THE DEVELOPMENT OF NEURAL PROCESSES FOR RHYME IN YOUNG CHILDREN By Valerie Rose A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of Communicative Sciences and Disorders—Master of Arts 2018 THE DEVELOPMENT OF NEURAL PROCESSES FOR RHYME IN YOUNG CHILDREN ABSTRACT By Valerie Rose Phonological awareness abilities have been established as strong predictors of literacy development. One early developing skill involved in phonological awareness is rhyme detection, an ability that emerges around the age of 3 years. The neural processes underlying this skill have been assessed using event-related potentials (ERPs), revealing a pattern of response that is stable and adult-like in children as young as age 6 years. The current longitudinal study aimed to examine the development of these effects in children younger than previously assessed, from age 5 to 6 years. To do so, ERPs elicited by nonword rhyming and non-rhyming pairs were acquired in 12 children with typical development. Findings revealed an ERP pattern that was different than the adult response in both 5- and 6-year-olds, though 6-year-olds demonstrated a transition toward the more mature response. Additionally, verbal working memory abilities were found to be associated with more mature ERP responses. These results indicate that the neural processes supporting rhyme detection are not yet mature in 5-year-olds with significant maturation from age 5 to 6 years. These findings extend understanding of the typical trajectory of the neural responses supporting rhyme to children aged 5 years. In future studies, this trajectory could be used to determine whether rhyme processing is different or delayed among populations with communication or reading disorders, with the potential to inform diagnoses and treatments to improve rhyme and other early literacy skills. ACKNOWLEDGEMENTS A special thank you goes out to all of those involved in the current research project, including Dr. Chris Weber, Barb Brown, The Purdue Stuttering Project, and the Brain Systems for Language Lab. We are also grateful for the families and participants who gave their time to support this potentially life-impacting research. This work is supported by grant funding from NIH-NIDCD DC00559 (PI: Dr. Anne Smith & Dr. Christine Weber). iii TABLE OF CONTENTS LIST OF TABLES LIST OF FIGURES INTRODUCTION Defining Phonological Awareness The Development of Phonological Awareness Phonological Awareness & Reading Rhyme Detection Neurophysiological Evaluations of Rhyme Development of Neural Processes for Real Word Rhyme Development of Neural Processes for Nonword Rhyme Current Study METHOD Participants Behavioral Testing Stimuli Procedure Electroencephalographic Recordings & Measures Statistical Analyses Behavioral Performance Neural Processes for Rhyme Primes. Rhyming & non-rhyming. 100-300 ms. 300-500 ms. 500-700 ms. 700-1000 ms. Rhyming only. Non-rhyming only. Behavioral Performance & ERPs Primes. 100-300 ms. 300-500 ms. 500-700 ms. 700-1000 ms. Rhyme effect. 100-300 ms. 300-500 ms. 500-700 ms. RESULTS iv vi vii 1 2 4 4 7 8 11 13 15 18 18 19 21 22 24 25 27 27 27 27 29 31 33 34 34 35 37 38 38 38 38 38 38 39 39 39 39 700-1000 ms. DISCUSSION Primes Rhyming & Non-rhyming Targets Relationships Between ERPs & Behavioral Tasks Limitations & Future Directions Implications for Practice CONCLUSION APPENDICES APPENDIX A: List One: Prime-Target APPENDIX B: List Two: Prime-Target REFERENCES 39 40 40 41 46 48 49 50 51 52 55 58 v LIST OF TABLES Table 1 Reported Reliability Coefficients for Standardized Assessments Table 2 Demographic Information and Behavioral Scores 21 23 vi LIST OF FIGURES Figure 1 Five- and Six-year-old Response to Primes Figure 2 Prime, Time x A/P, 300-500 ms Figure 3 Five-year-old Response to Rhyming and Non-rhyming Figure 4 Six-year-old Response to Rhyming and Non-rhyming Figure 5 Rhyming and Non-rhyming, Time x Cond x Lat, 100-300 ms Figure 6 Rhyming and Non-rhyming, Time x A/P x Lat, 700-1000 ms Figure 7 Rhyming and Non-rhyming, Time x Cond x A/P x Lat, 300-500 ms 28 29 30 31 33 35 37 vii Literacy skills are crucial to all members of society as they aid in passing along and INTRODUCTION acquiring information. This skill is necessary to be successful in nearly all aspects of life including academics, career advancement, and building social bonds. Though this skill is vitally important, it is estimated that a third of American 4th graders will struggle to achieve adequate literacy skills (National Assessment of Educational Progress, 2015). As literacy skills are critical for academic and long-term success, it is important to understand factors that influence and support the development of reading. Phonological awareness, which is the sensitivity to sound units in spoken language, is a skill that begins to develop prior to formal reading instruction and continues to develop into the early school years. This skill has been identified as a reliable predictor of literacy development (NELP, 2009). One skill involved in phonological awareness, rhyme detection, is predictive of both later developing measures of phonological awareness and literacy skills (e.g., Bryant, Maclean, Bradley, & Crossland, 1990). The early-emergence of rhyme detection skills relative to other measures of phonological awareness makes it an appropriate choice for the analysis of phonological awareness skills, and the development of these skills, in young children. Although rhyme detection has been studied extensively at the behavioral level, revealing the typical emergence of this skill, our current understanding of the neural mechanisms that support the development of this skill is limited. Previous studies of the neural processes for rhyme, assessed using event-related brain potentials (ERPs), have revealed neural processing patterns which are not apparent at the behavioral level. However, these studies have been limited to cross-sectional studies of children aged 6 and older and adults. Thus, there is a need for a more complete understanding of the developmental trajectory of the neural processes supporting rhyme detection. It is important to understand how neural processes for rhyme 1 develop in typically developing children because it has been shown that children with developmental language disorders, such as language impairment (LI) or stuttering, have deficits in rhyme and phonological awareness skills. Determining the typical developmental trajectory is the first step toward understanding what differences may be found in children with communication disorders, which could ultimately inform early assessment as well as intervention programs designed to strengthen these skills. Defining Phonological Awareness Generally, phonological awareness is an awareness of the sound units in spoken language. Many skills have been proposed to reflect the processes involved in phonological awareness. Skills vary depending on the size of the sound units being analyzed, specifically sound units may be at the word level, syllable level, intrasyllable level (i.e., onset and rime), and/or phoneme level. Skills also vary depending on the cognitive operation to be performed on the determined sound unit, including skills involving detection (e.g., detection of a specific sound or rhyming words), manipulation (e.g., swapping initial and final phonemes in a word), blending (e.g., combining individual phonemes), and segmenting (e.g., breaking down a blended unit into its constituent sounds). Though skills such as these have been used to measure phonological awareness, there has been some debate as to the exact nature of phonological awareness and whether some of these skills belong to the broader construct of phonological awareness or reflect distinct sound unit abilities. A notable disagreement is whether awareness of larger sound units such as syllables and intrasyllabic units (i.e., onsets and rimes) should be considered skills of phonological awareness alongside awareness of smaller sound units (i.e., phonemes), which is referred to as phonemic awareness, or whether awareness of larger and smaller sound units are distinct skills. 2 There are several ways that these potential differences have been evaluated, including correlations among proposed phonological awareness measures via exploratory factor analyses (EFA) (Muter, Hulme, Snowling, & Taylor, 1997;Wagner & Torgesen, 1987), correlations between various phonological awareness measures and reading abilities (Bryant et al., 1990; Muter et al., 1997), and confirmatory factor analysis (CFA) of the relationships between factors such as phonological awareness abilities at the word level, syllable level, rime level, and phoneme level (Anthony et al., 2002; Anthony & Lonigan, 2004). Studies utilizing EFA have reported mixed findings regarding the nature of the construct. Findings from Wagner and Torgesen (1987) supported phonological awareness as a single construct encompassing tasks involving both large and small sound units. In contrast, other EFA findings have suggested that awareness of larger sound units is a separate skill from awareness of smaller sound units (Muter et al., 1997). Correlations between specific phonological awareness abilities (i.e., rhyme and phonemic awareness) and reading skills have also yielded mixed findings regarding the nature of phonological awareness. Bryant and colleagues (1990) found that sensitivity to rhyme was correlated with both later developing measures of phonological awareness, like phonemic awareness, and reading abilities. This contrasts with findings by Muter and colleagues (1997), which revealed that phonemic awareness abilities, including identification and deletion of phonemes, was an important predictor of literacy while sensitivity to rhyme was not. Despite mixed findings from studies using EFA and studies examining how factors of phonological awareness correlate with reading abilities, several well-designed studies using CFA indicate that phonological awareness should be considered a single construct which encompasses tasks involving both larger (i.e., words, syllables, and intrasyllabic units) and smaller sound units (i.e., phonemes) (Anthony et al., 2002; Anthony & Lonigan, 2004). These findings support a broad 3 definition of phonological awareness as a single ability which can be reflected by many skills. This has led some to prefer the term phonological sensitivity to phonological awareness, a definition which asserts that awareness of smaller sound units reflects a higher level of sensitivity and awareness of larger sound units reflects a lower level of sensitivity (Stanovich, 1992). The Development of Phonological Awareness Consistent with this broader conceptualization of phonological awareness, it is thought that development of this ability is marked by proficiency in tasks of varying linguistic complexity. Studies have revealed that children first become sensitive to words, then syllables, then intrasyllabic units, and eventually, phonemes (e.g., Anthony, Lonigan, Burgess, Driscoll, & Phillips, 2003). Studies have shown that the progression of phonological awareness skills begins as early as age 3 and continues to develop into the early school years (e.g., Anthony et al., 2003). In addition to better skills with smaller linguistic units with age, the developmental trajectory of phonological awareness is also believed to be guided by the type of cognitive operation required to complete the task (e.g., detection, manipulation, blending, segmenting). Generally, children can detect differences between units before they are able to manipulate sounds, and children are able to blend sound units before they are able to segment them (Anthony et al., 2003). It is also thought that the development of each ability does not occur in strict stages, as was once believed. Instead, children continue to refine existing skills as new skills emerge (Anthony et al., 2003). Phonological