ABSTRACT THE ELECTROPHYSIOLOGICAL RESPONSES OF GUSTATORY NEURONS OF THE MUD PUPPY (NECTURUS MACULOSUS) By David William Samanen The taste responses of the single fibers of the glossopharyngeal nerve of the mud puppy, Necturus maculosus, to an extended concentration series of NaCl, HCl, quinine hydrochloride and sucrose and to thermal stimulation were examined. The isolation of each neuron (unit) was evaluated by computer analysis with a program, FREDSAM, that analyzed the action potential amplitudes of the experimental records. An enumer- ator program, SAM-COUNT, using the neuron's amplitude window, counted impulses over the tests, and pre- and post-stimulus distilled water rinses. Latency was evaluated by monitoring the time of arrival of the stimulus solution to the neuron's receptive field. Stimulus-response (SR) and latency functions were calculated for each test series. The aform of the gustatory response was also observed, i.e., whether the neuron responded with increased activity to stimulation, with decreased activity, or with increased activity during the water rinse. The forms of the gustatory responses, their SR functions, and latency functions were found to vary among different stimuli and con— centration as well as among nerve fibers. However, these parameters Were observed in specific combinations, the most unique being a constant David William Samanen SR and latency function for a single chemical series. Combinations of the response parameters were found which defined nine types of taste response. Most neurons responded to more than one of the chemical stimuli and therefore possessed multiple gustatory sensitivity. Their responses to at least two of the stimuli were of different type showing that the neurons also possessed a degree of chemospecificity. THE ELECTROPHYSIOLOGICAL RESPONSES OF GUSTATORY NEURONS OF THE MUD PUPPY (NECTURUS MACULOSUS) By David William Samanen A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Physiology 1976 .....__...- ~— .a..——..~.. ACKNOWLEDGEMENTS God bless Rudy, Marc, Pat and John, Dr. WOlterink and Dr. Hoffert, Drs. Hatton, Cunningham, and Pittman, and especially Carol, Steven, Jeffrey, Kay, Mom and Dad, and this dissertation. ********* ii TABLE OF CONTENTS Page I. INTRODUCTIONOOOOOOOOOOOOOIOOOOOOOOOOOO0.0000000000000000. 1 II. LITERATURE “VIEWOOOOOOOOO0.00...OOOOOOOOOOOOOOOOOOOOOOOO 3 Introduction......... ..... ............................ 3 Anatomy............................................... 3 General............................................ 3 Taste Organs of Necturus........................... 4 Innervation of the Mud Puppy's Taste System........ 9 Whole Nerve Electrophysiology......................... 14 Pattern of the Whole Nerve Taste Response.......... 14 Determination of Response Magnitude................ 15 Gustatory Stimulus-Response Functions.............. 18 Single Neuron E1ectrophysiology....................... 22 Character of the Taste Response.................... 22 Increasing Activity Upon Stimulation............ 22 Decreasing Activity Upon Stimulation............ 24 Rinse Responses................................. 26 Determining Response Magnitude..................... 26 Patterns and Phases of the Taste Response.......... 28 Latency of Gustatory Responses..................... 29 Multiple Gustatory Sensitivity of Taste Neurons.... 30 Selection of Stimulus Concentration............. 33 Functions and Relations with Single Concentra- tion Tests................................... 35 Single Unit SR Functions........................ 45 Mechanical and Thermal Sensitivity of Taste Neurons 55 III. STATMT OFTHEPROBLmOOOOOOOOOOOOOCOOOOOOOOOOOOOOOOOOO 57 IV. METHODSOOOOOOOOOOOOIOOOOIOOOOOOO0..OOOOCOOOOOOOOOOCOOOOOO 59 Computer Determination of the Isolation of Neurons, FREDSAM............................................ 6O Stimuli and Protocol.................................. 68 Stimulus Delivery..................................... 71 Data Analysis......................................... 71 Photographic....................................... 