Will It End? Endocrinology, 145(3), 1069–1074. https://doi.org/10.1210/en.2003-1462 Toran-Allerand, C. D., Guan, X., MacLusky, N. J., Horvath, T. L., Diano, S., Singh, M., Connolly, E. S., Nethrapalli, I. S., & Tinnikov, A. A. (2002). ER-X: A novel, plasma membrane-associated, putative estrogen receptor that is regulated during development and after ischemic brain injury. Journal of Neuroscience, 22(19), 8391–8401. https://doi.org/10.1523/jneurosci.22-19-08391.2002 Turek, F. W., Joshu, C., Kohsaka, A., Lin, E., Ivanova, G., Mcdearmon, E., Laposky, A., Losee-olson, S., Easton, A., Jensen, D. R., Eckel, R. H., Takahashi, J. S., & Bass, J. (2005). Obesity and Metabolic Syndrome in Circadian Clock Mutant Mice. 308(May), 1043–1046. Valenstein, E. S., Cox, V. C., & Kakolewski, J. W. (1968). Modification of Motivated Behavior Elicited by Electrical Stimulation of the Hypothalamus. Science, 159(3819), 1119–1121. Varma, M., Chai, J. K., Meguid, M. M., Laviano, A., Gleason, J. R., Yang, Z. J., & Blaha, V. (1999). Effect of estradiol and progesterone on daily rhythm in food intake and feeding patterns in Fischer rats. Physiology and Behavior, 68(1–2), 99–107. https://doi.org/10.1016/S0031-9384(99)00152-3 Williams, C. L. (2011). Sex differences in counting and timing. Frontiers in Integrative Neuroscience, 5(DECEMBER), 1–4. https://doi.org/10.3389/fnint.2011.00088 Zamorano, F., Billeke, P., Hurtado, J. M., López, V., Carrasco, X., Ossandón, T., & Aboitiz, F. (2014). Temporal constraints of behavioral inhibition: Relevance of inter- stimulus interval in a go-nogo task. PLoS ONE, 9(1). https://doi.org/10.1371/journal.pone.0087232 Zhou, D., Shen, Z., Strack, A. M., Marsh, D. J., & Shearman, L. P. (2005). Enhanced running wheel activity of both Mch1r- and Pmch-deficient mice. Regulatory Peptides, 124(1–3), 53–63. https://doi.org/10.1016/j.regpep.2004.06.026 144 APPENDIX Chapter 2 Supplemental Materials Table S2.1 Results from the overall model examining estrous cycle within vehicle. F df p Pre-peak: Estrous 0.372 239.983 0.543 Time 65.264 1326.782 <.001 Time2 0.304 1938.129 0.581 Time3 3.499 2670.165 0.061 Estrous x time 0.004 1326.782 0.95 Estrous x time2 0.326 1938.129 0.568 Estrous x time3 0.045 2670.165 0.831 Post-peak: Estrous 0.089 174.311 0.766 Time 32.686 499.686 <.001 Time2 14.096 637.7 <.001 Time3 3.36 1241.077 0.067 Estrous x time 0.033 499.686 0.856 Estrous x time2 0.375 637.7 0.54 Estrous x time3 0.194 1241.077 0.66 145 Table S2.2 Results from the interaction model examining effects of time within vehicle treated rats tested in P/E and M/D. Regression standard F df p slope (b) error (se) Pre-peak P/E: Intercept 29.241 244.997 <.001 0.410928 0.075992 Time 30.25 1377.804 <.001 0.062273 0.011322 Time2 0.538 2065.577 0.463 0.000628 0.000856 Time3 1.106 2717.457 0.293 -0.000122 0.000116 Pre-peak M/D: Intercept 40.38 234.978 <.001 0.475981 0.074904 Time 35.333 1271.433 <.001 0.063246 0.01064 Time2 0 1764.583 0.988 -1.089E-05 0.00072 Time3 2.864 2582.134 0.091 -0.000154 9.0901E-05 Post-peak P/E: Intercept 5.864 174.361 0.016 0.156647 0.064689 Time 15.128 503.14 <.001 -0.022848 0.005874 Time2 9.642 633.292 0.002 0.00059 0.00019 Time3 0.955 1245.794 0.329 1.3229E-05 1.354E-05 Post-peak M/D: Intercept 8.095 174.261 0.005 0.183905 0.064639 Time 17.622 496.179 <.001 -0.024352 0.005801 Time2 4.881 642.053 0.028 0.000424 0.000192 Time3 2.626 1236.227 0.105 2.1591E-05 1.3323E-05 146 Table S2.3 Results from the overall models examining estrous cycle within vehicle at 5s intervals. F df p 5s Estrous 0.075 276.246 0.784 Time 52.608 1437.004 <.001 Time2 4.052 2679.487 0.044 Time3 9.858 2790.363 0.002 Estrous x time 0.003 1437.004 0.959 Estrous x time2 0.026 2679.487 0.871 Estrous x time3 0.021 2790.363 0.884 10s Estrous 0.061 242.194 0.806 Time 83.33 1338.594 <.001 Time2 0.193 2293.232 0.661 Time3 9.858 2790.363 0.002 Estrous x time 0.021 1338.594 0.884 Estrous x time2 0.02 2293.232 0.887 Estrous x time3 0.021 2790.363 0.884 15s Estrous 2.978 272.248 0.086 Time 44.83 1346.098 <.001 Time2 11.401 2776.309 <.001 Time3 7.24 2794.018 0.007 Estrous x time 0.357 1346.098 0.55 Estrous x time2 0.589 2776.309 0.443 Estrous x time3 0.297 2794.018 0.586 20s Estrous 2.111 243.658 0.148 Time 35.574 862.511 <.001 Time2 0.024 2084.304 0.876 Time3 0.025 2511.09 0.874 Estrous x time 0.968 862.511 0.325 Estrous x time2 1.87 2084.304 0.172 Estrous x time3 0.529 2511.09 0.467 25s Estrous 0.77 238.624 0.381 Time 64.073 1170.708 <.001 Time2 0.243 1827.767 0.622 Time3 1.826 2597.625 0.177 Estrous x time 0.467 1170.708 0.494 Estrous x time2 0.185 1827.767 0.667 Estrous x time3 0.388 2597.625 0.534 30s Estrous 0.466 242.314 0.496 Time 61.366 1344.277 <.001 147 Table S2.3 (cont’d) Time2 0.15 1777.786 0.698 Time3 3.536 2615.946 0.06 Estrous x time 0.022 1344.277 0.882 Estrous x time2 0.271 1777.786 0.602 Estrous x time3 0.052 2615.946 0.82 35s Estrous 0.466 242.314 0.496 Time 61.366 1344.277 <.001 Time2 0.15 1777.786 0.698 Time3 3.536 2615.946 0.06 Estrous x time 0.022 1344.277 0.882 Estrous x time2 0.271 1777.786 0.602 Estrous x time3 0.052 2615.946 0.82 40s Estrous 26.48 246.256 0.496 Time 27.575 1344.277 <.001 Time2 0.357 1777.786 0.698 Time3 1.099 2615.946 0.06 Estrous x time 38.429 1344.277 0.882 Estrous x time2 34.288 1777.786 0.602 Estrous x time3 0.011 2615.946 0.82 148 Table S2.4 Results from the interaction models examining estrous cycle within vehicle at 5s intervals. Regression standard F df p slope (b) error (se) 5s P/E: Intercept 4.574 282.849 0.033 0.17128 0.080086 Time 25.903 1434.992 <.001 0.053641 0.01054 Time2 2.125 2674.622 0.145 0.002801 0.001921 Time3 4.417 2784.029 0.036 -0.000198 9.4009E-05 5s M/D: Intercept 6.655 269.572 0.01 0.201969 0.078291 Time 26.708 1439.024 <.001 0.054401 0.010526 Time2 1.931 2685.523 0.165 0.002383 0.001715 Time3 5.759 2799.817 0.016 -0.00018 7.5037E-05 10s P/E: Intercept 41.436 242.554 <.001 0.484817 0.075316 Time 38.488 1344.466 <.001 0.066834 0.010773 Time2 0.043 2325.075 0.837 -0.000163 0.000788 Time3 4.417 2784.029 0.036 -0.000198 9.4009E-05 10s M/D: Intercept 46.133 241.835 <.001 0.511036 0.075239 Time 45.714 1331.197 <.001 0.064724 0.009573 Time2 0.175 2258.8 0.676 -0.000318 0.000761 Time3 5.759 2799.817 0.016 -0.00018 7.5037E-05 15s P/E: Intercept 90.402 270.813 <.001 0.789514 0.083037 Time 25.041 1518.543 <.001 0.050189 0.01003 Time2 6.273 2779.731 0.012 -0.003116 0.001244 Time3 4.309 2803.359 0.038 -0.000196 9.4613E-05 15s M/D: Intercept 69.405 274.174 <.001 0.599857 0.072003 Time 19.823 1175.501 <.001 0.04196 0.009424 Time2 5.392 2768.679 0.02 -0.001962 0.000845 Time3 2.933 2778.355 0.087 -0.00013 7.6059E-05 20s P/E: Intercept 22.163 241.854 <.001 0.396332 0.084187 Time 23.088 889.565 <.001 0.042124 0.008767 Time2 1.207 2111.114 0.272 0.001151 0.001048 Time3 0.152 2497.711 0.697 2.4309E-05 6.2355E-05 20s M/D: Intercept 46.018 245.496 <.001 0.569005 0.083878 Time 12.994 834.079 <.001 0.030195 0.008376 Time2 0.707 2059.542 0.401 -0.000916 0.001089 Time3 0.419 2526.084 0.518 -3.789E-05 5.8558E-05 25s P/E: Intercept 25.567 242.948 <.001 0.393172 0.077758 Time 27.216 1416.741 <.001 0.060161 0.011532 Time2 0.358 1936.315 0.55 0.000518 0.000866 Time3 1.078 2665.554 0.299 -0.000124 0.00012 25s M/D: Intercept 41.353 234.214 <.001 0.488577 0.075977 Time 43.699 778.542 <.001 0.050696 0.007669 Time2 0.002 1675.061 0.961 3.5218E-05 0.000713 149 Table S2.4 (cont’d) Time3 1.378 1806.83 0.241 -4.582E-05 3.9037E-05 30s P/E: Intercept 26.48 246.256 <.001 0.430583 0.053715 Time 27.575 1396.705 <.001 -0.036651 0.053715 Time2 0.357 1911.721 0.55 0.06142 0.007841 Time3 1.099 2650.47 0.295 0.00022 0.000567 30s M/D: Intercept 44.883 201.424 <.001 0.458512 0.06844 Time 