Awareness & Reading A strong relationship between phonological awareness abilities and literacy development is well established (for reviews see Wagner & Torgesen, 1987; Castles & Coltheart, 2004; Lonigan & Shanahan, 2009). Phonological awareness is a skill that begins to develop prior to 4 literacy acquisition that can be used to predict literacy outcomes. As one study acknowledges, the establishment of phonological awareness as a factor in literacy development does not discredit other, well-established factors in learning to read and spell, such as knowledge of letters and their corresponding sounds (Castles & Coltheart, 2003), but contributes to our understanding of how sound-based skills may aid in the development of literacy. In a comprehensive review of the relationship between phonological awareness and reading, Wagner and Torgesen (1987) concluded that phonological awareness plays a critical role in learning to read. Longitudinal studies demonstrate a strong relationship between phonological awareness and reading, with numerous studies finding a correlation between the two abilities even after controlling for intelligence (e.g., Bradley & Bryant, 1985; Mann & Liberman, 1984; Stanovich, Cunningham, & Cramer, 1984). However, experimental phonological awareness training studies have revealed mixed findings, with some evidence that phonological awareness training is effective in increasing reading and spelling abilities (e.g., Fox & Routh, 1975; Treiman & Baron, 1983) and other studies finding no effect of phonological awareness training alone on reading and spelling abilities (e.g., Bradley & Bryant, 1985). Despite these contradictory training results, the general conclusion based on the extensive amount of longitudinal evidence is that phonological awareness is a casual factor in learning to read (Wagner & Torgesen, 1987). As outlined in a more recent review, Castles and Coaltheart (2003) also state that the extensive longitudinal evidence supporting the relationship between phonological awareness and literacy is important. However, the authors point out that correlational evidence cannot be used to establish a causal relationship between phonological awareness and reading due to the possibility of a third untested variable. Therefore, the authors also turn to training studies to 5 determine whether a causal relationship between the two skills exists. Like Wagner and Torgesen (1987), Castles and Coaltheart (2003) included training studies with both significant findings and null findings, and ultimately concluded that, although a few studies came close to establishing a causal relationship between the skills (e.g., Lundberg, Frost, & Peterson, 1988; Schneider, Kuespert, Roth, & Vise, 1997), these studies were not able to clearly determine that reading skills were not present or emerging at the time of intervention. Thus, Castles and Coaltheart (2003) concluded that the findings from across training studies indicate that phonological awareness and reading development may not have a direct causal relationship. In 2009, the National Early Literacy Panel (NELP) published a comprehensive meta- analysis of the early literacy development literature, including the exploration of numerous factors which may shape early reading skills. Phonological awareness was found to have a moderate relationship with traditional literacy outcome measures of decoding, reading comprehension, and spelling even when variables such as age, socioeconomic status, alphabet knowledge, oral language, intelligence, and prior coding ability were controlled. This study also identified alphabet knowledge, the knowledge of letter names and their respective sounds, as a key factor in literacy development. Other factors with moderate to strong relationships with literacy outcomes included rapid automatic naming (RAN) of letters and digits, RAN of objects and colors, writing/writing name, and phonological short-term memory, even when several variables were controlled in multivariate analyses. While several factors were identified as predictors of reading skills, this meta-analysis demonstrates that phonological awareness is a moderately strong and reliable predictor of literacy. Longitudinal correlational studies and, arguably, a few training studies have demonstrated that there is indeed a link between phonological awareness skills and learning to 6 read. While numerous factors have been established as predictors of literacy achievement, a strong and well-established relationship exists between phonological awareness and reading. The well-documented relationships between phonological awareness and reading has led to phonological awareness as a target for early reading intervention; improving phonological awareness skills could result in enhanced reading skills. Furthermore, because these skills can be targeted earlier than formal reading instruction, they serve as a target for pre-literacy training and intervention programs. Rhyme Detection One early developing aspect of phonological awareness is the detection of rhyme, a skill present as early as preschool-age (3- to 5-years) (e.g., Bradley & Bryant, 1983; Gathercole, Willis, & Baddley, 1991; Maclean, Bryant, & Bradley, 1987). The detection of rhyme consists of a cognitive operation in which parts of a syllable or word are compared to parts of a different syllable or word. To detect a rhyme, the beginning sound of a word, the onset, is separated from the rest of the word, the rime. Then, the rimes of the two words are compared to determine whether they match. Rhyme detection tasks are used in studies involving phonological awareness for several reasons. First, the ability to perform this task develops in children as young as 3 years old (Maclean, Bryant, & Bradley, 1987). This is earlier than other aspects of phonological awareness, such as those involving sensitivity to individual phonemes (e.g., phoneme tapping), which are difficult for children younger than 5 or 6 years old (Liberman, Shankweiler, Fischer, & Carter, 1974). This relatively earlier emergence of rhyme skills makes rhyme detection an appropriate choice for studies involving young children or studies aiming to assess longitudinal development of phonological awareness skills. Second, rhyme detection tasks are useful because 7 they have been shown to be predictive of later developing measures of phonological awareness, such as phonemic awareness (Bryant et al., 1990; Carroll, Snowling, Stevenson, & Hulme, 2003). Phonemic awareness tasks have been shown to be strongly correlated with letter-sound- correspondence (Muter et al., 1997), a skill which has been identified as critical in learning to read (NELP, 2009). Therefore, rhyme detection skills have been proposed to be important to later developing literacy skills through this indirect relationship. Third, rhyme detection has also been theorized to have a direct influence on reading abilities (Bryant et al., 1990) because it allows children to learn to read by comparing new words with previously acquired words (e.g., learning to reading a new word strand because it has the same rime as familiar words like hand and stand) (Goswami, 1988). Due to the early emergence of rhyme detection, its relationship to later developing measures of phonological awareness, and its connections to literacy achievement, rhyme detection tasks are often deemed an appropriate choice for the study of the development of phonological awareness abilities in young children. Neurophysiological Evaluations of Rhyme One methodology used in the examination of the neural processes underlying rhyme detection is electroencephalography (EEG), with rhyme processing evaluated by event-related brain potentials (ERPs) (Luck, 2014). With this technique, electrical signals from the populations of neuron firing in synchrony are recorded at the level of the scalp. ERPs are EEG that is time-locked to specific stimuli then averaged together. ERPs provide information about neural processes with high temporal resolution, with neural responses occurring within milliseconds of a stimulus. However, ERPs do not have strong spatial resolution, making it difficult to determine from which part of the brain a response originates (Kutas & Federmeier, 2011; Luck, 2014). 8 ERPs are thought to index neural processes involved in specific cognitive tasks (Luck, 2014) and can be used to identify differences between conditions and/or groups of participants (Kutas & Federmeier, 2011). Neurophysiological recordings have potential to detect subtle differences in the processes involved in achieving a given behavior and allow for assessment of at least some aspects of processing involved in a given behavior. For example, using ERPs allows for analysis of word processing even during a passive listening task for which no behavioral response is required. ERPs compared between two different words or two groups can reveal whether different cognitive processing strategies may have been used, such as faster or slower processing or recruitment of more neurons to complete a task. An ERP component previously found to reflect the neural processes underlying rhyme is the N400. The N400 component is characterized by a negative peak around 400 ms after the onset of a stimulus, typically a word (Kutas & Federmeier, 2011). This negativity generally occurs between 200 and 600 ms, is more prominent over centro-parietal sites, and tends to appear in the right hemisphere in adults (Kutas & Federmeier, 2011). The N400 has been elicited by a variety of stimuli and under many different conditions, and is thought to index semantic processing, or processing of word meaning (Kutas & Federmeier, 2011). The amplitude of the N400 reflects the difficulty of semantic integration within a given context, with increased N400 amplitudes reflecting a higher degree of word processing difficulty (Kutas & Federmeier, 2011). In rhyme tasks, larger N400s are typically elicited by non-rhyming targets compared to rhyming targets; this difference is known as the rhyme effect (RE) (e.g. Rugg, 1984). The RE is thought to reflect a phonological mismatch between two words (e.g., Grossi, Coch, Coffey-Corina, Holcomb, & Neville, 2001). 