71 iii TABLE OF CONTENTS—-continued V. VI. VII. VIII. Post-Experimental Computer Data Analysis, SAM-COUNT. Graphical Analysis of Enumerated Response Activity..... Response Latency....... Measuring Delivery Time............................. Conduction Velocity................................. Determining Physiological Latency................... RESULTSOOOOOOOOOOOOOOOOO0.0. Spontaneous Activity................................... Forms of Gustatory Response............................ Gustatory Stimulus-Response (SR) Functions............. Neurons with Low Level Spontaneous Activity......... Neurons with High Level Spontaneous Activity........ Rinse Response SR Functions......................... Gustatory Latency Functions............................ Combined SR-Latency Functions.......................... Mechanical and Thermal Responses....................... DISCUSSIWOOOOOOOOO0.00...OOOOOOOOOOOOOOOOOIOOO0.0.0.0.... Mechanical and Thermal Sensitivity of the Gustatory Neurons............................................. The Expanded Gustatory Response........................ Considerations of Stimulus-Response (SR) Functions.. Considerations of Latency Functions................. Varied Response Forms............................... Multiple Chemosensitivity of the Taste Neurons......... Considerations of the Origins of the Taste Response.... Single Unit Activity and the Whole Nerve Response...... SUMYOOOOOOOOOCCOOOOOOOOOOOOOOIOOOOOOOOOO00......0...... RECWENDATIONS FOR FUTURE STUDYOOOOOOOOOOOOOOOOOOOO0.0.0. BIBLIWRAPHYOOOOOOOOOO0.0.0.0....0.00.0000...OOOOOOOOOOOOOOOOOOOO APPENDICES I. II. III. IV. SR AND LATENCY FUNCTIONS... OPRATION 0F FRESAMOOOOOOOIOOOOOIOIOIOOOOOOO0.0.0.0000... OPEMTION AND EVALUATION OF SAM.COUNTOIOOOO 00...... O... 0.. EVALUATION OF DELIVERY TIME MEASUREMENTS............ ...... iv Page 74 75 76 76 77 80 81 82 85 89 89 99 102 105 110 117 121 121 121 121 122 124 125 126 127 134 136 138 143 154 159 168 TABLE II. III. IV. LIST OF TABLES Relative effectiveness of four primary stimuli used in taste experiments on chorda tympani neurons.............. Distribution of SR and latency functions................. SAM—COUNT reprOducj-bflity.00.00.000.000.00000000000000... Enumeration product-moment correlation................... Page 34 108 156 157 10. 11. 12. 13. 14. 15. 16. 17. 18. LIST OF FIGURES Gustatory eminences of the mud puppy............ .......... Light micrograph of a gustatory eminence.................. The lingual distribution of gustatory emdnences........... Lingual innervation and gustatory sensitivity of the glossopt‘aryngeal nerveIOOOOOOOOOOOIOOOOOOOOOOOOOOOOOOOO0.. The phasic whole nerve taste response of Necturus......... Gustatory whole nerve SR functions of Necturus............ Latency function of single gustatory fibers to NaCl....... Multiple gustatory sensitivity of single neurons.......... Best taste profiles of nine single gustatory neurons...... SR functions for single neurons........................... The single neuronls SR functions to different stimuli..... Interrelated SR functions................................. Electronic equipment for gustatory single neural unit recordingOOOOOOOOOOOOOOO0..0.0.0.0.0...OOOOOOOOOOOOOOOO... Computer analysis (FREDSAM) of a well-isolated neural unit FREDSAM analysis of multiple isolated units............... Calibration of the computer and photographic records...... The stimulus delivery system.............................. Measurement of solution delivery time..................... vi ll l3 17 20 32 38 41 47 49 53 62 65 67 70 73 79 LIST OF FIGURES--continued FIGURE 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 35. Page Activity and responses of glossopharyngeal afferent neuronSOOOOOOOOOOOOOO000......OOOOOOOOOOOOOOOIOOO ......... 84 Depressed activity and rinse responses.................... 88 Gustatory responses with a positive SR and negative latency function..... ....... ........ ..... ...... ....... .... 92 Regular forms of SR and latency functions.......... ....... 94 The negative SR and positive latency function of a unit... 96 U-shaped SR functions and bell-shaped latency functions... 98 Gustatory responses of units with high level spontaneous act1v1ty0000000000000......OOOOOOOOOOOOOOOOOOOOOO0.0.0.... 101 Rinse response SR functions............................... 