28.515 421.285 <.001 -0.025844 0.00484 Time2 1.433 1373.365 0.231 -0.000377 0.000315 Time3 5.167 1388.455 0.023 2.6251E-05 1.1548E-05 35s P/E: Intercept 21.211 181.636 <.001 0.300112 0.065163 Time 28.374 469.314 <.001 -0.029491 0.005536 Time2 0.668 908.95 0.414 0.000178 0.000217 Time3 3.895 1317.578 0.049 2.4911E-05 1.2622E-05 35s M/D: Intercept 25.441 178.927 <.001 0.323158 0.064069 Time 25.65 473.044 <.001 -0.027641 0.005458 Time2 0.008 982.323 0.929 1.7238E-05 0.000195 Time3 5.167 1388.455 0.023 2.6251E-05 1.1548E-05 40s P/E: Intercept 26.48 246.256 <.001 0.393932 0.076553 Time 27.575 1396.705 <.001 0.060252 0.011474 Time2 0.357 1911.721 0.55 0.000515 0.000862 Time3 1.099 2650.47 0.295 -0.000125 0.000119 40s M/D: Intercept 8.598 176.354 0.004 0.188667 0.064342 Time 22.18 488.893 <.001 -0.0255 0.005414 Time2 4.576 719.182 0.033 0.000411 0.000192 Time3 5.167 1388.455 0.023 2.6251E-05 1.1548E-05 150 Table S2.5 Results from the overall model examining drug treatment within P/E. F df p Pre-peak: Drug 0.047 232.575 0.828 Time 59.179 1365.987 <.001 Time2 1.688 2041.091 0.194 Time3 2.223 2713.98 0.136 Drug x time 0.075 1365.987 0.785 Drug x time2 0.038 2041.091 0.845 Drug x time3 0.03 2713.98 0.863 Post-peak: Drug 25.845 536.238 <.001 Time 14.223 688.659 <.001 Time2 2.251 1379.884 0.134 Time3 0.043 170.805 0.837 Drug x time 0.046 536.238 0.831 Drug x time2 0.103 688.659 0.749 Drug x time3 0.027 1379.884 0.87 151 Table S2.6 Results from the interaction model examining drug treatment within P/E. Regression standard F df p slope (b) error (se) Pre-peak P/E: Intercept 26.474 237.682 <.001 0.410428 0.079768 Table S2.6 (cont’d) Time 29.161 1450.839 <.001 0.062061 0.011493 Time2 0.515 2153.858 0.473 0.00062 0.000863 Time3 1.07 2754.569 0.301 -0.000121 0.000117 Pre-peak M/D: Intercept 24.178 227.471 <.001 0.386111 0.078523 Time 30.183 1271.489 <.001 0.057801 0.010521 Time2 1.368 1880.777 0.242 0.000839 0.000717 Time3 1.23 2626.688 0.268 -9.571E-05 8.631E-05 Post-peak P/E: Intercept 5.325 169.781 0.022 0.158222 0.068564 Time 14.321 527.6 <.001 -0.022939 0.006062 Time2 8.878 672.116 0.003 0.000581 0.000195 Time3 0.961 1350.168 0.327 1.349E-05 1.3761E-05 Post-peak M/D: Intercept 4.007 171.824 0.047 0.138164 0.06902 Time 11.625 544.719 <.001 -0.021089 0.006185 Time2 5.633 703.883 0.018 0.00049 0.000207 Time3 1.294 1406.277 0.256 1.6784E-05 1.4757E-05 152 Table S2.7 Results from the overall models examining drug (VEH, CNO) during P/E at 5s intervals. F df p 5s Drug 0.028 264.307 0.867 Time 42.406 1497.739 <.001 Time2 3.528 2696.84 0.06 Time3 7.052 2804.575 0.008 Drug x time 0.326 1497.739 0.568 Drug x time2 0.093 2696.84 0.761 Drug x time3 0.541 2804.575 0.462 10s Drug 0.409 235.14 0.523 Time 73.886 1318.461 <.001 Time2 0.02 2365.82 0.888 Time3 7.052 2804.575 0.008 Drug x time 0.486 1318.461 0.486 Drug x time2 0.203 2365.82 0.653 Drug x time3 0.541 2804.575 0.462 15s Drug 3.986 260.022 0.047 Time 47.484 1397.062 <.001 Time2 8.029 2774.13 0.005 Time3 6.15 2801.157 0.013 Drug x time 0.087 1397.062 0.768 Drug x time2 2.232 2774.13 0.135 Drug x time3 0.798 2801.157 0.372 20s Drug 0.206 236.562 0.65 Time 45.14 904.926 <.001 Time2 1.156 2142.29 0.282 Time3 0.033 2544.887 0.855 Drug x time 0.051 904.926 0.821 Drug x time2 0.323 2142.29 0.57 Drug x time3 0.184 2544.887 0.668 25s Drug 0.011 231.584 0.916 Time 58.42 1221.622 <.001 Time2 0.808 1933.017 0.369 Time3 1.367 2649.329 0.242 Drug x time 0.764 1221.622 0.382 Drug x time2 0.001 1933.017 0.979 Drug x time3 0.555 2649.329 0.456 30s Drug 0.029 233.978 0.866 Time 54.628 1386.611 <.001 Time2 1.311 1892.158 0.252 153 Table S2.7 (cont’d) Time3 2.187 2663.537 0.139 Drug x time 0.047 1386.611 0.828 Drug x time2 0.05 1892.158 0.824 Drug x time3 0.024 2663.537 0.876 35s Drug 0.029 233.978 0.866 Time 54.628 1386.611 <.001 Time2 1.311 1892.158 0.252 Time3 2.187 2663.537 0.139 Drug x time 0.047 1386.611 0.828 Drug x time2 0.05 1892.158 0.824 Drug x time3 0.024 2663.537 0.876 40s Drug 23.628 238.903 0.866 Time 26.484 1386.611 <.001 Time2 0.361 1892.158 0.252 Time3 1.033 2663.537 0.139 Drug x time 22.088 1386.611 0.828 Drug x time2 28.365 1892.158 0.824 Drug x time3 1.139 2663.537 0.876 154 Table S2.8 Results from the interaction model examining the effects of drug (VEH, CNO) during P/E at 5s intervals. Regression standard F df p slope (b) error (se) 5s VEH: Intercept 4.25 272.913 0.04 0.17285 0.083848 Time 25.045 1485.706 <.001 0.053567 0.010704 Time2 2.055 2724.516 0.152 0.002767 0.00193 Time3 4.317 2811.623 0.038 -0.000196 9.4318E-05 5s CNO: Intercept 3.49 255.719 0.063 0.153154 0.081982 Time 17.676 1509.901 <.001 0.044934 0.010688 Time2 1.473 2655.395 0.225 0.001994 0.001643 Time3 2.759 2789.303 0.097 -0.000111 6.6821E-05 10s VEH: Intercept 37.592 234.913 <.001 0.485368 0.079163 Time 36.926 1412.648 <.001 0.066541 0.01095 Time2 0.047 2405.077 0.828 -0.000172 0.000794 Time3 4.317 2811.623 0.038 -0.000196 9.4318E-05 10s CNO: Intercept 27.365 235.369 <.001 0.413813 0.079106 Time 37.552 1197.173 <.001 0.056555 0.009229 Time2 0.177 2324.852 0.674 0.00033 0.000783 Time3 2.759 2789.303 0.097 -0.000111 6.6821E-05 15s VEH: Intercept 86.278 266.525 <.001 0.789023 0.084945 Time 24.597 1553.309 <.001 0.050051 0.010092 Time2 6.238 2791.729 0.013 -0.003108 0.001245 Time3 4.271 2816.533 0.039 -0.000196 9.4628E-05 15s CNO: Intercept 56.229 252.022 <.001 0.562726 0.075044 Time 22.889 1243.134 <.001 0.045948 0.009604 Time2 1.798 2711.189 0.18 -0.000962 0.000718 Time3 1.884 2764.624 0.17 -9.198E-05 6.7007E-05 20s VEH: Intercept 20.569 235.503 <.001 0.393868 0.086844 Time 22.801 919.974 <.001 0.042411 0.008882 Time2 1.228 2189.665 0.268 0.001166 0.001052 Time3 0.166 2550.355 0.684 2.5425E-05 6.2496E-05 20s CNO: Intercept 27.245 237.646 <.001 0.449436 0.086104 Time 22.366 888.405 <.001 0.039646 0.008383 Time2 0.143 2084.395 0.706 0.000359 0.000951 Time3 0.035 2537.763 0.852 -1.025E-05 5.4917E-05 25s VEH: Intercept 22.91 236.113 <.001 0 0 Time 26.162 1493.523 <.001 0.390766 0.081639 Time2 0.361 2026.694 0.548 0.059899 0.011711 Time3 1.015 2712.69 0.314 0.000526 0.000875 25s CNO: Intercept 25.472 226.98 <.001 -0.000121 0.00012 Time 37.343 793.768 <.001 0.402856 0.079821 Time2 0.467 1800.66 0.495 0.047608 0.007791 155 Table S2.8 (cont’d) Time3 0.459 1866.364 0.498 0.000496 0.000726 30s VEH: Intercept 23.628 238.903 <.001 0.381842 0.056477 Time 26.484 1477.058 <.001 0.009557 0.056477 Time2 0.361 2006.41 0.548 0.058256 0.007882 Time3 1.033 2702.394 0.309 0.00065 0.000568 30s CNO: Intercept 31.07 201.025 <.001 0.411853 0.073888 Time 23.141 453.943 <.001 -0.024265 0.005044 Time2 0.851 1499.161 0.357 -0.000355 0.000385 Time3 3.756 1556.809 0.053 2.548E-05 1.3147E-05 35s VEH: Intercept 19.306 176.092 <.001 0.303885 0.069161 Time 26.58 491.347 <.001 -0.029539 0.00573 Time2 0.514 967.362 0.474 0.000159 0.000222 Time3 3.906 1426.449 0.048 2.535E-05 1.2827E-05 35s CNO: Intercept 16.841 179.341 <.001 0.284831 0.069407 Time 20.503 502.759 <.001 -0.025906 0.005721 Time2 0.013 1138.357 0.909 2.6976E-05 0.000235 Time3 3.756 1556.809 0.053 2.548E-05 1.3147E-05 40s VEH: Intercept 23.628 238.903 <.001 0.391399 0.080521 Time 26.484 1477.058 <.001 0.059964 0.011652 Time2 0.361 2006.41 0.548 0.000524 0.000872 Time3 1.033 2702.394 0.309 -0.000122 0.00012 40s CNO: Intercept 5.379 171.928 0.022 0.159158 0.068626 Time 16.277 539.627 <.001 -0.023726 0.005881 Time2 4.183 771.938 0.041 0.000409 0.0002 Time3 3.756 1556.809 0.053 2.548E-05 1.3147E-05 156 Table S2.9 Results from the overall model examining drug treatment within M/D. F df p Pre-peak: Drug 0.245 236.66 0.621 Time 70.478 1263.682 <.001 Time2 0.257 1744.578 0.612 Time3 4.72 2554.977 0.03 Drug x time 0.003 1263.682 0.959 Drug x