9 The N400 component has been found to change across development. In a study evaluating the development of semantic processing, participants aged 5-25 years were asked to listen to or read sentences that ended with a semantically expected or semantically unexpected word (Holcomb, Coffey, & Neville, 1992). It was found that the N400, which was larger to semantically unexpected words, decreased in amplitude and latency with age. The authors suggest that the trend may be due to the fact that children rely more heavily on the context of a sentence than adults and therefore, require greater recruitment of neurons to process or integrate unexpected information, reflected by the larger amplitude. In another study, participants between the ages of 6 and 13 and adults listened to sentences that were correct, were semantically unexpected, or were syntactically incorrect (Hahne, Eckstein, & Friederici, 2004). ERPs elicited by the semantic condition revealed larger N400s elicited by semantically unexpected sentences in both adults and children. Additionally, similar to the study by Holcomb and colleagues (1992), the latency of the N400 was related to age, such that the 7-to-8-year-old group showed a later N400 than groups of children aged 10 and 13 years; the older children showed N400 latencies similar to those of adults. However, decreased N400 amplitudes with age, as seen in an earlier study (Holcomb et al., 1992) were not observed. The authors proposed that this difference was a result of the type of sentences used in the study, which provided less contextual information than the sentences in the other study. ERP studies of rhyme typically involve reading or listening to pairs of words and determining whether the target word rhymes with the prime, or first word. Both words in the pair elicit an N400, indexing processing of the word. A larger N400 amplitude is typically elicited by non-rhyming compared to rhyming target words and peaks around 400 ms in adults. This difference is known as the rhyme effect, which is most prominent over the right hemisphere 10 in adults (Rugg, 1984). The RE is thought to index the neural processes supporting rhyme judgments, with the increased N400 amplitude elicited by non-rhyming targets indicating that target words that do not rhyme require a higher degree of, or more effortful processing (Grossi et al., 2001). The rhyme effect has been found to develop in patterns consistent with the development of the N400 observe for semantic processing tasks (Hahne et al., 2004; Holcomb et al., 1992). Grossi and colleagues (2001) found that while the N400 is elicited by both rhyming and non- rhyming targets (with larger amplitudes elicited by non-rhyming targets) in children, the N400 elicited by rhyming targets is small or absent in adults. However, several aspects of the RE including the onset, size, and distribution have been shown to be stable and adult-like in children as young as 6 years old (Coch, Grossi, Coffey-Corina, Holcomb, Neville, 2002; Coch, Grossi, Skendzel, & Neville, 2005; Grossi et al., 2001). Development of Neural Processes for Real Word Rhyme The development of neural processes for rhyme has been explored using cross-sectional analyses of real word rhyme processing in both the visual (Grossi et al., 2001) and auditory domains (Coch et al., 2002). In a study of visual real word rhyme processing, participants, who were native speakers of English, were divided into separate age groups: 7-8, 9-10, 11-12, 13-14, 15-16, 17-18, 19-20, and 21-23. ERPs were recorded while participants read a list of 70 word pairs deemed appropriate for their group reading level and judge if the pairs of words rhymed or not, indicating their response by pressing a “yes” or “no” button. Adults (aged 18 or older) exhibited a significant N400 elicited by non-rhyming targets but not rhyming targets, a significant RE. The RE was most prominent across posterior sites over the right hemisphere. The mean amplitude of the RE was not found to correlate with 11 reading or spelling scores, in contrast to predictions. The researchers speculated this may be due to a lack of sensitivity to phonological awareness in the behavioral measures used (Grossi et al., 2001). This cross-sectional study revealed important findings regarding development of neural processes involved in visual real word rhyme. The characteristic RE was present in both children and adults. However, in children under the age of 18, an N400 was elicited by both rhyming and non-rhyming targets, with a smaller N400 elicited by rhyming targets, while in adults, the N400 was present only for the non-rhyming targets (Grossi et al., 2001). This was interpreted to be consistent with previous findings of the development of the N400, which suggest that its amplitude decreases with age (Holcomb et al., 1992). There were also no significant differences between the RE distribution or the timing of the onset of the RE between any of the age groups. This was interpreted as showing that adult-like processes for rhyme are present in children as young as 7 years. Additionally, the onset did not vary with age, which was in contrast to findings from a separate study evaluating semantic processing in the same participants (Holcomb et al., 1992). The authors concluded that the neural processes for rhyme observed in this study were specific to phonological processing, and distinct from the neural processes observed in a study of semantics. Another cross-sectional study utilized a similar paradigm as Grossi and colleagues (2001) to investigate the neural processes of real word rhyme in the auditory domain (Coch et al., 2002). In this study, participants were monolingual native speakers of English and were divided into separate age groups: 7-8, 9-10, 11-12, 13-14, 15-16, 17-18, 19-20, and 21 years and older. ERPs were recorded while participants listened to 70 pairs of words, half of which rhymed and half of which did not, and indicate if each pair rhymed or not by pressing the designated button. 12 This auditory rhyme study also found the RE, as seen in previous studies. Though the auditory RE was largely similar to the RE demonstrated in studies of visual rhyme (Grossi et al., 2001; Rugg, 1984), the auditory RE was found to have a bilateral distribution, whereas the visual RE was more prominent over the right hemisphere. In addition to the posterior RE, which was consistent with previous studies, the auditory rhyme study also reported a left anterior N400 that was larger for rhyming compared to non-rhyming targets, a pattern not observed in visual rhyme studies. Neither the posterior RE nor the anterior reversed RE were found to be correlated with behavioral measures reading, reaction time, and accuracy on the ERP task. The researchers suggested that this may be due to the fact that the link between behavioral measures of rhyme and reading have been established most strongly in children younger than those assessed in this study (e.g., Bryant et al., 1990). The cross-sectional nature of this study provided important findings about the development of the neural indices for auditory rhyme. Coch et al. (2002) found that neither the classic posterior RE nor the anterior reversed RE varied in amplitude, distribution, or latency across age groups, indicating the presence of adult-like neural processes for rhyme in children as young as 7 years in the auditory modality. Development of Neural Processes for Nonword Rhyme Nonword rhyme pairs have also been utilized in the study of neural processes for rhyme in an effort to reduce the semantic context of rhyme stimuli. Studies exploring the neural responses elicited by real words and nonwords have revealed that, though semantic and phonological processing are indeed distinct from one another (e.g., Henson, 2001; Simos, Brier, Fletcher, Foorman, Castillo, 2002; Xu et al., 2001), they overlap in timing and amplitude in ERP measures (Perrin & Garcia-Larrea, 2003). Perrin & Garcia-Larrea (2003) revealed that semantic 13 processing elicits a more robust neural response than phonological processing, and thus, semantic processing overshadows the neural responses elicited by phonological processing. In an effort to reduce the overlap of these processes and measure phonological processes in a more isolated way, nonwords are often utilized. Nonwords are typically designed to mimic the phonological forms of real words such that they are pronounceable for speakers of the language upon which the nonwords are based. However, the defining characteristic of a nonword is that it does not hold meaning. Thus, researchers often use nonwords as a method of measuring the ways in which the brain processes phonological information with less overshadowing by semantic processing. A cross-sectional study of auditory rhyme investigated the neural processes underlying nonword rhyme (Coch et al., 2005). In this study, ERPs were acquired from adults and 6-, 7-, and 8-year-old children while they listen to 88 pairs of nonwords, half of which rhymed (e.g., grood-bood) and half of which did not (e.g., ked- voo). Participants gave verbal responses indicating whether the word pairs rhymed. In the nonword rhyme task, a posterior RE was elicited, consistent with previous findings in real word rhyme tasks. However, the onset of the RE in this study was found to be later than found in real word rhyme studies (Coch et al., 2002), which the authors interpreted as nonword rhyme being a more challenging task. Additionally, like Coch and colleagues (2002), the nonword rhyme study reported the finding of an anterior reversed RE, though in this study, the left hemisphere distribution of this effect elicited by the real word task was not observed. The authors postulated that the left-greater-than-right asymmetry seen in the real word task may reflect semantic processing, a process not required in nonword rhyme (Coch et al., 2005). 14 As this study looked at nonword rhyme processing in both school-aged children and adults, it allowed for some comparison across development. This study differed from the previous studies of neural development of rhyme because it included younger participants than previously assessed, children aged 6 years. The auditory nonword rhyme study found no difference between the RE in the 6-year-olds compared to the older groups. Further, consistent with previous studies (Coch et al., 2002; Grossi et al., 2001), the RE observed in this study evidenced no change in the direction, amplitude, distribution, or peak latency with age, indicating that this effect is adult-like as early as six years of age. One key difference in this study was that the onset of the RE was later than previously observed and the timing of the RE onset also varied with age. The onset of the RE in adults was at 300 ms, in 8-year-olds it was at 480 ms, and in 6- and 7-year-olds it was at 360 ms. Further, when children were divided into high and low groups based on performance on a standardized assessment of phonological awareness, the onset of the RE was found to be later in children who scored lower on these measures. This suggested that the onset of the RE may be sensitive to measures of phonological awareness, a finding not observed in previous studies (Coch et al., 2002; Grossi et al., 2001). Current Study Cross-sectional studies in both the visual and auditory modalities reveal that the rhyme effect is present and stable by age 6 years. This suggests that the development of neural processes underlying rhyme occurs at an earlier age, potentially between the ages of 3 and 5 years. To date, understanding of the neural processes underlying rhyme prior to age 6 is limited. Information about the processes involved in phonological awareness during the early childhood years is especially important because this skill is foundational for later developing literacy skills. Studying the neural mechanisms involved in rhyme, by using words with limited semantic 15 context, may reveal subtle patterns in the processes involved in phonological awareness which are not apparent at the behavioral level. Thus, neural measurements have the potential to provide a more sensitive means of assessing children for early phonological awareness deficits, marked by atypical or immature neural processing of rhyme, which may lead to later literacy failure. More accurate and earlier identification of such deficits will inform treatment decisions and yield better treatment outcomes. Furthermore, understanding of the development of neural processes for rhyme have been limited to cross-sectional studies of children aged 6 years and older. A longitudinal assessment of this skill, including children younger than previously assessed, will add to the existing developmental trajectory of rhyme processing. This will allow for future comparisons of rhyme processing between varied groups of children (e.g., children who stutter) to determine if processing of phonological information is immature or atypical in disordered populations. Such information may potentially help researchers identify causes for such disorders and/or devise intervention strategies. The current study aimed to assess the developmental changes in the neural processes supporting rhyme between ages 5 and 6 years. Based on previous findings from an auditory nonword rhyme study (Coch et al., 2005), it was hypothesized that the RE would be adult-like in the 6-year-old children. However, based on the behavioral literature, which indicates that the ability to detect rhyme begins sometime during the preschool years (Bradley & Bryant, 1983; Gathercole et al., 1991), it was hypothesized that the RE would be reduced in 5- compared to 6- year-olds, which would indicate less mature rhyme processing at age 5. Additionally, previous findings reported that the onset of the RE varied with age, with the onset occurring later in children than adults (Coch et al., 2005). Thus, we hypothesized that the onset of the RE will occur later for children at age 5 than at age 6. Additionally, previous studies found reduced 16 N400 amplitudes with age (Hahne et al., 2004; Holcomb et al., 1992). Although this has not been reported in previous rhyme studies, we hypothesized that children will exhibit larger overall N400 amplitudes at younger compared to older ages. The current study also aimed to evaluate relationships between the RE and performance on behavioral tasks reflecting verbal working memory skills and rhyme abilities, including forward digit span, word span, nonword repetition, and rhyme discrimination. Previous studies suggested that the absence of an established relationship between the RE and phonological awareness measures may be due to the fact that this relationship exists behaviorally at a younger age (Coch et al., 2002; Grossi et al., 2001). Therefore, we hypothesized that a relationship between performance on a rhyme discrimination task and the RE would be present in 5-year-olds but not observed in 6-year-olds. Additionally, we hypothesized that nonword repetition abilities and verbal working memory skills would correlate with the RE in both 5- and 6-year-olds, as rhyme discrimination requires phonological working memory abilities. 