104 Response magnitude versus latency......................... 107 Gustatory responses defined by SR and/or latency function. 113 Combined SR functions to the four primary stimuli for several neuronSOOOOOOOO0..O.I...OOIOIOOOOIOOOOOOOOOO0.0... 116 Thermal responses of gustatory neurons.................... 120 Summed SR functions of the four primary stimuli for 29 gustatory neurons 0f NecturuSOOO...OCOOOOOOOOOOOOOOOOOOOOO 129 Summed response patterns for 29 gustatory neurons of NecturUSOOOOO0.0...OOOOOOOOOOOOOOOOOCOOOOOOOOOOOOOOOOO0.0. 132 Action potential alterations by background electrical DaiseOOOOOOOOO0.0COOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO ......... 146 Diphasic amplitude measurement and display... ............. 149 FREDSAM displays of units with poor isolation, low activity ...................... . ........................... 152 vii LIST OF FIGURES-continued FIGURE 36. 37. 38. 39. 40. 41. 42. 43. 44. Dual determination of delivery time....................... Comparison of phototransistor and fluid switch determina- tions Of delivery time...-0.0.0....OOOOOOOOOOOOOOIOOOIOOOO Effects of saline concentration on delivery time measure- mentCOOOOOOOOOOOOOOOOOOOOOOCOOOO0.000.000...00.......00... Complex SR and latency functions.......................... Positive SR functions..................................... SR functions with zero and negative slope................. U-shaped and bell—shaped SR functions..................... Latency functions with negative, zero, and positive slope. Latency functions with positive and negative slopes combinedOOOOOOOOOO0.0....0...OOOOOOOOOOOOOIOIOO...00...... viii Page 162 165 167 170 172 174 176 178 180 I. INTRODUCTION Electrophysiological recordings from single gustatory neurons from many species have revealed a variety of neurophysiological taste responses. These have differed in form: 1) many neurons showing increased activity to stimulation, 2) others responding with decreased activity to stimulation, and 3) many neurons having increased activity with the presentation of the post-stimulus rinse. The response magni- tude has been found to vary with the concentration of the stimulus in manners giving rise to widely different stimulus-response or SR func- tions. The responses of several neurons regularly increase with increasing stimulus concentration. Other fibers respond with regularly decreasing responses as stimulus concentration is increased. Some neurons have SR curves with both positive and negative slopes, that are U-shaped or bell-shaped. Yet the majority of experiments in taste sensory physiology have been performed using single, unequal concentra- tions of different stimuli to define the mechanisms of the first order peripheral afferent fibers of the taste system. A systematic investi- gation using an extended concentration range of chemical stimuli has not been reported for any animal. The latency of the gustatory response of the single taste neuron has recently been studied as a parameter of the taste system (T. Sato, 1976). An inverse relationship between stimulus concentration and response latency has been found but with implications that various forms of a latency-concentration function may exist. Accordingly, the experiments of this study were undertaken in order to define the taste response of single neurons according to three parameters: 1) the form of the response, 2) its SR function, and 3) its latency function. The mud puppy has been chosen for taste research owing to the unusually large size of the cells of its taste organs, the taste buds. Intracellular recordings from the cells of the taste buds of Necturus have been completed in an attempt to discover their responses to chemi— cal stimulation (West, 1976). However, only the general sensitivity of the gustatory neurons were known as revealed from whole nerve studies. The detailed responses of the individual neurons of the mud puppy's taste system are presented in this study. II. LITERATURE REVIEW Introduction This review will consider initially the anatomy of the taste system of the mud Puppy, Necturus maculosus. The innervation of its tongue by cranial nerves