time2 0.276 1744.578 0.6 Drug x time3 0.04 2554.977 0.841 Post-peak: Drug 0.134 174.29 0.715 Time 42.529 483.83 <.001 Time2 8.639 621.096 0.003 Time3 8.033 1189.501 0.005 Drug x time 0.208 483.83 0.649 Drug x time2 0.093 621.096 0.76 Drug x time3 0.209 1189.501 0.647 157 Table S2.10 Results from the interaction model examining drug treatment within M/D. Regression standard F df p slope (b) error (se) Pre-peak P/E: Intercept 41.713 236.243 <.001 0.476073 0.073712 Time 35.954 1253.967 <.001 0.063287 0.010555 Time2 0 1742.861 0.99 -9.397E-06 0.000715 Time3 2.906 2566.263 0.088 -0.000154 9.0431E-05 Pre-peak M/D: Intercept 33.044 237.077 <.001 0.424479 0.073843 Time 34.539 1273.325 <.001 0.062524 0.010639 Time2 0.51 1746.152 0.475 0.000534 0.000747 Time3 1.886 2544.245 0.17 -0.000128 9.3263E-05 Post-peak P/E: Intercept 8.359 175.169 0.004 0.183521 0.063476 Time 18.027 490.337 <.001 -0.024339 0.005732 Time2 5.037 632.904 0.025 0.000426 0.00019 Time3 2.66 1212.484 0.103 2.1551E-05 1.3215E-05 Post-peak M/D: Intercept 11.749 173.407 <.001 0.216243 0.063087 Time 24.849 477.177 <.001 -0.027995 0.005616 Time2 3.631 608.518 0.057 0.000346 0.000182 Time3 5.776 1164.074 0.016 2.9845E-05 1.2418E-05 158 Table S2.11 Results from the overall models examining drug (VEH, CNO) during M/D at 5s intervals. F df p 5s Drug 0.011 272.127 0.916 Time 52.751 1439.288 <.001 Time2 2.243 2686.017 0.134 Time3 7.21 2797.677 0.007 Drug x time 0.007 1439.288 0.935 Drug x time2 0.194 2686.017 0.659 Drug x time3 0.41 2797.677 0.522 10s Drug 0.015 242.855 0.903 Time 79.58 1368.551 <.001 Time2 0.398 2268.58 0.528 Time3 7.21 2797.677 0.007 Drug x time 0.249 1368.551 0.618 Drug x time2 0.002 2268.58 0.967 Drug x time3 0.41 2797.677 0.522 15s Drug 0.003 267.535 0.955 Time 42.39 1218.29 <.001 Time2 7.58 2775.304 0.006 Time3 4.652 2778.648 0.031 Drug x time 0.108 1218.29 0.742 Drug x time2 0.205 2775.304 0.651 Drug x time3 0.038 2778.648 0.846 20s Drug 0.984 245.306 0.322 Time 41.781 842.279 <.001 Time2 0 2076.088 0.994 Time3 0.217 2520.711 0.642 Drug x time 2.055 842.279 0.152 Drug x time2 1.413 2076.088 0.235 Drug x time3 0.187 2520.711 0.666 25s Drug 0.544 237.419 0.462 Time 73.132 1081.248 <.001 Time2 0.339 1627.39 0.56 Time3 2.723 2435.242 0.099 Drug x time 0.68 1081.248 0.41 Drug x time2 0.258 1627.39 0.611 Drug x time3 0.576 2435.242 0.448 30s Drug 0.289 241.027 0.591 Time 68.489 1271.952 <.001 Time2 0.214 1581.628 0.643 159 Table S2.11 (cont’d) Time3 4.605 2522.825 0.032 Drug x time 0.005 1271.952 0.944 Drug x time2 0.369 1581.628 0.544 Drug x time3 0.069 2522.825 0.793 35s Drug 0.289 241.027 0.591 Time 68.489 1271.952 <.001 Time2 0.214 1581.628 0.643 Time3 4.605 2522.825 0.032 Drug x time 0.005 1271.952 0.944 Drug x time2 0.369 1581.628 0.544 Drug x time3 0.069 2522.825 0.793 40s Drug 40.132 240.594 0.591 Time 35.059 1271.952 <.001 Time2 0.011 1581.628 0.643 Time3 2.992 2522.825 0.032 Drug x time 30.971 1271.952 0.944 Drug x time2 33.446 1581.628 0.544 Drug x time3 0.548 2522.825 0.793 160 Table S2.12 Results from the interaction model examining the effects of drug (VEH, CNO) during M/D at 5s intervals. Regression standard F df p slope (b) error (se) 5s VEH: Intercept 6.763 270.26 0.01 0.201852 0.077616 Time 26.978 1431.826 <.001 0.054403 0.010474 Time2 1.947 2679.35 0.163 0.002384 0.001709 Time3 5.803 2796.342 0.016 -0.00018 7.4773E-05 5s CNO: Intercept 7.489 273.993 0.007 0.213534 0.078029 Time 25.78 1446.785 <.001 0.053192 0.010476 Time2 0.539 2692.143 0.463 0.0013 0.00177 Time3 1.997 2798.881 0.158 -0.000111 7.8399E-05 10s VEH: Intercept 46.957 242.333 <.001 0.510958 0.074565 Time 46.202 1323.282 <.001 0.064736 0.009524 Time2 0.176 2248.267 0.675 -0.000318 0.000757 Time3 5.803 2796.342 0.016 -0.00018 7.4773E-05 10s CNO: Intercept 44.46 243.377 <.001 0.498137 0.074708 Time 34.107 1411.791 <.001 0.05788 0.009911 Time2 0.223 2288.355 0.637 -0.000362 0.000766 Time3 1.997 2798.881 0.158 -0.000111 7.8399E-05 15s VEH: Intercept 65.951 262.798 <.001 0.597001 0.073513 Time 19.594 1181.061 <.001 0.042038 0.009497 Time2 5.33 2778.188 0.021 -0.001952 0.000845 Time3 2.919 2788.104 0.088 -0.00013 7.6109E-05 15s CNO: Intercept 65.765 272.288 <.001 0.60286 0.074339 Time 22.824 1254.954 <.001 0.046511 0.009735 Time2 2.555 2772.581 0.11 -0.001401 0.000876 Time3 1.829 2769.721 0.176 -0.000109 8.0282E-05 20s VEH: Intercept 47.198 248.056 <.001 0.569551 0.082903 Time 13.158 830.66 <.001 0.030154 0.008313 Time2 0.721 2043.716 0.396 -0.00092 0.001083 Time3 0.428 2512.435 0.513 -3.813E-05 5.8282E-05 20s CNO: Intercept 30.093 242.584 <.001 0.453438 0.082657 Time 30.029 853.225 <.001 0.047341 0.008639 Time2 0.692 2107.97 0.406 0.000909 0.001093 Time3 0.001 2528.022 0.982 -1.42E-06 6.1797E-05 25s VEH: Intercept 43.087 236.642 <.001 0.488805 0.074467 Time 44.774 769.515 <.001 0.050692 0.007576 Time2 0.003 1647.817 0.956 3.8773E-05 0.000707 Time3 1.414 1782.856 0.235 -4.603E-05 3.8717E-05 25s CNO: Intercept 29.862 238.184 <.001 0.41077 0.075169 Time 32.971 1294.461 <.001 0.061511 0.010712 Time2 0.55 1610.184 0.458 0.000569 0.000767 161 Table S2.12 (cont’d) Time3 1.688 2533.933 0.194 -0.000124 9.5777E-05 30s VEH: Intercept 40.132 240.594 <.001 0.439596 0.052244 Time 35.059 1266.146 <.001 0.028109 0.052244 Time2 0.011 1576.641 0.917 0.062125 0.007507 Time3 2.992 2532.835 0.084 0.000244 0.000528 30s CNO: Intercept 58.358 200.336 <.001 0.505951 0.066231 Time 34.196 405.872 <.001 -0.027563 0.004714 Time2 5.194 1262.226 0.023 -0.000651 0.000286 Time3 11.019 1316.905 <.001 3.6055E-05 1.0861E-05 35s VEH: Intercept 26.542 180.851 <.001 0.322462 0.06259 Time 26.539 467.307 <.001 -0.027681 0.005373 Time2 0.012 966.937 0.914 2.0772E-05 0.000192 Time3 5.272 1358.022 0.022 2.6259E-05 1.1437E-05 35s CNO: Intercept 32.896 178.094 <.001 0.356366 0.062133 Time 34.946 464.813 <.001 -0.031369 0.005306 Time2 0.38 866.968 0.538 -0.00011 0.000179 Time3 11.019 1316.905 <.001 3.6055E-05 1.0861E-05 40s VEH: Intercept 40.132 240.594 <.001 0.467705 0.073829 Time 35.059 1266.146 <.001 0.062653 0.010581 Time2 0.011 1576.641 0.917 -7.605E-05 0.00073 Time3 2.992 2532.835 0.084 -0.00016 9.2703E-05 40s CNO: Intercept 10.275 177.588 0.002 0.201271 0.062791 Time 32.748 472.845 <.001 -0.029767 0.005202 Time2 5.235 693.796 0.022 0.000431 0.000188 Time3 11.019 1316.905 <.001 3.6055E-05 1.0861E-05 162 Chapter 3 Supplemental Materials Table S3.1 Results from the overall model examining estrous cycle within vehicle. F df p Pre-peak: Estrous 0.052 312.224 0.82 Time 64.172 1864.046 <.001 Time2 0.811 2685.781 0.368 Time3 0.497 3596.903 0.481 Estrous x time 0.586 1864.046 0.444 Estrous x time2 0.039 2685.781 0.844 Estrous x time3 0.614 3596.903 0.433 Post-peak: Estrous 1.069 226.476 0.302 Time 61.607 695.345 <.001 Time2 4.325 897.092 0.038 Time3 19.627 1787.695 <.001 Estrous x time 1.111 695.345 0.292 Estrous x time2 1.494 897.092 0.222 Estrous x time3 2.855 1787.695 0.091 163 Table S3.2 Results from the interaction model examining effects of time within vehicle treated rats tested in P/E and M/D. Regression standard F df p slope (b) error (se) Pre-peak P/E: Intercept 40.047 314.713 <.001 0.436405 0.068961 Time 36.805 1917.777 <.001 0.059504 0.009808 Time2 0.231 2748.883 0.631 0.000348 0.000724 Time3 0.977 3620.92 0.323 -9.412E-05 9.5202E-05 Pre-peak M/D: Intercept 36.539 309.734 <.001 0.414294 0.068538 Time 27.523 1806.867 <.001 0.049126 0.009364 Time2 0.648 2613.667 0.421 0.000543 0.000675 Time3 0.004 3563.911 0.952 4.9866E-06 8.329E-05 Post-peak P/E: Intercept 9.45 225.855 0.002 0.182097 0.059237 Time 23.254 690.099 <.001 -0.024849 0.005153 Time2 5.59 887.017 0.018 0.00039 0.000165 Time3 3.864 1764.49 0.049 2.2316E-05 1.1352E-05 Post-peak M/D: Intercept 20.48 