17 Participants METHOD Participants in the current study included 12 typically developing children (7 males) with electrophysiological data recordings at both ages 5 (Mean: 5.3 years) and 6 years (Mean: 6.4 years) (for a summary of participant demographic information see Table 2). All children participated in a larger study conducted at Purdue University and the University of Iowa and were followed longitudinally, completing testing at one-year increments. Thus, all children were evaluated at ages 5 and 6 years. This larger study focused on the development of stuttering as well as contributing factors such as language and motor speech abilities, and included both children who stutter and children with no history of stuttering. All participants were native monolingual speakers of English with no known history of neurological impairment or injury. To be included in the study, all children were required to pass a bilateral hearing screening at 1000, 2000, and 4000 Hz at 20 dBHL and to have normal or corrected-to-normal vision. All children exhibited social skills within the typical range and did not exhibit behaviors associated with autism spectrum disorder, as measured by the Child Autism Rating Scale. The Hollingshead Four Factor Index of Social Status (1975), which includes factors of maternal/paternal educational achievement and maternal/paternal occupational prestige, was used to calculate the socioeconomic status (SES). SES is calculated by a combination of highest education and occupation level by mother and father. For education, highest education of high school graduate receives a score of 4, partial college a score of 5, college graduate a score of 6, and graduate degree a score of 7. Calculated SES scores can range from 8-66, with 8-19 reflecting lower SES and 55-66 reflecting higher SES. For the current participants, mean maternal and paternal education levels were 6.4 (.26) and 5.9 (.31), respectively, and the mean 18 SES was 48.8 (3.64). These scores indicate that on average, the participants came from higher SES families. Behavioral Testing All children completed a comprehensive battery of behavioral tasks. The battery included the following assessments. The Columbia Mental Maturity Scale (CMMS; Blum, Burgemeister, & Lorge, 1972) was used to assess non-verbal intelligence (for a summary of scores see Table 2). The CMMS demonstrates evidence of high internal consistency, test-retest reliability, and is relatively error- free (for all assessment reliability correlation coefficients see Table 1). The CMMS also cites evidence of validity based upon correlations with other assessments including the 1964 Stanford Achievement Test (SAT) (r= .57-.47), the Otis-Lennon Mental Ability Test (r= .62-.69), and the Stanford-Binet Intelligence Scale (r= .67). The Test for Auditory Comprehension of Language-third edition (TACL-3; Carrow- Woolfolk, 1998) was used to measure participant receptive language abilities. The TACL-3 demonstrates internal consistency in the optimal range, low standard error of measurements (indicating a high degree of test reliability), strong stability over time, and meets a high level of interrater reliability. The Structured Photographic Expressive Language Test-third edition (SPELT-3; Dawson & Stout, 2003) was utilized to assess expressive language. The SPELT-3 reports evidence of test-retest reliability, interrater reliability, and internal consistency. To ensure high content validity and concurrent validity were achieved, the assessment items and scores on this assessment were compared to widely-used measures of the same construct including the Index of Productive Syntax (IPSyn; Scarborough, 1900) and the Comprehensive Assessment of Spoken 19 Language (CASL; Carrow-Woolfolk, 1999). Additionally, the assessment is reported to have achieved high construct validity as the means and standard deviations are shown to reflect a typical language development course, with test scores increasing with age, standard deviations decreasing with age, and rate of score increase greater in younger children. The Bankson-Bernthal Test of Phonology (BBTOP; Bankson & Bernthal, 1990) was used to assess articulation and phonological processes. The BBTOP also provides evidence of reliability including measures of internal consistency, standard error of measurement, stability, and interrater agreement. The content and construct validity of the BBTOP are also adequate, as the assessment was designed based upon typical development of speech sounds and measures of correlations among subtest scores are moderately high to very high. The Test of Auditory Processing Skills- third edition (TAPS-3; Martin & Brownell, 2005) was used to test participant auditory verbal working memory. The TAPS-3 reports moderate to high internal consistency measures, sufficient stability of scores over time, and standard error of measurement scores reflecting little error. The TAPS-3 also reports adequate content, construct, and criterion validity based upon comparisons to previous editions of the assessment, comparison of scores to development and cognition, and careful selection/analysis of test items. All children included in the current study had nonverbal intelligence, receptive and expressive language, speech sound production, and phonological abilities within the normal range on the measures administered. Additionally, participants were determined to be able to discriminate between rhyming and non-rhyming real words based on performance on a subtest of the Phonological Awareness Test- second edition (PAT-2; Robertson & Saltzer, 2007) (for a summary of scores see Table 2). This assessment reports evidence of test-retest reliability and internal consistency for both the 20 rhyme discrimination task alone and phonological awareness overall, interrater reliability, and standard error of measurements. The PAT-2 demonstrates validity because test items were designed to reflect phonological awareness skills necessary for children and because it demonstrates a strong ability to identify children struggling to read from children with typical literacy development. Table 1 Reported Reliability Coefficients for Standardized Assessments Assessment Test-retest Reliability Internal Consistency Temporal Stability Interrater Reliability Columbia Mental Maturity Scale .85 .90 Test for Auditory Comprehension of Language- third edition .91-.94 .86-.97 .95 Structured Photographic Expressive Language Test- third edition .94 .86 Bankson-Bernthal Test of Phonology .92-.98 Test of Auditory Processing Skills- third edition .96 .69-.94 Phonological Awareness Test- second editiona .63, .90 .70, .96 .97-.99 .85-.99 .97 aReliability coefficients reported for the PAT-2 included those for the rhyme discrimination subtest and overall phonological awareness scores, respectively. Table 1. Test reliability data for the standardized assessments administered. Stimuli All children completed an ERP paradigm designed by Coch et al. (2005) and previously presented in study of older children (Mohan & Weber, 2015) which utilized words containing no 21 semantic content. The paradigm consisted of 88 pairs of pronounceable nonwords, with half of the nonword pairs rhyming (e.g., demp-semp) and half non-rhyming (e.g., bry-pag). All nonwords were one syllable, except for one pair, and their structure was roughly based upon real word rhyme pairs used in previous studies (Coch et al., 2002; Grossi et al., 2001). Non-rhyming pairs consisted of the prime from one rhyming pair with the target from a different rhyming pair. Additionally, the word pairs were divided into two separate lists, each with 44 pairs of rhyming and 44 pairs of non-rhyming, and counterbalanced across participants (see Appendix for complete lists of stimuli). The stimuli were recorded by a female, native speaker of American- English. Each nonword was digitized using PRAAT software, with a recording sampling rate of 22.5 kHz. The average length of each nonword stimulus was 582.4 ms (SD = 72.1). Procedure All children came in for multiple sessions each year, with behavioral and ERP data acquisition occurring on separate days. Behavioral testing was administered by a certified speech-language pathologist (SLP) across two 1.5-hour sessions. The SLP had many years of experience in conducting standardized behavioral assessments in clinical practice and research settings. To measure ERP data, the child was fitted with an electrode cap designed to record neural responses from the brain. The electrode cap was secured on the child’s head and prepared while s/he played video games or watched a video. Stimuli were then presented auditorily via headphones in a sound-attenuating booth. An experimenter sat in the booth with the child to monitor movements and provide the child with periodic reinforcement. The child was asked to listen to the words and determine if the words rhymed or did not. Ten times throughout the paradigm (approximately every 7-9 trials) the experimenter asked the child, “Did those two rhyme?” and recorded the child’s verbal response before proceeding to the next trial. The total 22 number of correctly identified word pairs (as rhyming or non-rhyming) are reported in Table 2. Children completed the same task but with different lists of rhyming/non-rhyming pairs at ages 5 and 6 years. Table 2 Demographic Information and Behavioral Scores Age 5 Age 6 Mean SE Mean SE Age (years) 5.3 .07 6.4 .08 ERP Booth Scoresa 8.5 .52c 9.3 .47d Rhyme Discrimination Scores (PAT-2)b 9.5 .19 9.4 .29 Maternal Education 6.4 .26 Paternal Education 5.9 .31 Socioeconomic Status 48.8 3.64 Nonverbal IQ (CMMS) 114.8 2.48 Note. Scores of maternal education, paternal education, SES, and nonverbal IQ remained constant between ages 5 and 6. aERP booth score out of 10 possible points. bRhyme discrimination score out of 10 possible points. cAt age 5, there is one child missing ERP booth score dAt age 6, there are two children missing ERP booth scores Table 2. Summary of demographic information, ERP booth performance, and rhyme discrimination abilities. 