VII, IX, and X will be described. Then will follow a discussion of the neurophysiology of the taste system of Necturus as has been revealed by whole nerve electrophysiological recording. Finally, the taste responses of single gustatory neurons of several species will be reviewed. This will emphasize: 1) the forms of the single neuron's taste response (increased or decreased activity on stimulation or increased activity on post-stimulus rinsing), 2) the latency of the single fiber's taste response, and 3) the sensitivity of taste neurons to several chemical stimuli. For the latter, those experiments which use single concentrations of stimuli will be con- trasted with those which use a range of concentrations, defining stimulus—response or SR functions. Anatomy General The mud puppy, Necturus maculosus, is a completely aquatic, caudate amphibian. It prefers dark, cool habitats living on the bottom of the fresh water lakes and rivers of Eastern North America consuming spawn, small fish, aquatic insects, and insect larvae. The adult mud puppy, possessing bushy, external gills, resembles the larval salamander or frog. Necturus at birth is 2.3 cm long, matures after eight years and may reach an adult length of 49 cm. Animals have lived as long as 23 Years in captivity. The physiologist's keen interest in Necturus is directed to the unusually large size of the cells of many of its organs. Knowledge of cellular function and response of the vertebrate kidney and retina have been advanced by studies of this animal. Taste Organs of Necturus Under the dissecting microscope, the mud puppy tongue glistens with a mucous coat and is smooth, lacking any obvious papillae (the well-differentiated structures which support the taste organs of most vertebrates). Only under indirect illumination or reflected glare can the lingual surface be resolved to show many low relief elevations or mounds (30 u height, 250 u diameter). These lingual eminences, identi- fied by Farbman and Yonkers (1971), each hold a single taste bud. Figure l is a scanning electronmicrograph of two of these eminences (Samanen and Bernard, 1975). At the center of each mound can be seen the irregular surface of a taste bud, i.e., the taste organ of Necturus. The light micrograph of a sectioned eminence (Figure 2) shows that the taste bud extends completely to the lingual surface in contrast to the buds of most mammals which lie below the epithelium, contacting the surface environment only by way of a narrow taste pore of 5 u diameter. .munonHam Hmnwcwa some «0 sou mnu um mummnam one woman pawnwm a mo mummuzm umanwouuw use .samumouowaouuomao wcaaamom mafia aw doom menace no maowum>oao moHHou 30H one aHfiuH3 oHH moan manna muH .mmaawmwa woumwuamumwMHvIHHmB mxoma.mmmmmmmm .zamsa was onu mo moonoaflam muoumumno .H ouswwm .wnan mcavwsaou saws voawmum .xofinu : m .onmmau vowvonfio comm Bonn mama mm3 nowuoom any .vsn ago manomaam.anaosm uonv humHHamwo cam mumnHm m>uma wswawmuaoo .mHHHmmn awaken < .mummunm Hmswaaa onu ou waouxm mHHou voumwaoao omens «mmmmmmmm.wo amwuo mummu may .vsn mummu owuoa .onswm m moHH moamnwao some casuaz .ouamafiam huoumumsm m mo namuwouofia uswfiq .N ounwflm (For examples, see Murray and Murray, 1967 for the rabbit, and Farbman, 1965 for the rat.) In Necturus, a dermal papilla with a capillary supports each bud. In the oral cavity, gustatory eminences are found on the entire tongue, on much of the mucosa of the pharynx and gill arches. Figure 3 shows the distribution of 658 lingual eminences for one animal (Samanen, Kryda, and Bernard, 1975). The distribution is graded longitudinally, becoming more concentrated distally. Taste buds not associated with lingual eminences have not been reported. No extra- oral taste buds have been reported for the mud puppy. The dimensions of the mud puppy‘s taste buds (90-120 u wide by 100-150 u height) are twice as large as those of other vertebrates (40-80 wide by 50-80 H height). Their unusually large size arises both from their greater number of cells (80-100 versus 30-80) and the larger dimensions of those cells, reported to be twice as great in diameter and length (Farbman and Yonkers, 1971). The termination of