227.096 <.001 0.268842 0.059406 Time 39.348 700.57 <.001 -0.032558 0.00519 Time2 0.359 906.822 0.549 0.000101 0.000169 Time3 18.213 1809.924 <.001 4.9833E-05 1.1677E-05 164 Table S3.3 Results from the overall models examining estrous cycle within vehicle at 5s intervals. F df p 5s Estrous 0.014 350.773 0.906 Time 66.102 2005.66 <.001 Time2 0.229 3666.405 0.632 Time3 2.141 3791.698 0.144 Estrous x time 0.138 2005.66 0.71 Estrous x time2 0.47 3666.405 0.493 Estrous x time3 1.146 3791.698 0.284 10s Estrous 0.141 315.39 0.708 Time 76.278 1923.956 <.001 Time2 1.357 3169.915 0.244 Time3 2.141 3791.698 0.144 Estrous x time 0.942 1923.956 0.332 Estrous x time2 0.009 3169.915 0.925 Estrous x time3 1.146 3791.698 0.284 15s Estrous 0.737 350.252 0.391 Time 51.601 1703.555 <.001 Time2 2.65 3742.014 0.104 Time3 1.316 3773.399 0.251 Estrous x time 0.784 1703.555 0.376 Estrous x time2 0.136 3742.014 0.713 Estrous x time3 0.005 3773.399 0.944 20s Estrous 0.115 317.171 0.734 Time 72.56 1228.604 <.001 Time2 1.573 2947.16 0.21 Time3 0.001 3411.074 0.98 Estrous x time 0.001 1228.604 0.97 Estrous x time2 0.037 2947.16 0.848 Estrous x time3 0.036 3411.074 0.849 25s Estrous 0.056 314.772 0.812 Time 75.904 1623.396 <.001 Time2 2.696 2504.759 0.101 Time3 0.494 3455.586 0.482 Estrous x time 0.632 1623.396 0.427 Estrous x time2 0.853 2504.759 0.356 Estrous x time3 0.849 3455.586 0.357 30s Estrous 0.066 315.963 0.798 Time 59.668 1879.115 <.001 Time2 1.367 2481.12 0.242 165 Table S3.3 (cont’d) Time3 0.18 3556.533 0.672 Estrous x time 0.526 1879.115 0.469 Estrous x time2 0.203 2481.12 0.652 Estrous x time3 0.808 3556.533 0.369 35s Estrous 0.066 315.963 0.798 Time 59.668 1879.115 <.001 Time2 1.367 2481.12 0.242 Time3 0.18 3556.533 0.672 Estrous x time 0.526 1879.115 0.469 Estrous x time2 0.203 2481.12 0.652 Estrous x time3 0.808 3556.533 0.369 40s Estrous 36.236 318.673 0.798 Time 33.984 1879.115 <.001 Time2 0.241 2481.12 0.242 Time3 0.768 3556.533 0.672 Estrous x time 32.501 1879.115 0.469 Estrous x time2 25.798 2481.12 0.652 Estrous x time3 1.412 3556.533 0.369 166 Table S3.4 Results from the interaction models examining estrous cycle within vehicle at 5s intervals. Regression standard F df p slope (b) error (se) 5s P/E: Intercept 9.886 358.985 0.002 0.225773 0.071805 Time 36.127 1990.93 <.001 0.055749 0.009275 Time2 0.611 3686.695 0.435 0.001254 0.001605 Time3 2.631 3788.321 0.105 -0.000124 7.6198E-05 5s M/D: Intercept 11.295 342.587 <.001 0.237632 0.070708 Time 30.111 2020.504 <.001 0.050879 0.009272 Time2 0.024 3639.375 0.877 -0.000223 0.001436 Time3 0.099 3796.651 0.753 -1.915E-05 6.0933E-05 10s P/E: Intercept 57.587 315.049 <.001 0.520415 0.068578 Time 41.966 2005.676 <.001 0.059018 0.00911 Time2 0.792 3195.418 0.373 -0.0006 0.000674 Time3 2.631 3788.321 0.105 -0.000124 7.6198E-05 10s M/D: Intercept 49.814 315.731 <.001 0.484057 0.068584 Time 34.322 1824.33 <.001 0.047212 0.008059 Time2 0.573 3144.216 0.449 -0.00051 0.000674 Time3 0.099 3796.651 0.753 -1.915E-05 6.0933E-05 15s P/E: Intercept 72.05 368.058 <.001 0.576479 0.067915 Time 19.978 1696.239 <.001 0.037746 0.008445 Time2 1.548 3734.019 0.214 -0.001185 0.000953 Time3 0.477 3782.444 0.49 -5.373E-05 7.7773E-05 15s M/D: Intercept 97.987 332.97 <.001 0.658089 0.066481 Time 32.317 1710.798 <.001 0.048362 0.008507 Time2 1.113 3755.57 0.292 -0.000748 0.000709 Time3 0.945 3757.677 0.331 -6.067E-05 6.242E-05 20s P/E: Intercept 37.346 314.217 <.001 0.447649 0.073251 Time 35.028 1235.916 <.001 0.045422 0.007675 Time2 0.549 2972.221 0.459 0.000643 0.000868 Time3 0.013 3412.389 0.909 -6.005E-06 5.2313E-05 20s M/D: Intercept 43.503 320.169 <.001 0.482825 0.073203 Time 37.602 1220.953 <.001 0.045828 0.007474 Time2 1.077 2920.538 0.299 0.000875 0.000843 Time3 0.024 3409.637 0.876 7.7887E-06 5.0085E-05 25s P/E: Intercept 35.708 316.637 <.001 0.417402 0.069851 Time 33.774 1956.412 <.001 0.05753 0.009899 Time2 0.241 2543.598 0.623 0.000362 0.000737 Time3 0.759 3579.588 0.384 -8.529E-05 9.7869E-05 25s M/D: Intercept 32.428 312.887 <.001 0.394041 0.069196 Time 47.351 1131.302 <.001 0.047912 0.006963 Time2 3.532 2460.631 0.06 0.001294 0.000689 167 Table S3.4 (cont’d) Time3 0.091 2430.566 0.763 1.1477E-05 3.8086E-05 30s P/E: Intercept 36.236 318.673 <.001 0.405134 0.048879 Time 33.984 1948.092 <.001 0.012537 0.048879 Time2 0.241 2531.879 0.623 0.05262 0.006812 Time3 0.768 3572.948 0.381 0.000588 0.000503 30s M/D: Intercept 83.732 260.712 <.001 0.570606 0.062358 Time 27.612 578.741 <.001 -0.022652 0.004311 Time2 25.909 1979.757 <.001 -0.001422 0.000279 Time3 29.08 2055.239 <.001 5.4712E-05 1.0146E-05 35s P/E: Intercept 29.31 231.372 <.001 0.318903 0.058904 Time 35.058 655.659 <.001 -0.028859 0.004874 Time2 0.07 1269.628 0.792 -4.535E-05 0.000172 Time3 8.92 1930.204 0.003 3.0415E-05 1.0184E-05 35s M/D: Intercept 53.012 232.633 <.001 0.428638 0.058871 Time 45.882 663.752 <.001 -0.032768 0.004838 Time2 11.892 1383.642 <.001 -0.000601 0.000174 Time3 29.08 2055.239 <.001 5.4712E-05 1.0146E-05 40s P/E: Intercept 36.236 318.673 <.001 0.417671 0.069385 Time 33.984 1948.092 <.001 0.057558 0.009874 Time2 0.241 2531.879 0.623 0.000361 0.000736 Time3 0.768 3572.948 0.381 -8.567E-05 9.7738E-05 40s M/D: Intercept 18.797 229.701 <.001 0.256609 0.059187 Time 51.932 691.095 <.001 -0.034677 0.004812 Time2 1.656 1014.523 0.198 0.000219 0.000171 Time3 29.08 2055.239 <.001 5.4712E-05 1.0146E-05 168 Table S3.5 Results from the overall model examining drug treatment within P/E. F df p Pre-peak: Drug 0.015 318.271 0.901 Time 69.944 1815.933 <.001 Time2 1.046 2659.667 0.307 Time3 1.361 3591.193 0.244 Drug x time 0.158 1815.933 0.691 Drug x time2 0.091 2659.667 0.763 Drug x time3 0.111 3591.193 0.739 Post-peak: Drug 42.308 703.798 <.001 Time 13.471 901.344 <.001 Time2 4.233 1787.802 0.04 Time3 0.076 232.687 0.783 Drug x time 0.031 703.798 0.861 Drug x time2 0.193 901.344 0.661 Drug x time3 0.244 1787.802 0.622 169 Table S3.6 Results from the interaction model examining drug treatment within P/E. Regression standard F df p slope (b) error (se) Pre-peak P/E: Intercept 41.493 320.893 <.001 0.436706 0.067796 Time 36.973 1874.617 <.001 0.059616 0.009804 Time2 0.234 2695.946 0.628 0.000351 0.000726 Time3 0.989 3593.605 0.32 -9.517E-05 9.5689E-05 Pre-peak M/D: Intercept 39.857 315.642 <.001 0.424872 0.067299 Time 32.978 1754.791 <.001 0.054214 0.009441 Time2 0.985 2614.687 0.321 0.000645 0.00065 Time3 0.402 3587.998 0.526 -5.283E-05 8.3331E-05 Post-peak P/E: Intercept 9.788 229.979 0.002 0.181344 0.057965 Time 23.566 677.152 <.001 -0.024802 0.005109 Time2 5.774 864.892 0.016 0.000394 0.000164 Time3 3.828 1701.462 0.051 2.2186E-05 1.134E-05 Post-peak M/D: Intercept 12.059 235.367 <.001 0.204062 0.058764 Time 19.016 728.996 <.001 -0.023504 0.00539 Time2 7.705 932.986 0.006 0.000501 0.000181 Time3 1.063 1854.648 0.303 1.3599E-05 1.3187E-05 170 Table S3.7 Results from the overall models examining drug (VEH, CNO) during P/E at 5s intervals. F df p 5s Drug 0.126 364.738 0.722 Time 62.882 1978.373 <.001 Time2 1.344 3592.099 0.246 Time3 4.117 3736.59 0.043 Drug x time 0.347 1978.373 0.556 Drug x time2 0.001 3592.099 0.978 Drug x time3 0.149 3736.59 0.699 10s Drug 0.503 321.812 0.479 Time 82.151 1784.365 <.001 Time2 0.307 3104.424 0.579 Time3 4.117 3736.59 0.043 Drug x time 0.106 1784.365 0.745 Drug x time2 0.468 3104.424 0.494 Drug x time3 0.149 3736.59 0.699 15s Drug 1.994 354.253 0.159 Time 52.164 1729.808 <.001 Time2 3.736 3660.579 0.053 Time3 1.783 3736.784 0.182 Drug x time 0.926 1729.808 0.336 Drug x time2 0 3660.579 0.986 Drug x time3 0.074 3736.784 0.785 20s Drug 0.207 320.817 0.649 Time 70.181 1186.233 <.001 Time2 0.024 2780.328 0.877 Time3 0.871 3343.444 0.351 Drug x time 