23 Electroencephalographic Recordings & Measures ERPs were recorded using 32 Ag-Cl electrodes secured in an elastic Quick-cap. Electrodes were positioned across the scalp, with eight lateral sites (F7/F8, FT/FT8, TP7/TP8, P7/P8), 12 medial sites (FP1/FP2, F3/F4, FC3/FC4, CP3/CP4, P3/P4, O1/O2), and six midline sites (FZ, FCZ, CZ, CPZ, PZ, OZ). Electrodes were placed above and below the left eye to monitor for vertical eye movements and were also placed near the left and right outer canthi to monitor for horizontal movements. Additionally, electrodes placed over the left and right mastoids served as a linked reference electrode. Data were recorded from 0.1-100 Hz with a 512 Hz sampling rate. Offline, data were down sampled to 256 Hz and low-pass filtered at 40 Hz. EEGLAB and ERPLAB was used to analyze the EEG data. Independent component analysis (ICA), a data decomposition tool, was used to identify and remove vertical and horizontal eye movement artifact present in the continuous EEG data. Two graduate research assistants, trained by the Michigan State University Brain Systems for Language (BSL) Laboratory director, independently determined which components to remove. In the event of conflicting judgments regarding which components to remove, the BSL Lab director completed an independent rating and any remaining discrepancies were resolved between the BSL Lab director and the first author. To measure and average the ERPs, the data were epoched from 100 ms prior to the onset of the stimulus to 1000 ms after stimulus onset. An automatic artifact rejection procedure was used to detect and remove high voltage artifact. This procedure automatically marks eye artifact greater than ± 100 µV and artifact across other sites greater than ± 200 µV for rejection. A second step of manual artifact identification and rejection occurred to ensure all eye and movement artifact and voltage drifts were eliminated. EEG data was time- locked to both the prime and targets words. All primes were averaged together because at the 24 time of the presentation of the prime, the children did not know whether the following word would rhyme or not. Rhyming and non-rhyming targets were averaged separately. Comparisons between neural responses to primes, rhyming targets, and non-rhyming targets provided data regarding the neural processes for rhyme. Statistical Analyses A paired t-test with a within-subject factor of time (age 5/age 6) was used to evaluate changes in performance on behavioral measures including ERP booth rhyme detection scores, real word rhyme discrimination, and real word rhyme production. Neurophysiological measures included the mean amplitudes of the ERPs elicited by prime and target words within specified time windows. The mean amplitudes of the ERPs elicited by prime and target stimuli were measured in four separate time windows, including measures from 100-300 ms, 300-500 ms, 500-700ms, and 700-1000 ms post onset. This allowed for measurement of the early, middle, and later aspects of the broad N400 elicited in the current study. A mixed-effects repeated measures analysis of variance (ANOVA) was used to analyze ERP data elicited by the primes and targets. For the N400, the ANOVA included within-subject factors of Time (Time - two levels: age 5, age 6) and Condition (Cond - two levels: rhyming and non-rhyming), along with factors of scalp distribution, including Hemisphere (Hemi - two levels: left and right), Anterior-Posterior (A/P - five levels: frontal, anterior temporal, temporal, parietal, and occipital), and Laterality (Lat - two levels: lateral, medial). Alpha was set at p < .05. For all within-subject factors with more than two degrees of freedom, the Huynh-Feldt correction was utilized. Partial-eta squared (np 2), an index of effects size, was reported for all significant effects. 25 To evaluate relationships between behavioral rhyme performance and the neural processes supporting rhyme, Pearson’s correlations were employed to evaluate relationships between behavioral performance on the rhyme task and ERP measures. Difference waves (the RE) were computed by subtracting the N400 elicited by rhyming targets from the N400 elicited by non-rhyming targets. Composite ERP measures were calculated for both the primes and the RE across electrode sites at which these components were most prominent, specifically by taking the mean amplitude across the following electrode sites: F3, FC3, C3, CP3, P3, F4, FC4, C4, and CP4. Then, the composite ERP measure was correlated with performance on the forward digit span and word span tasks from Test of Auditory Processing Skills- third edition (TAPS-3; Martin & Brownell, 2005), nonword repetition as measured by the Dollaghan and Compbell Nonword Repetition Task (Dollaghan & Campbell, 1998), and the rhyme discrimination subtest of the Phonological Awareness Test- second edition (PAT-2; Robertson & Saltzer, 2007). 26 Behavioral Performance RESULTS Behavioral ERP performance was high at age 5, average 8.5 (SE .52) and age 6, average 9.3 (SE .47). Hence, there was no significant difference in ERP booth performance between age groups. Additionally, it was found that there were no significant differences between performance on rhyme detection and production tasks at age 5 and performance on these tasks at age 6. Neural Processes for Rhyme Primes. Grand average ERPs elicited by prime words at age 5 and 6 are illustrated in Figure 1. No significant effects or interactions were observed for the 100-300 ms or the later 500-700 ms time windows. For the 300-500 ms time window, analysis of the mean amplitudes of ERP responses to the primes revealed a trend toward larger N400 mean amplitudes over centro-parietal and parietal sites in 6-year-olds compared to 5-year-olds (Time x A/P: F(1,11) = 3.146, p = .071, p 2 = .222, (Fig. 2). This interaction was also significant in the 700-1000 ms time window (Time x A/P: F(1,11) = 3.611, p = .041, p 2 = .247), again with 6-year-olds exhibiting a more negative response to primes over centro-parietal and parietal sites compared to 5-year-olds. In the late time window (700-1000 ms), a larger negativity was elicited by primes over parietal sites in the left hemisphere in 6- compared to 5-year-olds (Time x Hemi x A/P: F(1,11) = 2.843, p = .054, p 2 = .205). 27 Figure 1 Five- and Six-year-old Response to Primes Figure 1. Grand average ERPs elicited by the prime words in 5- and 6-year-olds. N400 amplitudes elicited by primes were larger in 5-year-olds than 6-year-olds. Note that negativity is plotted upward in this and all subsequent ERP plots. 28 Figure 2 Prime, Time x A/P, 300-500 ms Figure 2. A significant interaction of time by anterior-posterior scalp distribution in the 300-500 ms time window revealed larger N400 mean amplitudes in response to primes over centro- parietal and parietal sites in 6-year-olds compared to 5-year-olds. Note that negativity is plotted upwards in this and all subsequent plots. Rhyming & non-rhyming. Grand average ERPs elicited by rhyming and non-rhyming targets for 5- and 6-year-olds are illustrated in Figures 3 and 4, respectively. 29 Prime300-500 msTime x A/PFrontalFronto-centralCentralCentro-parietalParietalMean Amplitude mV-2-101235-year-olds6-year-olds Figure 3 Five-year-old Response to Rhyming and Non-rhyming Figure 3. Grand average ERPs elicited by rhyming and non-rhyming target words in 5-year-olds reveal a broadly distributed reversed RE pattern. 30 Figure 4 Six-year-old Response to Rhyming and Non-rhyming Figure 4. Grand average ERPs elicited by rhyming and non-rhyming target words in 6-year-olds reveal a transition toward the RE over mid-lateral sites. 100-300 ms. Analyses comparing the ERP mean amplitudes elicited by rhyming and non-rhyming conditions revealed a trend toward 5-year-olds exhibiting a larger N400 amplitude elicited by rhyming targets than non-rhyming targets (a reversed RE), which was more pronounced over anterior sites, while the reversed RE was not present in the 6-year-olds in this 31 time window (Time x Cond x Hemi x A/P: F(1,11) = 2.504, p = .088, p 2 = .185). Another interaction trending toward significance found in this time window was characterized by an effect of time by condition by lateral scalp distribution, (Time x Cond x Lat: F(1,11) = 4.117, p = .067, p 2 = .272). This interaction reflected a trend toward a reversed RE across both lateral and mid-lateral sites in 5-year-olds, with a smaller N400 amplitude over mid-lateral sites elicited by non-rhyming targets. In contrast, 6-year-olds exhibited N400 responses to rhyming and non- rhyming target that were similar in amplitude, with slightly larger N400 amplitudes elicited by non-rhyming targets, indicating a possible transition toward a typical RE pattern over mid-lateral sites (Fig. 5). 32 Figure 5 Rhyming and Non-rhyming, Time x Cond x Lat, 100-300 ms Figure 5. A significant interaction of time by condition by laterality in the 100-300 ms time window revealed a transition toward the RE over mid-lateral sites in 6-year-olds. 300-500 ms. In this time window, another trend existed between the 5- and 6-year-olds involving time by condition by hemisphere by anterior-posterior scalp distribution, F(1,11) = 3.037, p = .062, p 2 = .216. This trend suggested a reversed RE is present over central sites in the left hemisphere of 5-year-olds, whereas this effect is occurring across several sites in both the left and right hemispheres in the 6-year-olds. Another trend observed revealed larger overall N400 mean amplitudes across both conditions over frontal and fronto-central mid-lateral sites in 5- compared to 6-year-olds; N400 amplitudes over lateral sites were similar between age groups (Time x A/P x Lat: F(1,11) = 2.445, p = .081, p 2 = .182). 33 Rhyming & Non-rhyming100-300 msTime x Cond x Lat5-year-olds L ML L MLMean Amplitudes mV-1012345Rhyming Non-rhyming6-year-olds 500-700 ms. A significant interaction present in this time window revealed greater differences in N400 amplitudes across both conditions between lateral and mid-lateral electrode sites, especially over centro-parietal and parietal regions in 5-year-olds as well as larger N400 amplitudes over frontal mid-lateral sites in 6-year-olds (Time x A/P x Lat: F(1,11) = 4.106, p = .014, p 2 = .272). 700-1000 ms. In the late time window, a significant interaction revealed a prolonged N400 response over frontal, fronto-central, and central lateral sites in 5- compared to 6-year-olds, (Time x A/P x Lat: F(1,11) = 2.710, p = .045, p 2 = .198) (Fig. 6). 34 Figure 6 Rhyming and Non-rhyming, Time x A/P x Lat, 700-1000 ms Figure 6. A significant interaction of time by anterior-posterior scalp distribution in the 700- 1000 ms time window revealed that the N400 is prolonged over frontal, fronto-central, and central lateral sites in 5- compared to 6-year-olds. Rhyming only. In order to determine whether changes occurred independently in neural responses to rhyming and non-rhyming targets separately, step-down ANOVAs were conducted for each condition (rhyming and non-rhyming). In the early time window (100-300 ms), 5-year- olds tended to exhibit larger N400 amplitudes elicited by rhyming targets over the left hemisphere compared to 6-year-olds (Time x Hemi x A/P x Lat: F(1,11) = 2.328, p = .071, p 2 = 35 .175). In the 300-500 ms, 500-700 ms, and 700-1000 ms time windows, significant interactions of time, anterior-posterior scalp distribution, and laterality reveal larger N400 amplitudes elicited by rhyming targets over mid-lateral sites in 5- compared to 6-year-olds as well as more variability in N400 amplitudes over lateral and mid-lateral sites in 5-year-olds than 6-year-olds (F(1,11) = 2.710, p = .046, p 2 = .198 (Fig. 7); F(1,11) = 2.710, p = .005, p 2 = .198; and F(1,11) = 2.710, p = .014, p 2 = .198, respectively). 