nerve fibers within the taste buds is also described by these authors. Innervation of the Mud Puppyis Taste System Cranial nerves VII, IX, and X supply the mud puppy's gustatory afferent fibers (Figure 4). The facial nerve (VII) arises in part from the bulbar fasciculus communis, considered by Strong (1895) and Kingsbury (1895) to be an homologue of the mammalian solitary fasciculus, the locus of termination of primary gustatory afferent fibers. Drflner (1901) stated that the ramus alveolaris (VII) courses toward the tongue but largely terminates between the tongue and mandible. This contrasts with the major lingual innervation by the mammalian chorda tympani (VII). Figure 3. 10 The lingual distribution of gustatory eminences. 658 lingual eminences, identified for one mud puppy, are seen to cover the dorsal lingual surface. A slight longitudinal gradation exists, more gustatory eminences being concentrated at the distal tip (top of figure). 11 1mm Figure 3. 12 .mu>hoc Hmomchumnmommoaw was Hmaomm ecu up vuumzm madam u>aunuoou m mumuwwsw Am munwam mam uxuu uumv umuwuuauoaoo umoa mum moan momma uuogz .mHu on» as hua>wuawaom mo unwaomn .Aomnoamou adaaxma u ooav nzonm mum Humz z m.o nufia coaumaaawum oumuocnm ou mumdoamuu o>aumamu one .mnfivuooou u>uon paces Hmowwoaowmmnm Iouuooao he vuawauouov we m>uua Hwowahumnmommoaw ecu mo huH>HuHmcmw wuoumumsm Hmcowwmu uau macaw ma unwwu onu OH .Amw_vsm NM..mo>uua nous Hmfinonmun nuanu was vacuum unuv o>uoa mswm> onu was .AaMMMMIdM «MM was aMMMIdM «MM .monocmun UHumaouulumoa use long onuv o>uuc Hmmwcmumcaommoaw mnu .A.>H< .m .HH> .mwumaou>aw mason mauv m>uua Hmwomw onu mo monocmun mo coaumnfiaumu was omunoo uumawxoumam onu 305m ozouum may .mmmmwmmm.mo auumhm mumMu one on musnwuu Iaou defies mm>uoc Hmacmuo ecu mo mmnuamum one mum umua onu do vuumuumDHHH .w>uus Hmowczumnmommoaw unu mo >uH>HuHmaom mucumumsw mam soaum>nucaa Hmnwcfiq .q muswwm 13 .v 2:»: 43.. .e .x. a... e .x. A .5 .e .5. l4 Gesteland and Dixon have recorded from the facial nerve in Necturus (Dr. Robert Gesteland, personal communication). This gives electro- physiological confirmation to a possible role of VII in the mud puppy‘s taste system. The glossopharyngeal nerve (IX) has been consistently described as part of the taste system of caudate amphibians (Drfiner, 1901 and Francis, 1934). The rami pre— and post-trematicus (IX) terminate in epithelium above the lingual cartilages over most of the distal tongue, though the lateral and longitudinal extent of their regions of distribu- tion are not known. Using punctate chemical stimulation 5-6 ul, 0.5 M NaCl, contrasted with distilled water) and whole nerve electrophysio- logical recording, Samanen, Kryda and Bernard (1975) found the area defined in Figure 4 to be the gustatory receptive field of IX. The decreasing sensitivity at the most distal area (5-10 mm caudally from the tip) is consistent with Gesteland and Dixon's recording from VII. The vagus nerve (X), also called the second and third branchial arch nerves of Necturus (Drfiner, 1901), terminates above the pharyngeal branchial cartilages of the proximal oral cavity (Drfiner, 1901). While electrophysiological confirmation of its role in the mud puppy's taste system is lacking, it may provide gustatory innervation to the pharyngeal mucosa and gill arches. Whole Nerve Electrophysiology Pattern of the Whole Nerve Taste Response The whole nerve response has been recorded in many animals. The activity from the many neurons is summated by an electronic integrator. 15 The summated taste response typically contains an initial phasic increase of activity upon stimulation, followed after 1-3 seconds by a lower tonic level of activity during continued stimulation. However, the typical phasic-tonic pattern varies among species, among stimuli, and even stimulus concentration. For example, NaCl and HCl test solu- tions produce the typical pattern from the chorda tympani of the cat, while quinine and sucrose do not (Pfaffmann, 1955). Tateda (1961) found the phasic response to occur proportionally with all suprathres- hold concentrations of NaCl (from the catfish facial nerve) while the tonic response had a higher threshold (above 