0.035 1186.233 0.852 Drug x time2 0.829 2780.328 0.363 Drug x time3 0.572 3343.444 0.449 25s Drug 0.002 320.943 0.962 Time 64.935 1842.733 <.001 Time2 0.848 2444.441 0.357 Time3 1.251 3521.115 0.263 Drug x time 0.11 1842.733 0.741 Drug x time2 0.039 2444.441 0.844 Drug x time3 0.049 3521.115 0.825 30s Drug 0.002 320.943 0.962 Time 64.935 1842.733 <.001 Time2 0.848 2444.441 0.357 171 Table S3.7 (cont’d) Time3 1.251 3521.115 0.263 Drug x time 0.11 1842.733 0.741 Drug x time2 0.039 2444.441 0.844 Drug x time3 0.049 3521.115 0.825 35s Drug 0.002 320.943 0.962 Time 64.935 1842.733 <.001 Time2 0.848 2444.441 0.357 Time3 1.251 3521.115 0.263 Drug x time 0.11 1842.733 0.741 Drug x time2 0.039 2444.441 0.844 Drug x time3 0.049 3521.115 0.825 40s Drug 37.645 324.669 0.962 Time 34.076 1842.733 <.001 Time2 0.234 2444.441 0.357 Time3 0.788 3521.115 0.263 Drug x time 37.351 1842.733 0.741 Drug x time2 30.864 2444.441 0.844 Drug x time3 0.709 3521.115 0.825 172 Table S3.8 Results from the interaction model examining the effects of drug (VEH, CNO) during P/E at 5s intervals. Regression standard F df p slope (b) error (se) 5s VEH: Intercept 10.112 366.856 0.002 0.225035 0.070768 Time 36.203 1962.42 <.001 0.055797 0.009273 Time2 0.618 3663.962 0.432 0.001268 0.001613 Time3 2.632 3777.54 0.105 -0.000124 7.6631E-05 5s CNO: Intercept 7.115 362.651 0.008 0.189404 0.071007 Time 27.005 1994.524 <.001 0.048083 0.009253 Time2 0.73 3503.502 0.393 0.001331 0.001557 Time3 1.513 3669.495 0.219 -8.455E-05 6.8744E-05 10s VEH: Intercept 59.42 321.141 <.001 0.520183 0.067482 Time 42.171 1970.822 <.001 0.059154 0.009109 Time2 0.779 3150.778 0.378 -0.000597 0.000676 Time3 2.632 3777.54 0.105 -0.000124 7.6631E-05 10s CNO: Intercept 45.098 322.486 <.001 0.452518 0.067384 Time 39.987 1601.515 <.001 0.055049 0.008705 Time2 0.008 3059.067 0.928 6.2493E-05 0.000687 Time3 1.513 3669.495 0.219 -8.455E-05 6.8744E-05 15s VEH: Intercept 77.936 389.52 <.001 0.581562 0.065876 Time 20.192 1662.679 <.001 0.037642 0.008377 Time2 1.631 3706.615 0.202 -0.001221 0.000956 Time3 0.503 3763.414 0.478 -5.54E-05 7.8081E-05 15s CNO: Intercept 102.134 327.542 <.001 0.718424 0.071088 Time 32.533 1795.789 <.001 0.049214 0.008628 Time2 2.2 3585.583 0.138 -0.001199 0.000809 Time3 1.472 3696.353 0.225 -8.385E-05 6.9103E-05 20s VEH: Intercept 37.813 316.547 <.001 0.44834 0.07291 Time 34.719 1216.471 <.001 0.045339 0.007695 Time2 0.539 2933.514 0.463 0.000641 0.000873 Time3 0.014 3386.008 0.905 -6.304E-06 5.2673E-05 20s CNO: Intercept 45.586 325.116 <.001 0.49542 0.073376 Time 35.543 1153.524 <.001 0.043367 0.007274 Time2 0.301 2611.675 0.583 -0.000454 0.000828 Time3 1.565 3291.273 0.211 -6.034E-05 4.8241E-05 25s VEH: Intercept 37.645 324.669 <.001 0.418801 0.068258 Time 34.076 1902.558 <.001 0.057683 0.009882 Time2 0.234 2472.417 0.628 0.000357 0.000738 Time3 0.788 3538.491 0.375 -8.73E-05 9.8341E-05 25s CNO: Intercept 37.351 317.227 <.001 0.414209 0.067775 Time 30.864 1781.003 <.001 0.053128 0.009563 Time2 0.709 2409.244 0.4 0.000552 0.000655 173 Table S3.8 (cont’d) Time3 0.468 3497.548 0.494 -5.848E-05 8.55E-05 30s VEH: Intercept 37.645 324.669 <.001 0.416505 0.048095 Time 34.076 1902.558 <.001 0.002296 0.048095 Time2 0.234 2472.417 0.628 0.055406 0.006876 Time3 0.788 3538.491 0.375 0.000454 0.000493 30s CNO: Intercept 64.292 277.695 <.001 0.509811 0.063581 Time 41.36 611.286 <.001 -0.028026 0.004358 Time2 0.699 1863.603 0.403 -0.000307 0.000367 Time3 3.999 1962.765 0.046 2.4304E-05 1.2154E-05 35s VEH: Intercept 30.269 235.877 <.001 0.317345 0.057681 Time 35.629 643.627 <.001 -0.028854 0.004834 Time2 0.05 1236.636 0.823 -3.83E-05 0.000171 Time3 8.83 1863.174 0.003 3.0282E-05 1.0191E-05 35s CNO: Intercept 37.535 249.374 <.001 0.365056 0.059586 Time 34.688 657.822 <.001 -0.029268 0.004969 Time2 0.068 1475.878 0.794 5.8029E-05 0.000222 Time3 3.999 1962.765 0.046 2.4304E-05 1.2154E-05 40s VEH: Intercept 37.645 324.669 <.001 0.418801 0.068258 Time 34.076 1902.558 <.001 0.057683 0.009882 Time2 0.234 2472.417 0.628 0.000357 0.000738 Time3 0.788 3538.491 0.375 -8.73E-05 9.8341E-05 40s CNO: Intercept 14.665 235.367 <.001 0.223203 0.058286 Time 26.508 715.757 <.001 -0.026865 0.005218 Time2 5.876 1004.491 0.016 0.000423 0.000174 Time3 3.999 1962.765 0.046 2.4304E-05 1.2154E-05 174 Table S3.9 Results from the overall model examining drug treatment within M/D. F df p Pre-peak: Drug 0.105 294.992 0.746 Time 67.789 1597.905 <.001 Time2 1.376 2346.847 0.241 Time3 0.066 3438.556 0.797 Drug x time 0.011 1597.905 0.915 Drug x time2 0.013 2346.847 0.911 Drug x time3 0.139 3438.556 0.709 Post-peak: Drug 1.211 228.086 0.272 Time 50.393 728.068 <.001 Time2 4.991 927.629 0.026 Time3 13.232 1882.963 <.001 Drug x time 2.574 728.068 0.109 Drug x time2 2.039 927.629 0.154 Drug x time3 4.351 1882.963 0.037 175 Table S3.10 Results from the interaction model examining drug treatment within M/D. Regression standard F df p slope (b) error (se) Pre-peak P/E: Intercept 36.375 308.48 <.001 0.414123 0.068663 Time 27.6 1818.035 <.001 0.049077 0.009342 Time2 0.651 2628.126 0.42 0.000542 0.000673 Time3 0.004 3571.668 0.948 5.3991E-06 8.2972E-05 Pre-peak M/D: Intercept 45.004 281.479 <.001 0.445084 0.066346 Time 43.28 1324.524 <.001 0.050366 0.007656 Time2 0.77 1912.845 0.38 0.000448 0.00051 Time3 0.45 2864.897 0.502 -2.962E-05 4.4144E-05 Post-peak P/E: Intercept 20.411 226.262 <.001 0.269175 0.059581 Time 39.441 704.575 <.001 -0.03258 0.005188 Time2 0.348 913.556 0.556 9.9636E-05 0.000169 Time3 18.347 1827.397 <.001 4.9897E-05 1.1649E-05 Post-peak M/D: Intercept 8.564 229.898 0.004 0.175996 0.06014 Time 14.517 750.736 <.001 -0.020569 0.005398 Time2 6.294 940.218 0.012 0.000453 0.00018 Time3 1.087 1929.013 0.297 1.3527E-05 1.2973E-05 176 Table S3.11 Results from the overall models examining drug (VEH, CNO) during M/D at 5s intervals. F df p 5s Drug 0.06 325.897 0.806 Time 58.041 2004.507 <.001 Time2 0 3537.178 0.986 Time3 0.461 3774.461 0.497 Drug x time 0.031 2004.507 0.86 Drug x time2 0.074 3537.178 0.786 Drug x time3 0.015 3774.461 0.903 10s Drug 0.064 313.892 0.8 Time 89.679 1555.549 <.001 Time2 0.497 3085.682 0.481 Time3 0.461 3774.461 0.497 Drug x time 0.036 1555.549 0.849 Drug x time2 0.149 3085.682 0.699 Drug x time3 0.015 3774.461 0.903 15s Drug 0.17 312.698 0.681 Time 67.491 1772.53 <.001 Time2 2.773 3588.446 0.096 Time3 1.645 3729.05 0.2 Drug x time 0.063 1772.53 0.801 Drug x time2 0.066 3588.446 0.797 Drug x time3 0.247 3729.05 0.619 20s Drug 0.023 300.552 0.88 Time 62.963 1160.092 <.001 Time2 1.761 3093.825 0.185 Time3 0.071 3422.7 0.79 Drug x time 0.911 1160.092 0.34 Drug x time2 0.231 3093.825 0.631 Drug x time3 0 3422.7 0.997 25s Drug 0.085 305.558 0.77 Time 83.746 1097.628 <.001 Time2 4.186 2540.216 0.041 Time3 0.034 2598.449 0.854 Drug x time 0.611 1097.628 0.434 Drug x time2 0.457 2540.216 0.499 Drug x time3 0.088 2598.449 0.767 30s Drug 0.305 300.752 0.581 Time 63.08 1499.56 <.001 Time2 2.867 2374.49 0.091 177 Table S3.11 (cont’d) Time3 0.099 3465.124 0.753 Drug x time 0.208 1499.56 0.648 Drug x time2 0.017 2374.49 0.896 Drug x time3 0.163 3465.124 0.686 35s Drug 0.26 297.046 0.61 Time 65.656 1572.331 <.001 Time2 1.57 2186.788 0.21 Time3 0.027 3394.433 0.869 Drug x time 0.052 1572.331 0.82 Drug x time2 0.427 2186.788 0.513 Drug x time3 0.669 3394.433 0.414 40s Drug 32.286 312.328 0.61 Time 25.855 1572.331 <.001 Time2 1.424 2186.788 0.21 Time3 0.137 3394.433 0.869 Drug x time 43.777 1572.331 0.82 Drug x time2 43.336 2186.788 0.513 Drug x time3 0.248 3394.433 0.414 178 Table S3.12 Results from the interaction model examining the effects of drug (VEH, CNO) during M/D at 5s intervals. Regression standard F df p slope (b) error (se) 5s VEH: Intercept 11.311 341.528 <.001 0.237809 0.070708 Time 30.272 2028.294 <.001 0.05085 0.009242 Time2 0.025 3644.755 0.874 -0.000227 0.00143 Time3 0.097 3798.384 0.755 -1.892E-05 6.0661E-05 5s CNO: Intercept 9.767 310.301 0.002 0.213641 0.068362 Time 27.791 1980.823 <.001 0.048556 0.009211 Time2 0.058 3302.734 0.809 0.000258 0.00107 Time3 0.799 3609.4 0.372 -2.716E-05 3.0389E-05 10s VEH: Intercept 49.785 314.871 <.001 0.484024 0.068599 Time 34.462 1831.668 <.001 0.047163 0.008034 Time2 0.579 3154.949 0.447 -0.000511 0.000671 Time3 0.097 3798.384 0.755 -1.892E-05 6.0661E-05 10s CNO: Intercept 45.221 312.91 <.001 0.459476 0.068327 Time 62.156 1196.802 <.001 0.049099 0.006228 Time2 0.052 3011.282 0.819 -0.000149 0.000652 Time3 0.799 3609.4 0.372 -2.716E-05 3.0389E-05 15s VEH: Intercept 104.254 347.338 <.001 0.659207 0.064562 Time 33.441 1704.18 <.001 0.048445 0.008377 Time2 1.155 3747.043 0.283 -0.000755 0.000703 Time3 0.984 3747.841 0.321 -6.136E-05 6.1868E-05 15s CNO: Intercept 106.44 285.338 <.001 0.697703 0.067627 Time 34.165 1855.123 <.001 0.045564 0.007795 Time2 2.478 2547.544 0.116 -0.000553 0.000351 Time3 0.791 3632.277 0.374 -2.709E-05 3.0467E-05 20s VEH: Intercept 43.225 315.811 <.001 0.482585 0.073402 Time 37.679 1220.304 <.001 0.045854 0.00747 Time2 1.079 2930.84 0.299 0.000874 0.000841 Time3 0.025 3417.818 0.875 7.8304E-06 4.9967E-05 20s CNO: Intercept 58.137 282.733 <.001 0.497457 0.065242 Time 25.609 1098.065 <.001 0.036006 0.007115 Time2 0.738 3546.882 0.39 0.000409 0.000477 Time3 0.062 3434.377 0.803 8.0356E-06 3.2197E-05 25s VEH: Intercept 31.051 304.572 <.001 0.393216 0.070566 Time 46.668 1143.389 <.001 0.047918 0.007014 Time2 3.507 2509.313 0.061 0.001292 0.00069 Time3 0.101 2462.088 0.75 1.2156E-05 3.8168E-05 25s CNO: Intercept 36.706 306.573 <.001 0.422212 0.069689 Time 37.145 1049.11 <.001 0.040374 0.006624 Time2 0.994 2575.028 0.319 0.00065 0.000652 179 Table S3.12 (cont’d) Time3 0.007 2782.542 0.932 -2.851E-06 3.3272E-05 30s VEH: Intercept 31.605 308.532 <.001 0.418851 0.048832 Time 25.549 1822.154 <.001 -0.026961 0.048832 Time2 1.417 2450.906 0.234 0.045008 0.005667 Time3 0.14 3550.616 0.708 0.000759 0.000448 30s CNO: Intercept 43.925 266.797 <.001 0.431005 0.065032 Time 28.674 605.54 <.001 -0.023455 0.00438 Time2 1.982 1877.179 0.159 -0.000485 0.000345 Time3 5.831 2026.334 0.016 2.8554E-05 1.1825E-05 35s VEH: Intercept 52.759 231.923 <.001 0.429145 0.059082 Time 45.874 668.217 <.001 -0.03277 0.004838 Time2 12.027 1395.022 <.001 -0.000604 0.000174 Time3 29.265 2075.426 <.001 5.4761E-05 1.0123E-05 35s CNO: Intercept 25.665 245.861 <.001 0.308021 0.060801 Time 23.781 704.067 <.001 -0.023953 0.004912 Time2 0.272 1625.294 0.602 0.000111 0.000212 Time3 2.464 2248.109 0.117 1.8305E-05 1.1661E-05 40s VEH: Intercept 32.286 312.328 <.001 0.392254 0.069034 Time 25.855 1818.453 <.001 0.047638 0.009369 Time2 1.424 2441.21 0.233 0.000816 0.000683 Time3 0.137 3543.773 0.712 3.1275E-05 8.4637E-05 40s CNO: Intercept 10.463 231.408 0.001 0.193312 0.059763 Time 17.581 777.979 <.001 -0.021473 0.005121 Time2 4.967 1043.473 0.026 0.000385 0.000173 Time3 2.464 2248.109 0.117 1.8305E-05 1.1661E-05 180 Chapter 4 Supplemental Materials Table S4.1 Results from the overall model examining hormone effects under vehicle. F df p Pre-peak: Hormone 0.989 228.568 0.321 Time 68.467 1899.678 <.001 Time2 0.952 2616.922 0.329 Time3 0.054 3281.154 0.815 Hormone x time 0.279 1899.678 0.598 Hormone x time2 1.665 2616.922 0.197 Hormone x time3 1.454 3281.154 0.228 Post-peak: Hormone 0.149 180.123 0.7 Time 54.08 845.436 <.001 Time2 17.257 1213.586 <.001 Time3 8.753 2368.068 0.003 Hormone x time 0.127 845.436 0.722 Hormone x time2 0.213 1213.586 0.645 Hormone x time3 0.239 2368.068 0.625 181 Table S4.2 Results from the interaction model examining effects of time within vehicle treated rats tested under oil and EB. Regression standard F df p slope (b) error (se) Pre-peak Oil: Intercept 56.433 231.381 <.001 0.044892 0.008382 Time 28.681 1969.441 <.001 -0.000815 0.000542 Time2 2.263 2678.835 0.133 4.1289E-05 6.5437E-05 Time3 0.398 3292.53 0.528 0.446805 0.072532 Pre-peak EB: Intercept 37.947 225.761 <.001 0.05101 0.008004 Time 40.615 1825.575 <.001 0.000113 0.000473 Time2 0.057 2535.033 0.811 -6.11E-05 5.409E-05 Time3 1.276 3262.595 0.259 0.19498 0.066394 Post-peak Oil: Intercept 8.624 178.329 0.004 -0.028162 0.005068 Time 30.876 817.705 <.001 0.000451 0.000166 Time2 7.434 1165.589 0.006 2.8428E-05 1.0925E-05 Time3 6.771 2304.139 0.009 0.15864 0.066961 Post-peak EB: Intercept 5.613 181.913 0.019 -0.02556 0.005261 Time 23.603 872.302 <.001 0.000564 0.00018 Time2 9.838 1256.136 0.002 2.0369E-05 1.2357E-05 Time3 2.717 2418.393 0.099 0 0 182 Table S4.3 Results from the overall models examining the effects of hormone under treatment vehicle at 5s intervals. F df p 5s Hormone 2.326 255.152 0.128 Time 88.021 2093.576 <.001 Time2 0.829 3303.262 0.363 Time3 0.086 3341.039 0.77 Hormone x time 2.342 2093.576 0.126 Hormone x time2 5.66 3303.262 0.017 Hormone x time3 4.762 3341.039 0.029 10s Hormone 2.637 237.169 0.106 Time 82.254 1765.341 <.001 Time2 5.232 3090.815 0.022 Time3 0.086 3341.039 0.77 Hormone x time 1.302 1765.341 0.254 Hormone x time2 4.861 3090.815 0.028 Hormone x time3 4.762 3341.039 0.029 15s Hormone 1.871 227.118 0.173 Time 46.542 1763.111 <.001 Time2 4.121 3278.453 0.042 Time3 0.07 3327.746 0.792 Hormone x time 1.445 1763.111 0.23 Hormone x time2 0.063 3278.453 0.802 Hormone x time3 2.937 3327.746 0.087 20s Hormone 0.076 216.372 0.784 Time 38.137 1288.669 <.001 Time2 1.836 2963.595 0.176 Time3 0.36 3276.423 0.548 Hormone x time 0.022 1288.669 0.883 Hormone x time2 0.445 2963.595 0.505 Hormone x time3 0.769 3276.423 0.381 25s Hormone 0.191 214.574 0.663 Time 86.139 1159.747 <.001 Time2 0.16 2616.049 0.689 Time3 1.05 2667.625 0.306 Hormone x time 0.981 1159.747 0.322 Hormone x time2 0.315 2616.049 0.575 Hormone x time3 0.147 2667.625 0.702 30s Hormone 0.992 230.826 0.32 Time 65.486 1899.36 <.001 Time2 0.8 2494.223 0.371 183 Table S4.3 (cont’d) Time3 0.04 3275.139 0.841 Hormone x time 0.246 1899.36 0.62 Hormone x time2 1.284 2494.223 0.257 Hormone x time3 1.528 3275.139 0.216 35s Hormone 0.992 230.826 0.32 Time 65.486 1899.36 <.001 Time2 0.8 2494.223 0.371 Time3 0.04 3275.139 0.841 Hormone x time 0.246 1899.36 0.62 Hormone x time2 1.284 2494.223 0.257 Hormone x time3 1.528 3275.139 0.216 40s Hormone 52.476 233.594 0.32 Time 27.629 1899.36 <.001 Time2 1.809 2494.223 0.371 Time3 0.453 3275.139 0.841 Hormone x time 34.607 1899.36 0.62 Hormone x time2 38.588 2494.223 0.257 Hormone x time3 0.033 3275.139 0.216 184 Table S4.4 Results from the interaction models examining the effects of hormone under vehicle treatment at 5s intervals. Regression standard F df p slope (b) error (se) 5s Oil: Intercept 20.392 256.212 <.001 0.337654 0.074773 Time 59.475 2094.662 <.001 0.064232 0.008329 Time2 5.148 3318.274 0.023 -0.002853 0.001257 Time3 1.569 3343.975 0.21 6.6587E-05 5.3165E-05 5s EB: Intercept 5.584 254.097 0.019 0.176471 0.074678 Time 30.858 2092.489 <.001 0.046217 0.00832 Time2 1.136 3282.816 0.286 0.001274 0.001195 Time3 3.553 3329.726 0.06 -8.719E-05 4.6258E-05 10s Oil: Intercept 66.816 237.189 <.001 0.59582 0.072891 Time 30.029 1896.164 <.001 0.040699 0.007427 Time2 10.173 3099.307 0.001 -0.001854 0.000581 Time3 1.569 3343.975 0.21 6.6587E-05 5.3165E-05 10s EB: Intercept 34.637 237.148 <.001 0.428504 0.072809 Time 54.675 1631.544 <.001 0.052416 0.007089 Time2 0.003 3082.354 0.954 -3.41E-05 0.000586 Time3 3.553 3329.726 0.06 -8.719E-05 4.6258E-05 15s Oil: Intercept 88.334 230.58 <.001 0.669824 0.071269 Time 15.786 1849.595 <.001 0.029968 0.007543 Time2 1.262 3307.946 0.261 -0.000663 