36 Figure 7 Rhyming and Non-rhyming, Time x Cond x A/P x Lat, 300-500 ms Figure 7. An interaction of time by anterior-posterior scalp distribution by laterality revealed that N400 amplitudes were larger in 5-year-olds, especially over mid-lateral sites. Although the interaction was significant across both conditions, rhyming and non-rhyming conditions were plotted separately in order to also illustrate significant interaction of time by anterior-posterior by laterality for the rhyming condition only. Non-rhyming only. Step-down ANOVAs for the non-rhyming condition in the 100-300 ms time window revealed that 5-year-olds exhibited larger N400 amplitudes to non-rhyming 37 targets over lateral sites whereas 6-year-olds exhibited slightly larger N400 amplitudes to non- rhyming targets over mid-lateral sites (Time x Lat: F(1,11) = 4.011, p = .07, p 2 = .267). A similar interaction was observed for the later 700-1000 ms time window, with 5-year-olds exhibited larger N400 amplitudes over left and right hemisphere lateral sites. In contrast, 6-year- olds exhibited larger N400 amplitudes over left hemisphere lateral compared to mid-lateral site, but slightly larger N400s over right hemisphere mid-lateral compared to lateral sites (time by hemisphere by laterality, F(1,11) = 2.710, p = .033, p 2 = .350). No significant interactions were present in the 300-500 ms or 500-700 ms time windows. Behavioral Performance & ERPs Primes. 100-300 ms. In the early time window, a significant correlation was found between the ERP response to primes in 5-year-olds and performance on the nonword repetition task at age 5 (r = .551, p =.049). A significant correlation between the ERP response to primes in 6-year-olds and 6-year-old performance on an auditory word memory task was also observed (r = .579, p =.040). 300-500 ms. There were no significant correlations between the ERP response to primes and behavioral performance were observed in this time window. 500-700 ms. In this time window, a significant correlation was observed between the ERP response to primes in 5-year-olds and nonword repetition scores at age 6 (r = -.625, p =.027). 700-1000 ms. There were no significant correlations between the ERP response to primes and behavioral performance were observed in this time window. 38 Rhyme effect. 100-300 ms. In the early time window, significant correlations were observed between the RE in 6-year-olds and the auditory number memory task at age 6 (r = -.604, p =.032). The RE at age 6 was also correlated with nonword repetition scores at age 5 (r = -.782, p =.004) and at age 6 (r = -.675, p =.016). 300-500 ms. The RE at age 6 was also correlated with nonword repetition scores at age 5 (r = -.600, p =.033) and age 6 (r = - .792, p =.003) in this time window. 500-700 ms. A significant correlation was observed between the RE at age 6 and performance on an auditory word memory task at age 6 (r = -.151, p =.032). The RE at age 6 was also correlated with performance on a real word rhyme discrimination task at age 5 (r = - .581, p =.039) and nonword repetition task at age 6 (r = -.683, p =.015). 700-1000 ms. In the latest time window, the RE at age 6 was correlated with performance on the rhyme discrimination task at age 5 (r = -.634, p =.025), the auditory number memory task at age 5 (r = -.718, p =.010), and the auditory number memory task at age 6 (r = - .745, p =.007). 39 DISCUSSION The current study aimed to assess the development of neural processes for rhyme from ages 5 to 6 years. Our findings provide evidence that there are differences in rhyme processing in 5-year-olds when compared to 6-year-olds. For prime words, differences in scalp distribution of N400 components were observed from age 5 to 6 years, characterized by a more mature, focal N400 response in 6-year-olds. Children also exhibited maturation in the neural response to target words, as evidenced by a shift from a reverse rhyme effect toward a rhyme effect from age 5 to 6 years. The 6-year-olds also exhibited smaller amplitude and shorter duration N400 responses than 5-year-olds, largely driven by the N400 responses elicited by rhyming targets. Together, these findings reveal maturation of neural processes for rhyme from age 5 to 6 years in children with typical development, refining and extending our understanding of the developmental trajectory of neural processes supporting rhyming in young children. Primes The N400 component was observed in both age groups in response to the prime words. In two separate time windows, differences in the N400 distribution between the two age groups were observed, indicating that the N400 observed in 6-year-olds tended to be more prominent over centro-parietal and parietal sites than the N400 in 5-year-olds (Fig. 2). As previous studies of the N400 have revealed that this component is typically largest over centro-parietal sites in adults (e.g., Kutas & Federmeier, 2011), the current findings suggest that 6-year-olds are showing a more mature N400 pattern than 5-year-olds. This more focal response compared to the broad scalp distribution observed in 5-year-olds is hypothesized to reflect a more focal, less distributed neural processing system for words, which is thought to indicate more efficient, mature word processing. These findings show evidence of the development of neural processes 40 for nonwords, with a transition toward a more efficient neural processing system between the ages of 5 and 6 years. Though previous studies have established that maturation of the N400 occurs between 5 and 6 years (e.g., Holcombe, 1992), these are the first known longitudinal data to date providing evidence of development in neural responses to nonwords from age 5 to 6 years. Rhyming & Non-rhyming Targets The children in the current study did not demonstrate a significant change in accuracy on the ERP task from age 5 to 6 years. There are several potential reasons as to why significant differences in behavioral performance were not observed between groups. First, there was little variation in the data in general as most children performed at or near ceiling level on the task. This is likely due to the early emergence of rhyme detection abilities and because only children who could reliably make rhyme judgments were selected for participation in this study. Additionally, the participants’ rhyme judgments were only recorded 10 times throughout the ERP task to reduce fatigue and artifact caused by button press movements in young participants. However, a greater number of rhyme judgment trials may have revealed differences in performance from age 5 to age 6. Another potential reason that significant differences were not achieved is that some behavioral data were missing. Though evidence of greater behavioral rhyme performance with age was not observed in the current study, there was evidence of the maturation of neural processes supporting rhyme between ages 5 and 6 years. The N400 response elicited by target words, across both rhyming and non-rhyming conditions, also changes from age 5 to 6 years. In all but the earliest time window, an effect of time by anterior-posterior scalp distribution by laterality revealed a decrease in N400 amplitude with age. In the 300-500 ms time window, 5-year-olds exhibited larger N400s than 6-year-olds, 41 especially over mid-lateral sites. The interaction in the 500-700 ms time window revealed that more variability in N400 mean amplitudes between lateral and mid-lateral sites in 5- compared to 6-year-olds, although 6-year-olds exhibited a larger N400 than 5-year-olds over frontal mid- lateral sites. In the late time window, decreased N400 amplitudes from age 5 to 6 were observed primarily over lateral sites (Fig. 6). Though the specific patterns differed across lateral and mid- lateral sites in different time windows, overall, the N400s observed in 5-year-olds were generally larger and more broadly distributed across the scalp than those in 6-year-olds. This finding is consistent with patterns observed for prime words as well as in previous studies of the N400, which demonstrated decreases in N400 amplitudes with age (Holcomb et al., 1992). This pattern of development is theorized to be indicative of a reduction in the amount of neural resources required to process a word (i.e., less effortful processing; Kutas & Federmeier, 2011). As the current findings show reduced N400 amplitudes with age, they may reflect less effortful word processing in 6- compared to 5-year-olds. Further evidence of the maturation of word processing is provided by the N400 effect in the 700-1000 ms time window. The interaction of time by anterior-posterior scalp distribution by laterality indicates that N400 amplitudes were larger in 5- year-olds than 6-year-olds in the latest time window, a pattern indicating a more prolonged N400 response in 5-year-olds (Fig. 6). This pattern may suggest that 6-year-olds are able to process words more quickly than 5-year-olds, evidenced by a resolved N400 in the late time window in 6-year-olds (Holcomb et al., 1992). To better understand the current findings, and to evaluate development for rhyming and non-rhyming targets separately, step-down ANOVAs were conducted for each condition (rhyming and non-rhyming). These analyses revealed that the interactions of time, anterior- posterior scalp distribution, and laterality across both conditions, described above, were only 42 significant in the step-down analyses of the rhyming condition alone. The presence of these significant interactions for the rhyming condition but not for the non-rhyming condition alone indicate that the N400 amplitudes differences between age groups were largely driven by the N400s elicited by rhyming targets (Fig. 7). Taken together, the step-down analyses for rhyming targets revealed that N400 mean amplitudes elicited by rhyming targets were larger in the 5-year- olds than the 6-year-olds, whereas N400s elicited by non-rhyming targets were relatively stable from 5 to 6 years. The decreasing N400 amplitudes elicited by rhyming targets with time is consistent with the general development of the N400 component (Holcomb et al., 1992). We hypothesized that N400 amplitudes would decrease with age, based on previous findings (Holcomb et al., 1992). Generally, the current findings support this hypothesis, though this change was only for N400s elicited by the rhyming condition. The step-down analysis of non-rhyming targets revealed larger N400 amplitudes over lateral sites in 5-year-olds, whereas 6-year-olds presented with slightly larger N400s amplitudes over mid-lateral sites in the early time window, evidenced by an interaction of time by laterality. These findings suggest a transition to a more mature N400 pattern, with larger N400 amplitudes over mid-lateral sites, from age 5 to 6 years. The typical RE, marked by a larger N400 component elicited by non-rhyming words than rhyming words, was not observed in either age group in the current study. The larger N400 amplitude reflects a greater recruitment of neurons during the processing of a word, indicative of more effortful processing (e.g., Kutas & Federmeier, 2011), and the RE reflects the increased difficulty in processing non-rhyming compared to rhyming target words (e.g., Coch et al., 2005). The current findings reveal that 5- and 6-year-olds in the current study are not using more effortful processing for non-rhyming words, a different pattern than has been established in 43 previous studies of children aged 6 years and older (Coch et al., 2005). In contrast to our hypotheses, the current study provides evidence that mature processes for rhyme may not yet be well-established in at least some 6-year-olds. However, the current findings support, in part, our hypothesis that less mature rhyme processing would be observed in 5-year-olds. Taken together, these findings indicate that, while maturation is observed in neural processes for rhyme from age 5 to 6 in the current study, the typical, mature RE pattern in not yet reliable by age 6 years. Instead of the mature RE, children in the current study exhibited a reversal of the RE, with larger N400 amplitudes elicited by rhyming than non-rhyming target words or similar mean amplitudes elicited by both rhyming and non-rhyming targets. An interaction in the early time window (100-300 