0.15 M). The mud puppy's glossopharyngeal nerve shows only a phasic response for all stimuli at all concentrations tested (Samanen, 1973). This is seen in Figure 5 for KCl up to 0.1 M. Determination of Response Magnitude Experimenters have chosen one of the phases of the taste response to estimate the size of the entire response. For example, Beidler (1953) chose the tonic level of activity as a magnitude estimator while Tateda (1961) used only the phasic portion. Halpern and Tapper (1971) and Halpern and Marowitz (1973) have shown in the rat that taste quality information is available within the first second of stimulation. When conditioned to avoid 0.3 M NaCl, rats could detect, evaluate, and stop licking the conditioned stimulus within 138 to 600 msec. Faull and Halpern (1972) have shown a strong correlation between the initial rise of the phasic response and the concentration of the stimulus and note that this portion is most linear l6 .umvuoomu sump noonwaw>usu 0:» mo nommwuum nu ma Hum z H.o ou uncommon ofimmca us» no omen Hugues“ usu mo opunu unwfiam one .AHmamn nosoav cowumummvm nouns umumm cowumowannm “mama moaaaumav ou umonam uoa moon was sowumuunuoaoo ou Honoauuoaoum ma uH monsoon Awasawum Hmaumsu no Hmowamnooa on no: vamv soaumanawum xuou Imumnw maco ou uncommon mnu on ou voumvwmcou up coo uncommon ofimmna use .muH>Huom nausea aw ommuuonfi Hmauacw uaoumfimcoo m maamunou uncommon mummu m>umc macs: osu umnu 30cm moomuu vmmoaawuuesm one .nowunaom Hum mo noaumuucmoaoo onaHm m mo aoaumowanam vmumommu mnu ou noncommmu m>uoc Hmmwahumamommoaw woumaanm mnu mo mwuooou manna magnucoo Sou Home: unu ca Huang zoom .mmmmwmmm mo uncommou mummu o>uua oaosa owmmsm any .m muswfim 17 8m .m 9...»... his} a. on: .3. 3. 3. .s. 32?. U Names; a... .3. a... 18 over time. Smith (1975) expanded these findings, also for the rat, using both electrical and chemical stimulation. The magnitude of the phasic response was proportional to the rate of rise of stimulating current, or the delivery rate of 0.1 M NaCl. Current of different amplitude delivered with the same rate of rise had equal phasic responses and proportionally unequal tonic responses. Thus for magnitude estima— tion, use of either phasic or tonic component alone may result in dif- ferent conclusions. The tonic activity is proportional to steady concentration of the stimulus; whereas, the phasic is proportional to the rate of change of stimulus concentration and contains taste quality information. Gustatory Stimulus-Response Functions Gustatory intensity functions, called the stimulus-response or SR functions, have provided useful information about-an animal's taste system. These include the relative effectiveness and thresholds of stimuli. Figure 6 is a composite graph showing the intensity functions for eight stimuli in the mud puppy's glossopharyngeal taste system. For all chemicals, the response is a regularly increasing function of concentration. 801 at all concentrations is the most effective stimu- lus and has the lowest threshold. (The threshold is the concentration at which the response exceeds the water-adapted, pre-stimulus level of activity. The threshold for HCl, not fully attained in the function of Figure 6, is below 0.0003 N or -3.5 Log N.) Sucrose has the highest threshold and smallest responses. All other stimuli are of intermediate l9 .mnma .suGQEMm nouw< .Aumn mmuwv saonm ma hufi>wuom Houm3 mHsu mo AoawuumSUHousav amass vmuaaaa < .mswoOu anemone nouns was on soaumusomuun nouns so wswuuauoo muw>auow unu Scum moswsuuuov was mua>auum uswaomoe any .muwooan segue sues somauaeaoo moumuaaaomm was Hanswum Hmsuonu com Hmuwsmnooa ou momsonmou mo msoausnauusoo nouns Iaaaao mane .Homz z H.o ou uncommon unu mo o=Hm> unmouoe o no vommuunxu who moanawum some mo sowuwuusoosou some ou mumsommou u>uos mace: use no mmsHm>. .Aomouosm I .uosm .okuoaauouvmn oswsfisv I.mmmmV voumou mamowauno use you saonm sum omega .msusuooz mo msoauussm Mm u>uos oaos3 muoumumsu .o madman 21 effectiveness and no consistent correlations can be drawn from their SR functions. One use of SR curves is to choose equally effective concentra— tions of test stimuli. For example, interpolation from