0.00059 Time3 0.912 3337.318 0.34 5.1563E-05 5.3983E-05 15s EB: Intercept 58.163 223.601 <.001 0.533259 0.069922 Time 32.207 1680.373 <.001 0.042787 0.007539 Time2 3.489 3210.816 0.062 -0.000849 0.000455 Time3 2.306 3308.19 0.129 -7.034E-05 4.6316E-05 20s Oil: Intercept 47.165 213.844 <.001 0.559988 0.08154 Time 19.248 1308.143 <.001 0.030037 0.006847 Time2 0.233 2986.282 0.629 -0.000376 0.000777 Time3 0.037 3285.353 0.848 8.0749E-06 4.2065E-05 20s EB: Intercept 42.04 218.948 <.001 0.528293 0.081478 Time 18.898 1268.021 <.001 0.028639 0.006588 Time2 2.071 2939.866 0.15 -0.001104 0.000767 Time3 1.134 3266.122 0.287 -4.312E-05 4.0489E-05 25s Oil: Intercept 45.023 214.241 <.001 0.516575 0.076987 Time 49.391 1227.739 <.001 0.045515 0.006476 Time2 0.435 2641.484 0.51 -0.000351 0.000531 Time3 0.82 2489.824 0.365 -2.762E-05 3.0493E-05 25s EB: Intercept 37.174 214.908 <.001 0.469009 0.076924 Time 36.878 1087.451 <.001 0.036739 0.00605 Time2 0.014 2587.257 0.907 5.875E-05 0.0005 185 Table S4.4 (cont’d) Time3 0.26 2925.709 0.61 -1.258E-05 2.4684E-05 30s Oil: Intercept 52.476 233.594 <.001 0.479592 0.05165 Time 27.629 1960.103 <.001 -0.051432 0.05165 Time2 1.809 2559.404 0.179 0.04709 0.005819 Time3 0.453 3287.65 0.501 -0.00033 0.000369 30s EB: Intercept 53.043 209.425 <.001 0.510698 0.070122 Time 40.766 749.413 <.001 -0.027905 0.004371 Time2 2.649 2532.655 0.104 -0.000507 0.000311 Time3 8.261 2562.247 0.004 2.9958E-05 1.0423E-05 35s Oil: Intercept 32.562 178.834 <.001 0.373779 0.065503 Time 48.222 755.218 <.001 -0.032968 0.004748 Time2 0.653 1743.795 0.419 -0.000132 0.000163 Time3 16.356 2468.946 <.001 3.7278E-05 9.2176E-06 35s EB: Intercept 29.593 186.014 <.001 0.362248 0.066591 Time 40.064 785.885 <.001 -0.030726 0.004854 Time2 0.088 2047.915 0.767 -5.74E-05 0.000194 Time3 8.261 2562.247 0.004 2.9958E-05 1.0423E-05 40s Oil: Intercept 52.476 233.594 <.001 0.531024 0.073305 Time 27.629 1960.103 <.001 0.044203 0.00841 Time2 1.809 2559.404 0.179 -0.000748 0.000556 Time3 0.453 3287.65 0.501 4.5083E-05 6.6999E-05 40s EB: Intercept 10.23 179.537 0.002 0.210928 0.065946 Time 34.128 825.417 <.001 -0.029053 0.004973 Time2 5.716 1396.245 0.017 0.000392 0.000164 Time3 8.261 2562.247 0.004 2.9958E-05 1.0423E-05 186 Table S4.5 Results from the overall model examining drug treatment under oil. F df p Pre-peak: Drug 0.043 235.045 0.835 Time 62.86 1984.134 <.001 Time2 4.896 2748.843 0.027 Time3 0.011 3303.571 0.918 Drug x time 0.21 1984.134 0.646 Drug x time2 0.033 2748.843 0.856 Drug x time3 0.508 3303.571 0.476 Post-peak: Drug 0.015 179.519 0.903 Time 59.994 812.578 <.001 Time2 12.039 1141.348 <.001 Time3 13.452 2284.507 <.001 Drug x time 0.008 812.578 0.927 Drug x time2 0.168 1141.348 0.682 Drug x time3 0.003 2284.507 0.959 187 Table S4.6 Results from the interaction model examining drug treatment under oil. Regression standard F df p slope (b) error (se) Pre-peak VEH: Intercept 58.002 232.982 <.001 0.549409 0.072139 Time 28.918 1941.364 <.001 0.044943 0.008358 Time2 2.261 2647.08 0.133 -0.000813 0.000541 Time3 0.386 3285.654 0.534 4.0664E-05 6.5416E-05 Pre-peak CNO: Intercept 53.011 237.112 <.001 0.528101 0.072532 Time 33.974 2024.639 <.001 0.050464 0.008658 Time2 2.635 2831.924 0.105 -0.000959 0.000591 Time3 0.163 3315.017 0.686 -3.037E-05 7.5213E-05 Post-peak VEH: Intercept 8.858 179.132 0.003 0.194592 0.065382 Time 31.21 801.535 <.001 -0.028134 0.005036 Time2 7.565 1136.688 0.006 0.000453 0.000165 Time3 6.768 2258.629 0.009 2.8354E-05 1.0899E-05 Post-peak CNO: Intercept 9.882 179.905 0.002 0.205932 0.065509 Time 28.83 823.47 <.001 -0.027474 0.005117 Time2 4.657 1145.984 0.031 0.000357 0.000166 Time3 6.69 2308.808 0.01 2.9157E-05 1.1273E-05 188 Table S4.7 Results from the overall models examining drug (VEH, CNO) under oil at 5s intervals. F df p 5s Drug 0.27 263.323 0.604 Time 116.86 2059.452 <.001 Time2 3.134 3312.538 0.077 Time3 0.006 3342.554 0.938 Drug x time 0.022 2059.452 0.882 Drug x time2 1.729 3312.538 0.189 Drug x time3 2.917 3342.554 0.088 10s Drug 0.038 239.745 0.846 Time 73.819 1886.792 <.001 Time2 16.533 3089.808 <.001 Time3 0.006 3342.554 0.938 Drug x time 1.237 1886.792 0.266 Drug x time2 0.173 3089.808 0.677 Drug x time3 2.917 3342.554 0.088 15s Drug 1.421 247.703 0.234 Time 34.256 1916.4 <.001 Time2 11.668 3303.758 <.001 Time3 0.067 3339.375 0.796 Drug x time 0.049 1916.4 0.825 Drug x time2 4.036 3303.758 0.045 Drug x time3 2.278 3339.375 0.131 20s Drug 0.116 214.733 0.734 Time 44.995 1306.89 <.001 Time2 0.293 2971.81 0.588 Time3 0.031 3268.426 0.86 Drug x time 0.291 1306.89 0.59 Drug x time2 0.028 2971.81 0.867 Drug x time3 0.008 3268.426 0.928 25s Drug 0.036 227.944 0.85 Time 73.863 1711.661 <.001 Time2 2.182 2652.467 0.14 Time3 0.185 3247.976 0.667 Drug x time 0.043 1711.661 0.835 Drug x time2 0.372 2652.467 0.542 Drug x time3 0.076 3247.976 0.783 30s Drug 0.1 237.002 0.752 Time 57.804 1998.143 <.001 Time2 3.768 2616.35 0.052 189 Table S4.7 (cont’d) Time3 0.132 3292.072 0.717 Drug x time 0.086 1998.143 0.77 Drug x time2 0.012 2616.35 0.913 Drug x time3 0.256 3292.072 0.613 35s Drug 0.1 237.002 0.752 Time 57.804 1998.143 <.001 Time2 3.768 2616.35 0.052 Time3 0.132 3292.072 0.717 Drug x time 0.086 1998.143 0.77 Drug x time2 0.012 2616.35 0.913 Drug x time3 0.256 3292.072 0.613 40s Drug 53.965 234.793 0.752 Time 27.868 1998.143 <.001 Time2 1.821 2616.35 0.052 Time3 0.441 3292.072 0.717 Drug x time 46.884 1998.143 0.77 Drug x time2 29.937 2616.35 0.913 Drug x time3 1.949 3292.072 0.613 190 Table S4.8 Results from the interaction model examining the effects of drug (VEH, CNO) under oil at 5s intervals. Regression standard F df p slope (b) error (se) 5s VEH: Intercept 21.014 259.382 <.001 0.337325 0.073586 Time 60.059 2067.725 <.001 0.064296 0.008297 Time2 5.142 3312.566 0.023 -0.002849 0.001256 Time3 1.556 3343.664 0.212 6.6283E-05 5.3144E-05 5s CNO: Intercept 14.467 267.264 <.001 0.282978 0.074399 Time 56.824 2051.207 <.001 0.062562 0.008299 Time2 0.097 3312.513 0.756 -0.000421 0.001354 Time3 1.392 3341.222 0.238 -7.263E-05 6.1566E-05 10s VEH: Intercept 69.093 239.799 <.001 0.595876 0.071687 Time 30.413 1868.812 <.001 0.040781 0.007395 Time2 10.213 3076.714 0.001 -0.001854 0.00058 Time3 1.556 3343.664 0.212 6.6283E-05 5.3144E-05 10s CNO: Intercept 64.617 239.691 <.001 0.576195 0.07168 Time 43.585 1902.216 <.001 0.052909 0.008014 Time2 6.554 3102.243 0.011 -0.00151 0.00059 Time3 1.392 3341.222 0.238 -7.263E-05 6.1566E-05 15s VEH: Intercept 95.854 243.514 <.001 0.67282 0.068722 Time 16.141 1814.232 <.001 0.029936 0.007451 Time2 1.318 3299.395 0.251 -0.000674 0.000587 Time3 0.906 3332.906 0.341 5.1154E-05 5.3739E-05 15s CNO: Intercept 113.078 251.336 <.001 0.793772 0.074646 Time 18.129 2019.483 <.001 0.032294 0.007585 Time2 11.783 3306.278 <.001 -0.002597 0.000757 Time3 1.375 3342.429 0.241 -7.236E-05 6.1708E-05 20s VEH: Intercept 48.759 215.791 <.001 0.560826 0.080316 Time 19.365 1286.476 <.001 0.02991 0.006797 Time2 0.238 2964.181 0.626 -0.000377 0.000774 Time3 0.036 3277.869 0.851 7.8976E-06 4.1902E-05 20s CNO: Intercept 42.121 213.686 <.001 0.522088 0.080444 Time 25.807 1326.959 <.001 0.035143 0.006918 Time2 0.074 2980.288 0.786 -0.000199 0.000731 Time3 0.004 3258.497 0.951 2.5635E-06 4.21E-05 25s VEH: Intercept 49.844 222.827 <.001 0.517712 0.07333 Time 51.337 1170.748 <.001 0.045476 0.006347 Time2 0.436 2547.704 0.509 -0.000347 0.000526 Time3 0.948 2428.379 0.33 -2.934E-05 3.0142E-05 25s CNO: Intercept 44.698 233.068 <.001 0.497966 0.074483 Time 29.459 2088.145 <.001 0.047728 0.008794 Time2 1.916 2730.95 0.166 -0.000835 0.000603 191 Table S4.8 (cont’d) Time3 0.007 3307.415 0.934 -6.428E-06 7.7562E-05 30s VEH: Intercept 53.965 234.793 <.001 0.515161 0.051337 