ms) involving time, condition, hemisphere, and anterior-posterior scalp distribution revealed that 5-year-olds exhibited a reversed RE across the scalp, whereas 6-year- olds exhibited relatively small differences between conditions across the scalp. This pattern is opposite of the RE observed in older children and adults in previous studies (e.g., Coch et al., 2005). Another interaction in this time window which involved time, condition, and laterality, revealed a reversed RE over lateral and mid-lateral sites in 5-year-olds while 6-year-olds presented with similar N400 responses to rhyming and non-rhyming targets overall. This interaction also revealed that 6-year-olds exhibited slightly larger negativity elicited by non- rhyming targets over mid-lateral sites in this early time window (Fig. 5). This is interpreted as evidence of a transition toward the typical RE in 6-year-olds. In the 300-500 ms time window, an effect of time by condition by hemisphere by anterior-posterior scalp distribution revealed that the reversed RE is present in both age groups, with different scalp distribution patterns for each age group. This interaction also demonstrated that the distribution of the reversed RE across the scalp is more robust in 5- compared to 6-year-olds. A reversed RE has been 44 previously reported for both nonword and real word rhyme discrimination tasks, though primarily over anterior sites (Coch et al., 2005; Coch et al., 2002). In the current study, the reversed RE is broadly distributed across the scalp, especially in 5-year-olds. As this is the first study to date to assess neural processes supporting rhyme in children aged 5 years, this broadly distributed reversed RE pattern is interpreted as a less mature neural process for rhyme. Although both age groups exhibited less mature rhyme processing patterns than previous studies, the current findings suggest that 6-year-olds exhibited rhyme processes that were transitioning toward the mature RE, marked by a smaller reversed RE across the scalp with and slightly larger N400 elicited by non-rhyming targets (a typical RE) over mid-lateral sites in the early time window. Based on previous findings of word processing in toddlers, the current study proposes an explanation for the large negativity seen in both 5- and 6-year-olds in response to rhyming targets. In adults and older children, the reduced amplitude in response to rhyming targets reflects the benefit of priming, easier integration of information when similar to (i.e., rhyming with) the phonological prime word (e.g., Coch et al., 2005). As this pattern was not observed in the ERP responses from the participants in the current study, it appears that both 5- and 6- year- olds are not benefitting from the prime word. In fact, both age groups in the current study are demonstrating an unexpected larger N400 for rhyming targets. A previous study in children aged 13 to 20 months found larger N400-like negativities to comprehended (known) words compared to unknown words. This pattern in toddlers is thought to reflect the processes underlying word learning during a time when language abilities undergo rapid development (Mills, Coffey-Corina, & Neville, 1997). In the current study, the reversed rhyme effect may reflect processing of rhyming targets, indicated by larger N400s elicited by rhyming words, with 45 less neural resources engaged in processing non-rhyming targets, indicated by the smaller N400 elicited by non-rhyming words. Thus, the children in the current study are comprehending or recognizing the target word that rhymes while being less engaged by the target that does not rhyme. Given that this pattern occurs during a period during which rhyming and other phonological awareness skills are emerging and undergoing refinement, this pattern may reflect that at the ages of 5 and 6 years, children do not yet reliably benefit from phonological primes in the same way as older children and adults. Relationships Between ERPs & Behavioral Tasks The composite prime ERP measure was positively correlated with performance on the nonword repetition task and the auditory working memory task for strings of words. These correlations indicate that better performance on each task was associated with a smaller N400, which is thought to reflect more mature word processing skills (e.g., Kutas & Federmeier, 2011). Together, these findings suggest that children who have better auditory verbal working memory skills exhibit more mature neural processes for words. This finding is consistent with previous findings of cognitive processes that support language, as Hampton Wray and Weber-Fox (2013) also found evidence that a more mature N400 response was related to better verbal working memory skills. This study proposed that the ability to make semantic associations improve memory by reducing the cognitive load imposed by remembering a large string of digits or words (Hampton-Wray & Weber-Fox, 2013). This study also suggested that better verbal working memory allows for reduced reliance on semantic processing, which is reflected by the reduction of N400 amplitudes with better verbal working memory abilities (Hampton-Wray & Weber-Fox, 2013). 46 The RE was found to be correlated with performance on several behavioral tasks. The RE difference wave was found to be inversely related to nonword repetition scores, meaning that a more larger response to non-rhyming than rhyming targets, or a bigger, more mature RE pattern, was associated with better, more accurate, nonword repetition skills. This relationship suggests that children who are more proficient in repeating nonwords have more mature neural processes for a nonword rhyme task. Both tasks engage phonological knowledge, phonological working memory, and phonological parsing and combining, skills associated with phonological awareness. These correlations also revealed that nonword repetition task scores at age 5 were associated with the RE at age 6, which indicates that better performance on the nonword repetition task at age 5 may be predictive of more mature ERP patterns for rhyme processing at age 6. The RE was also found to be inversely related with auditory working memory skills for both string of numbers (digit span task) and words. These relationships indicate that stronger auditory working memory skills may be predictive of more mature auditory rhyme processing. In order to detect whether a pair of words rhyme, working memory must be engaged to allow the listener to separate the onset from the rime, hold the rime in their working memory, then compare the rimes of prime and target words to determine whether they match. Therefore, children who are better able to hold the rime in their memory may be better able to compete the rhyme task. The current study also revealed a relationship between the RE and a behavioral rhyme discrimination task. The relationship between the two measures indicates that stronger rhyme discrimination abilities at age 5 were predictive of a more mature RE at age 6. 47 Limitations & Future Directions Though the current findings provide evidence of maturation in rhyme processing abilities from age 5 to 6 years, there are several limitations to the current study. First, the study only involved children from age 5 to 6 years. This limited our ability to assess the maturation of neural processes closer to the age that rhyme detection abilities emerge, around the age of 3 years. Additionally, lack of information regarding neural processes at older ages limits our ability to determine whether the transitioning toward a mature RE observed in the 6-year-olds in the present study results in a mature RE. Future longitudinal studies including both younger (i.e., 3-4 years) and older children (i.e., 7-8 years) in addition to the ages assessed in the current study may provide more complete information about the development of neural processes supporting rhyme. Another limitation of the present study is that children performed at or near ceiling levels on behavioral assessments of rhyme detection, limiting our ability to determine if refinement in performance in these areas was related to a maturation of the neural response. It is possible that differences from age 5 to 6 may have been present in ERP booth performance if participants had been asked to provide responses to every word pair (i.e., 88 pairs) as opposed to only 10 times throughout the task in the current study. Thus, future studies should involve the collection of rhyme judgments for each trial. The SES of the participants was another limitation of the current study. The average SES was relatively high, with most mothers having at least a college degree. This limits our ability to generalize the findings of the present study to a broad population. Future studies should include children from a broader SES range to allow for better generalization of findings regarding neural processes for rhyme. 48 Lastly, power analyses revealed that the statistical power of the current findings are in the lower range. More participants would increase power in the current study. Thus, the current findings should be interpreted as preliminary in nature and future studies should include a greater number of participants. Implications for Practice Future studies that account for the described limitations, especially across broader age ranges, will be able to establish a more complete developmental trajectory for the neural processes for rhyme in children with typical development. This trajectory in typical development could then be used as a benchmark against which rhyme processes in children with language and/or communication disorders, such as children who stutter, children with language impairments, or children with reading difficulties could be compared. This would allow for the determination of whether children with atypical language and/or phonology skills may have delayed maturation or deviant/atypical phonological processes compared to typically developing children. Understanding how the development of neural processes for rhyme may differ in these populations has the potential to inform and help refine rhyme assessments and to provide targets for treatment to improve rhyme and other early literacy skills. 49 CONCLUSION The current longitudinal study revealed maturation in neural processes for rhyme in children with typical development from age 5 to 6 years. These findings extended our understanding of rhyme processing to younger children than previous studies and reveal that the neural processes supporting rhyme are not yet mature in 5-year-olds. Additionally, the study revealed relationships between auditory verbal working skills and neural processes for rhyme. The current findings may serve as pilot data, with future expanded studies having the potential to inform interventions targeting early rhyme and phonological awareness skills. 