Figure 6 at 150 relative response units shows that taste responses of equal magnitude can be expected from the glossopharyngeal nerve of Necturus for the following concentrations of stimulus; 0.2 M NaCl, 0.15 M KCl, 0.3 M NH Cl, and 0.02 N HCl (-0.7, -0.9, -0.4, and -l.85 Log concentration 4 respectively). The magnitude of the summated gustatory response has been well correlated with perceptual intensity. For example, recording from the human chorda tympani, the summated responses have shown: 1) that the electrophysiological and perceptual threshold for NaCl are the same, 2) that several sugars have the same relative electrophysiological and psychophysical effectiveness for any one person, 3) that electrophysio- logical adaptation closely parallels psychophysical adaptation, and 4) that gymnemic acid applied to the tongue abolishes the electro- physiological responses to sugars as well as the perception of sweet substances (Borg gt 31., 1969). Whole nerve studies are limited in determining the more subtle functions of the taste system. For example, little information regard— ing the neurophysiological determination of taste quality is available in the pattern of the whole nerve response, or in its SR or latency functions. For example, near threshold concentration, human observers can detect the presence of a taste stimulus but cannot identify it. 22 It is only at higher concentrations that the quality of the stimulus can be recognized. There is nothing in the whole nerve response that explains this distinction. Understanding the neurophysiological mechanisms underlying these differences may require knowledge of the function of the individual elements of the taste system, the neurons and receptor cells and their interactions. Single Neuron Electrophysiologz, The gustatory responses of single neurons have been recorded at many levels in the nervous system: from the peripheral nerve fibers, in the geniculate ganglion of the facial nerve, at the nuclei for second order neurons (the nucleus tractus solitarius, the sensory trigeminal nucleus), from third order cells at the ventral thalamic nuclei, and at the cerebral cortex. 0f the many preparations studied, the rat's chorda tympani is now classical by virtue of repetition. Individual gustatory neurons have been investigated in the rat, cat, hamster, guinea pig, rabbit, sheep, dog, frog, and rhesus, macaque, and squirrel monkeys, but not in humans. This review will emphasize the neurophysiology of the first order neurons. Character of the Taste Response IncreasingTActivity Upon Stimulation. Gustatory neurons generally respond with an increase in activity when stimulated. Yet their re- sponse patterns vary widely. The most typical pattern, first discovered by Pfaffmann (1941) in cat chorda tympani fibers, resembles the whole nerve response and consists of a phasic increase followed by a tonic 23 discharge throughout stimulation. He noted that the firing pattern of impulses was greatly irregular. This contrasted with both the regular tonic discharge of several sensory systems (e.g., the cat carotid baroreceptors) and the regularly decreasing frequencies of more rapidly adapting systems like the visual ganglion cells (Bronk, 1935). Fishman (1957) noted five distinct patterns when recording from rat chorda tympani fibers: 1) the typical phasic-tonic pattern, 2) an immediate rise to tonic discharge, 3) a slow rise to a peak response followed by gradual but complete decline, 4) a 0.2 second burst followed by quiescence and then a gradual build-up to maximum discharge, and 5) rhythmic bursting with regular intervals between impulse trains. All types occurred under constant stimulation. The response pattern has been found to vary with different stimuli, to vary with stimulus concentration, or to be invariant for a specific fiber. For example, a cat chorda tympani fiber responded with the typical phasic-tonic pattern to 0.1 M NH CI but with only phasic discharge (complete, rapid 4 adaptation) to 0.1 M NaCl (Beidler 33:4,, 1955). Ogawa £13.11; (1974), found only phasic responses for one rat chorda tympani neuron with near threshold NaCl stimulation. With greater stimulus concentration, the tonic discharge appeared, becoming more regular with increasing concen- tration. A different chorda fiber responded to 0.5 M sucrose, 0.1 M NaCl and 0.02 M quinine hydrochloride with regular bursting activity. Thus, even though the whole nerve response displays a particular pattern, the underlying activity of the individual units may be widely different. 