Time 27.868 1932.587 <.001 0.016271 0.051337 Time2 1.821 2524.81 0.177 0.046006 0.006051 Time3 0.441 3280.631 0.507 -0.000793 0.000409 30s CNO: Intercept 59.997 205.993 <.001 0.531771 0.068653 Time 35.484 652.231 <.001 -0.025283 0.004244 Time2 9.729 2246.42 0.002 -0.000925 0.000297 Time3 16.243 2404.209 <.001 4.0898E-05 1.0148E-05 35s VEH: Intercept 33.446 179.482 <.001 0.37312 0.064517 Time 48.911 740.757 <.001 -0.032968 0.004714 Time2 0.633 1710.729 0.426 -0.000129 0.000162 Time3 16.407 2429.377 <.001 3.7234E-05 9.1924E-06 35s CNO: Intercept 35.076 184.044 <.001 0.387344 0.065402 Time 43.117 743.791 <.001 -0.031466 0.004792 Time2 2.873 1714.504 0.09 -0.000312 0.000184 Time3 16.243 2404.209 <.001 4.0898E-05 1.0148E-05 40s VEH: Intercept 53.965 234.793 <.001 0.531432 0.072342 Time 27.868 1932.587 <.001 0.044235 0.008379 Time2 1.821 2524.81 0.177 -0.000749 0.000555 Time3 0.441 3280.631 0.507 4.443E-05 6.6924E-05 40s CNO: Intercept 12.237 179.229 <.001 0.22734 0.064988 Time 41.399 781.99 <.001 -0.031513 0.004898 Time2 3.511 1260.857 0.061 0.000302 0.000161 Time3 16.243 2404.209 <.001 4.0898E-05 1.0148E-05 192 Table S4.9 Results from the overall model examining drug treatment under EB. F df p Pre-peak: Drug 0.009 222.865 0.923 Time 81.818 1942.567 <.001 Time2 0.195 2705.888 0.659 Time3 2.74 3294.305 0.098 Drug x time 0.061 1942.567 0.806 Drug x time2 0.015 2705.888 0.902 Drug x time3 0.028 3294.305 0.867 Post-peak: Drug 38.734 921.003 <.001 Time 20.829 1337.277 <.001 Time2 3.32 2540.751 0.069 Time3 0.078 178.725 0.78 Drug x time 0.306 921.003 0.58 Drug x time2 0.023 1337.277 0.879 Drug x time3 0.315 2540.751 0.575 193 Table S4.10 Results from the interaction model examining drug treatment under EB. Regression standard F df p slope (b) error (se) Pre-peak VEH: Intercept 34.532 221.254 <.001 0.44632 0.075951 Time 39.413 1916.266 <.001 0.050933 0.008113 Time2 0.056 2643.801 0.813 0.000113 0.000477 Time3 1.236 3288.703 0.266 -6.044E-05 5.4361E-05 Pre-peak CNO: Intercept 32.638 224.479 <.001 0.435864 0.076294 Time 42.413 1968.226 <.001 0.053783 0.008258 Time2 0.146 2756.678 0.702 0.000201 0.000525 Time3 1.503 3298.63 0.22 -7.408E-05 6.0419E-05 Post-peak VEH: Intercept 5.127 179.467 0.025 0.15958 0.070476 Time 22.678 930.479 <.001 -0.025616 0.005379 Time2 9.345 1360.386 0.002 0.000559 0.000183 Time3 2.705 2554.415 0.1 2.053E-05 1.2482E-05 Post-peak CNO: Intercept 3.521 177.982 0.062 0.131802 0.070238 Time 16.282 911.418 <.001 -0.021436 0.005312 Time2 11.601 1312.691 <.001 0.000598 0.000176 Time3 0.837 2525.654 0.36 1.0863E-05 1.1877E-05 194 Table S4.11 Results from the overall models examining drug (VEH, CNO) under EB at 5s intervals. F df p 5s Drug 0.09 244.206 0.764 Time 63.403 2168.748 <.001 Time2 0.972 3311.441 0.324 Time3 4.539 3342.405 0.033 Drug x time 0.052 2168.748 0.82 Drug x time2 0.262 3311.441 0.609 Drug x time3 0.265 3342.405 0.607 10s Drug 0.074 229.523 0.786 Time 99.858 1737.485 <.001 Time2 0.271 3147.197 0.603 Time3 4.539 3342.405 0.033 Drug x time 0.111 1737.485 0.739 Drug x time2 0.184 3147.197 0.668 Drug x time3 0.265 3342.405 0.607 15s Drug 0.002 213.623 0.964 Time 57.87 1781.731 <.001 Time2 5.509 3269.353 0.019 Time3 2.942 3331.564 0.086 Drug x time 0.137 1781.731 0.711 Drug x time2 0.018 3269.353 0.892 Drug x time3 0.17 3331.564 0.68 20s Drug 0.063 218.898 0.802 Time 43.789 1328.744 <.001 Time2 4.795 3037.792 0.029 Time3 3.442 3295.643 0.064 Drug x time 0.309 1328.744 0.578 Drug x time2 0.056 3037.792 0.814 Drug x time3 0.159 3295.643 0.69 25s Drug 0.161 213.274 0.689 Time 85.052 1209.903 <.001 Time2 0.006 2703.579 0.938 Time3 1.456 2688.1 0.228 Drug x time 1.032 1209.903 0.31 Drug x time2 0.006 2703.579 0.938 Drug x time3 0.354 2688.1 0.552 30s Drug 0.021 224.479 0.885 Time 77.208 1949.524 <.001 Time2 0.141 2590.732 0.708 195 Table S4.11 (cont’d) Time3 2.687 3287.285 0.101 Drug x time 0.048 1949.524 0.827 Drug x time2 0.013 2590.732 0.908 Drug x time3 0.027 3287.285 0.869 35s Drug 0.021 224.479 0.885 Time 77.208 1949.524 <.001 Time2 0.141 2590.732 0.708 Time3 2.687 3287.285 0.101 Drug x time 0.048 1949.524 0.827 Drug x time2 0.013 2590.732 0.908 Drug x time3 0.027 3287.285 0.869 40s Drug 31.201 222.957 0.885 Time 37.391 1949.524 <.001 Time2 0.037 2590.732 0.708 Time3 1.213 3287.285 0.101 Drug x time 28.675 1949.524 0.827 Drug x time2 39.82 2590.732 0.908 Drug x time3 0.111 3287.285 0.869 196 Table S4.12 Results from the interaction model examining the effects of drug (VEH, CNO) under EB at 5s intervals. Regression standard F df p slope (b) error (se) 5s VEH: Intercept 5.131 245.411 0.024 0.177785 0.078484 Time 29.935 2169.769 <.001 0.046168 0.008438 Time2 1.104 3306.932 0.294 0.001263 0.001202 Time3 3.482 3338.758 0.062 -8.675E-05 4.6493E-05 5s CNO: Intercept 7.273 243.003 0.007 0.211043 0.078254 Time 33.517 2167.728 <.001 0.04889 0.008445 Time2 0.114 3315.801 0.735 0.0004 0.001183 Time3 1.312 3343.947 0.252 -5.301E-05 4.6281E-05 10s VEH: Intercept 31.377 229.559 <.001 0.429353 0.07665 Time 52.567 1716.271 <.001 0.05229 0.007212 Time2 0.004 3145.399 0.948 -3.843E-05 0.000591 Time3 3.482 3338.758 0.062 -8.675E-05 4.6493E-05 10s CNO: Intercept 35.798 229.487 <.001 0.458864 0.076693 Time 47.328 1759.575 <.001 0.048913 0.00711 Time2 0.453 3149.008 0.501 -0.000395 0.000587 Time3 1.312 3343.947 0.252 -5.301E-05 4.6281E-05 15s VEH: Intercept 51.55 212.905 <.001 0.527972 0.073535 Time 31.285 1749.575 <.001 0.042904 0.007671 Time2 3.342 3239.281 0.068 -0.000838 0.000458 Time3 2.268 3320.974 0.132 -7.023E-05 4.6635E-05 15s CNO: Intercept 51.95 214.337 <.001 0.532733 0.073913 Time 26.643 1815.587 <.001 0.038919 0.00754 Time2 2.268 3291.284 0.132 -0.000746 0.000496 Time3 0.847 3338.923 0.357 -4.298E-05 4.6708E-05 20s VEH: Intercept 39.844 217.982 <.001 0.526342 0.083385 Time 18.772 1315.816 <.001 0.028823 0.006652 Time2 2.002 2985.858 0.157 -0.001091 0.000771 Time3 1.099 3281.894 0.294 -4.262E-05 4.0644E-05 20s CNO: Intercept 44.406 219.819 <.001 0.555911 0.083423 Time 25.188 1341.268 <.001 0.034111 0.006797 Time2 2.811 3082.604 0.094 -0.001354 0.000807 Time3 2.454 3306.976 0.117 -6.593E-05 4.2086E-05 25s VEH: Intercept 34.497 213.1 <.001 0.46814 0.079705 Time 35.967 1145.385 <.001 0.036842 0.006143 Time2 0.013 2673.176 0.91 5.7302E-05 0.000504 Time3 0.239 2981.275 0.625 -1.215E-05 2.4845E-05 25s CNO: Intercept 28.06 213.448 <.001 0.42292 0.079838 Time 49.27 1270.38 <.001 0.045962 0.006548 Time2 0 2730.289 1 3.1457E-07 0.000536 197 Table S4.12 (cont’d) Time3 1.331 2487.083 0.249 -3.581E-05 3.1037E-05 30s VEH: Intercept 31.201 222.957 <.001 0.418593 0.054111 Time 37.391 1926.526 <.001 0.007851 0.054111 Time2 0.037 2530.854 0.848 0.051191 0.005826 Time3 1.213 3283.648 0.271 0.000136 0.000363 30s CNO: Intercept 38.771 201.475 <.001 0.455466 0.073148 Time 39.609 745.48 <.001 -0.027857 0.004426 Time2 0.996 2569.14 0.318 -0.000301 0.000302 Time3 6.139 2661.264 0.013 2.5397E-05 1.025E-05 35s VEH: Intercept 26.92 182.424 <.001 0.36421 0.070197 Time 37.924 837.968 <.001 -0.03065 0.004977 Time2 0.111 2163.455 0.739 -6.532E-05 0.000196 Time3 8.12 2685.022 0.004 3.0014E-05 1.0533E-05 35s CNO: Intercept 19.883 180.495 <.001 0.311823 0.06993 Time 34.515 821.138 <.001 -0.028966 0.00493 Time2 0.177 2028.637 0.674 7.9572E-05 0.000189 Time3 6.139 2661.264 0.013 2.5397E-05 1.025E-05 40s VEH: Intercept 31.201 222.957 <.001 0.426444 0.076345 Time 37.391 1926.526 <.001 0.04992 0.008164 Time2 0.037 2530.854 0.848 9.4153E-05 0.00049 Time3 1.213 3283.648 0.271 -6.154E-05 5.5888E-05 40s CNO: Intercept 6.122 176.211 0.014 0.172159 0.069578 Time 27.328 859.386 <.001 -0.026265 0.005024 Time2 7.731 1452.05 0.005 0.000461 0.000166 Time3 6.139 2661.264 0.013 2.5397E-05 1.025E-05 198