50 APPENDICES 51 List One: Prime-Target APPENDIX A jite–fauer trin–phy grize–yise nobe–drobe kow–deeb gox–brocks dreat–ged frield–geeled murze–thurze dat–lat yi–marp fum–zi grood–bood ked–voo dabe–lum nilled–dilled clate–pline kile–spile mun–lun shum–hane ky–tate jate–yate vease–meeze doan–pone bry–pag zare–jare crail–lale kun–gree demp–semp crute–doot zeer–pud yocks–toos nake–dake blug–kroar siff–piff pake–spake stam–glig gour–druze daf–coom vore–jore trum–pum 52 maft–yaft drere–vair bro–slore gines–rabe nin–rin floos–cho drig–stug sarp–cly gite–clite chole–thole nool–shull dorde–morde mag–yare gee–blail blauer–flam blane–vox sare–nare poom–dite foo–breet daip–laip fam–cham feap–neap vite–balf bome–slines prail–stobe nef–gef stee–kwee pooze–lauer chuz–luz moce–boce taid–chy throre–slin plol–groll bly–gry rine–clum doode–keer neeb–stide glir–flir poat–hoat blore–plo quo–zow mide–gome slair–jun poe–trow ji–claid 53 plew–snew 54 List Two: Prime-Target APPENDIX B foo–voo plol–neap nin–laip trum–dake bro–plo kile–pone dorde–gry glir–meeze plew–dilled vore–lun gour–lauer doan–lat murze–hoat floos–toos daip–piff daf–balf stee–spile poom–coom trin–slin bry–chy grize–morde feap–vair nobe–kwee jite–zite blug–stug gines–slines blore–slore crute–lale shum–lum chuz–semp nake–trow pake–brocks moce–rin crail–thole jate–yise fam–zow grood–nare siff–jore rine–pline sarp–marp stam–flam 55 zeer–keer blauer–fauer clate–tate nilled–groll pooze–druze mag–pag gite–thurze neeb–deeb chole–pum quo–jare ky–phy gox–luz demp–geeled gee–gree dreat–breet fum–clum bome–gome kow–cho doode–pud drig–glig poat–spake nef–doot vite–dite dabe–rabe maft–snew poe–flir prail–blail sare–cham ked–ged mun–gef yocks–vox krobe–stobe vease–boce yi–cly zare–yaft blane–hane bly–shull frield–bood taid–claid kun–jun mide–stide throre–kroar drere–yate dat–clite ji–zi nool–drobe 56 slair–yi 57 REFERENCES 58 REFERENCES Phonological sensitivity: A quasi-parallel progression of word structure units and cognitive operations. Reading Research Quarterly,38(4), 470-487. doi:10.1598/rrq.38.4.3 Evidence From Four Studies of Preschool and Early Grade School Children. Journal of Educational Psychology, 96(1), 43-55. http://dx.doi.org/10.1037/0022-0663.96.1.43 Directions in Psychological Science, 14(5), 255-259. doi:http://dx.doi.org.proxy1.cl.msu.edu/10.1111/j.0963-7214.2005.00376.x Anthony, J. L., Lonigan, C. J., Burgess, S. R., Driscoll, K., Phillips, B. M., & Cantor, B. G. (2002). Structure of Preschool Phonological Sensitivity: Overlapping Sensitivity to Rhyme, Words, Syllables, and Phonemes. Journal of Experimental Child Psychology,82(1), 65-92. doi:10.1006/jecp.2002.2677 Anthony, J. L., Lonigan, C. J., Driscoll, K., Phillips, B. M., & Burgess, S. R. (2003). Anthony, J. L., & Lonigan, C. J. (2004). The Nature of Phonological Awareness: Converging Anthony, J. L., & Francis, D. J. (2005). Development of phonological awareness. Current Bankson, N. W. & Bernthal, J. E. (1990). Bankson-Bernthal Test of Phonology [Measurement Bentin, S. (1992). Chapter 11 Phonological Awareness, Reading, and Reading Acquisition: A Blum, L.H., Burgemeister, B.B., & Lorge, I. (1972) Columbia Mental Maturity Scale- third Bradley, L., & Bryant, P. E. (1983). Categorizing sounds and learning to read: A causal Bradley, L., & Bryant, P. E. (1985). Rhyme and reason in reading and spelling. Ann Arbor: Bryant, P. E., Maclean, M., Bradley, L. L., & Crossland, J. (1990). Rhyme and alliteration, Burgomeister, B. B., Blume, L. H., & Lorge, I. (1972). Columbia Mental Maturity Scale phoneme detection, and learning to read. Developmental Psychology,26(3), 429-438. doi:10.1037//0012-1649.26.3.429 Survey and Appraisal of Current Knowledge. Advances in Psychology Orthography, Phonology, Morphology, and Meaning, 193-210. doi:10.1016/s0166-4115(08)62796-x edition [Measurement instrument]. Harcourt Brace Jovanovich Inc. connection. Nature, 301(5899), 419-421. doi:http://dx.doi.org.proxy1.cl.msu.edu/10.1038/301419a0 University of Michigan Press. Instrument]. Pro-ed Inc. [Measurement Instrument]. 59 children and adults. Journal of Cognitive Neuroscience, 17(1), 168-182. doi:http://dx.doi.org.proxy1.cl.msu.edu/10.1162/0898929052880020 success in learning to read? Cognition, 91(1), 77-111. doi:10.1016/s0010- 0277(03)00164-1 [Measurement Instrument]. Pro-ed Inc. Instrument]. Pearson Assessments. phonological awareness in preschool children. Developmental Psychology, 39(5), 913- 923. http://dx.doi.org/10.1037/0012-1649.39.5.913 developmental investigation of ERP auditory rhyming effects. Developmental Science, 5(4), 467-489. doi:http://dx.doi.org.proxy1.cl.msu.edu/10.1111/1467- 7687.00241 Carroll, J. M., Snowling, M. J., Stevenson, J., & Hulme, C. (2003). The development of Carrow-Woolfolk, E. (1999). Test for Auditory Comprehension of Language- Third Edition Carrow-Woolfolk, E. (1999). Comprehensive Assessment of Spoken Language [Measurement Castles, A., & Coltheart, M. (2004). Is there a causal link from phonological awareness to Coch, D., Grossi, G., Coffey-Corina, S., Holcomb, P. J., & Neville, H. J. (2002). A Coch, D., Grossi, G., Skendzel, W., & Neville, H. (2005). ERP nonword rhyming effects in Conner, W.H., & Lang, P. J. (1969). Cortical slow wave and cardiac rate responses in stimulus orientation and reaction time conditions. Journal of Experimental Psychology, 82, 310- 320. Dawson, J. & Stout, C. (2003). Structured Photographic Expressive Language Test- Third Dollaghan, C. & Campbell, T. (1998). Dollaghan and Campbell Nonword Repetition Task Fox, B. & Routh, D. K. (1975). Analyzing spoken language into words, syllables, and phonemes: Gathercole, S. E., Willis, C., & Baddeley, A. D. (1991). Differentiating phonological memory Goswami, U. (1988). Childrens use of analogy in learning to spell. British Journal of Grossi, G., Coch, D., Coffey-Corina, S., Holcomb, P. J., & Neville, H. J. (2001). Phonological Edition [Measurement Instrument]. Janelle Publications [Measurement Instrument]. A developmental study. Journal of Psycholinguistic Research, 4, 331-342. and awareness of rhyme: Reading and vocabulary development in children. British Journal of Psychology, 82(3), 387. Developmental Psychology,6(1), 21-33. doi:10.1111/j.2044-835x.1988.tb01077.x processing in visual rhyming: A developmental ERP study. Journal of Cognitive Neuroscience, 13(5), 610-625. 60 processes during children's language development. Journal of Cognitive Neuroscience, 16(7), 1302-1318. doi:http://dx.doi.org.proxy1.cl.msu.edu/10.1162/0898929041920504 A developmental analysis using event-related brain potentials. Developmental Neuropsychology, 8(2-3), 203-241. doi:http://dx.doi.org.proxy1.cl.msu.edu/10.1080/87565649209540525 proficiency account for variability in neural indices of semantic and syntactic processing in children. Developmental Cognitive Neuroscience, 5, pp. 149–171. doi:10.1016/j.dcn.2013.03.002. Hahne, A., Eckstein, K., & Friederici, A. D. (2004). Brain signatures of syntactic and semantic Hampton Wray, A., & Weber-Fox, C. (2013). Specific aspects of cognitive and language Holcomb, P. J., Coffey, S. A., & Neville, H. J. (1992). Visual and auditory sentence processing: Henson, R. N. A. (2001). Repetition effects for words and nonwords as indexed by event‐related Hollingshead, A. A. (1975). Four-factor index of social status. Unpublished manuscript, Yale Jarvilehto, T., & Fruhstorfer, H. (1970). Differentiation between slow cortical potentials Kutas, M., & Federmeier, K. D. (2011). Thirty Years and Counting: Finding Meaning in the Liberman, I. Y., Shankweiler, B., Fischer, F. W., & Carter, B. (1974). Explicit syllable and Lonigan, C. J., & Shanahan, T. (2009). Developing early literacy: Report of the national early literacy panel. executive summary. A scientific synthesis of early literacy development and implications for intervention. National Institute for Literacy. 1775 I Street NW Suite 730, Washington, DC 20006-2401. Luck, S. J., & Kappenman, E. S. (2012). The Oxford Handbook of Event-Related Potential Luck, S. J. (2014). An introduction to the event-related potential technique. Cambridge, MA: phoneme segmentation in the young child. Journal of Experimental Child Psychology, 18, 201-212. doi:10.1016/0022-0965(74)90101-5 associated with motor and mental acts in man. Experimental Brain Research, 11, 309- 317. N400 Component of the Event-Related Brain Potential (ERP). Annual Review of Psychology,62(1), 621-647. doi:10.1146/annurev.psych.093008.131123 fMRI: A preliminary study. Scandinavian Journal of Psychology, 42(3), 179-186. doi:10.1111/1467-9450.00229 University, New Haven, CT. Components. Oxford: Oxford University Press. The MIT Press. 61 phonological awareness in preschool children. Reading Research Quarterly, 23, 263–284. [Measurement Instrument]. Academic Therapy Publications acquisition. Child Development, 73(5), 1392-1407. childhood. Merrill-Palmer Quarterly, 33(3), 255-281. Journal of Learning Disabilities, 17, 592-599. to first graders' phonemic awareness, spelling acquisition, and word recognition skills (Order No. 9980726). Available from ProQuest Dissertations & Theses Global. (304677098). Lundberg, I., Frost, J., & Petersen, O. (1988). Effects of an extensive program for stimulating Maclean, M., Bryant, P., & Bradley, L. (1987). Rhymes, nursery rhymes, and reading in early Mann, V. A., & Liberman, J. Y. (1984). Phonological awareness and verbal short-term memory. Martin, N. & Brownell, R. (2005). Test of Auditory Processing Skills- Third Edition McBride-Chang, C., & Kail, R. V. (2002). Cross-cultural similarities in the predictors of reading McGinnis, K. L. (2000). Phonemic awareness training by a classroom teacher: Its relationship Mills, D. L., Coffey‐Corina, S., & Neville, H. J. (1997). Language comprehension and cerebral specialization from 13 to 20 months. Developmental Neuropsychology,13(3), 397-445. doi:10.1080/87565649709540685 Mohan, R., & Weber, C. (2015). Neural systems mediating processing of sound units of Muter, V., Hulme, C., Snowling, M., & Taylor, S. (1997). Segmentation, Not Rhyming, Predicts NAEP Nations Report Card - National Assessment of Educational Progress - NAEP. (2015). Perrin, F., & Garcia-Larrea, L. (2003). Modulation of the N400 potential during auditory Robertson, C., & Salter, W. (2007). Phonological Awareness Test- second edition Rugg, M. D. (1984). Event-related potentials in phonological matching tasks. Brain and Language, 23(2), 225-240. doi:http://dx.doi.org.proxy1.cl.msu.edu/10.1016/0093- 934X(84)90065-8 Scarborough, H.S. (1990). Index of Productive Syntax [Measurement Instrument]. Early Progress in Learning to Read. Journal of Experimental Child Psychology,65(3), 370-396. doi:10.1006/jecp.1996.2365 language distinguish recovery versus persistence in stuttering. Journal of Neurodevelopmental Disorders, 7(1), 13. Retrieved March 26, 2018, from http://nces.ed.gov/nationsreportcard/ phonological/semantic interaction. Cognitive Brain Research, 17, 36–47. [Measurement instrument]. LinguiSystems, Inc. 62 training phonological awareness in kindergarten: evidence from two German studies. Journal of Experimental Child Psychology, 66, 311–340. early reading. Journal of Educational Psychology, 86(2), 221-234. doi:http://dx.doi.org.proxy1.cl.msu.edu/10.1037/0022-0663.86.2.221 (2002). Brain mechanisms for reading words and pseudowords: An integrated approach. Cerebral Cortex (New York, N.Y. : 1991), 12(3), 297-305. doi:10.1093/cercor/12.3.297 awareness in kindergarten children: Issues of task comparability. Journal of Experimental Child Psychology, 38, 175-190. Schneider, W., Kuespert, P., Roth, E., & Vise, M. (1997). Short- and long-term effects of Stahl, S. A., & Murray, B. A. (1994). Defining phonological awareness and its relationship to Simos, P. G., Breier, J. I., Fletcher, J. M., Foorman, B. R., Castillo, E. M., & Papanicolaou, A. C. Stanovich, K. E., Cunningham, A. E., & Cramer, B. B. (1984). Assessing phonological Stanovich, K. E. (1992). Speculations on the causes and consequences of individual differences Treiman, R., & Baron, J. (1983). Phonemic-analysis training helps children benefit from Wagner, R. K., & Torgesen, J. K. (1987). The nature of phonological processing and its causal Xu, B., Grafman, J., Gaillard, W. D., Ishii, K., Vega-Bermudez, F., Pietrini, P., Reeves-Tyler, P., Dicamillo, P., & Theodore, W. (2001). Conjoint and extended neural networks for the computation of speech codes: The neural basis of selective impairment in reading words and pseudowords. Cerebral Cortex, 11(3), 267-277. doi:http://dx.doi.org.proxy1.cl.msu.edu/10.1093/cercor/11.3.267 in early reading acquisition. In P. B. Gough, L. C. Ehri, & R. Treiman (Eds.), Reading acquisition (pp. 307-342). Hillsdale, NJ, US: Lawrence Erlbaum Associates, Inc. spelling-sound rules. Memory & Cognition, 11, 382-389. role in the acquisition of reading skills. Psychological Bulletin, 101(2), 192-212. doi:http://dx.doi.org.proxy2.cl.msu.edu/10.1037/0033-2909.101.2.192 63