24 DecreasiggiActivity Upon Stimulation. Taste neurons are rarely completely inactive. They often fire at slow rates whether adapted to a rinse solution (usually distilled water) or in the complete absence of lingual stimulation. Mean rates of 0.5 to 0.8 Hz are common (rat and hamster chorda tympani neurons, Ogawa, Sato and Yamashita, 1968; Miller, 1971; Frank, 1973). Sato g£_§1, (1975) noted a greater fre- quency for the macaque monkey chorda fibers, 1.96 Hz. The activity varied among fibers, with a range of 0-9 Hz, and 7 of 67 units above 4 Hz (all determined from 5 second averages). Boudreau g£_gl, (1971) classified first order neurons of the cat's geniculate ganglion partly on their levels of spontaneous activity. Several units had very high rates of regular discharge (18—77 Hz), some in near synchrony with each other. These were not associated with lingual stimulation, several units responding to static displacement of the pharyngeal tissue, palate tissue, and eyeball, and several showing no response to any stimulus. Lingual units (responding to electrical, mechanical, or chemical stimu- lation of the tongue) varied widely in spontaneous activity, from 0-10 Hz (20 second averages). Their pattern of response consisted of various forms of bursting or a completely irregular discharge. Funakoski ggugl. (1972) observed rapid spontaneous activity of 15 Hz in the lingual sensory cells of the cerebral cortex of unanesthetized dogs and rats. Spontaneous activity can be depressed or completely suppressed with specific gustatory stimuli, and therefore this represents a second type of response. Decreased spontaneous activity has been observed in responses to CaClz, quinine, NaCl, HCl, and sodium saccharin (Pfaffmann, 25 1941, Boudreau £5 31,, 1971; and Sets g£_al,, 1975). In the macaque monkey, a particular chorda tympani fiber showed normal responses (increasing activity) to sodium saccharin concentrations greater than 0.003 M, while the spontaneous activity was proportionally depressed by 0.001 M and 0.0003 M.in the lower portion of the fiber's SR curve (Sate st 31,, 1975). Dog and rat cerebral cortical cells showed dif- ferential sensitivity to chemical stimulation by various changes in activity from spontaneous levels (0.1 M NaCl, 0.01 quinine, 0.5 M sucrose, or 0.05 N tartaric acid). Some cells were suppressed by all four, while others were selectively suppressed, excited (to rates above 40 Hz), or showed no response to one or more of the stimuli (Funakoshi gt; 1;. , 1972). The origin of the spontaneous activity is unknown. Pfaffmann (1941) observed that topical procaine hydrochloride abolished the spontaneous activity before the loss of taste responses. Sato £5.29: (1975) and Ogawa g£_al, (1968) observed a weak correlation coefficient (r a 0.4) between spontaneous activity and the response to cold stimu- lation in thermally sensitive gustatory neurons (rat, hamster, and macaque monkeys). They suggest that the spontaneous discharge repre- sents ongoing thermal stimulation for these homeotherms, whose lingual surface temperature was only 250 C. Cohen gt 31, (1955) suggested that spontaneous activity may result from specific neurons responding to the rinse solution even after extensive periods of adaptation. Specifically, they found fibers in the cat chorda tympani that fired continuously under Ringer's rinse and that were also sensitive to NaCl. 26 Rinse Responses. A third category of response occurs not on application of the stimulus, but upon its removal with a rinse solution. For example: 1) after testing with sodium saccharin (hamster chorda tympani) in a single fiber that was sensitive to sucrose, 2) after NaCl in one fiber responding to the halides (hamster chorda tympani), and 3) in 21 of 26 fibers of the cat's chorda tympani after 0.3 M NaCl (Ogawa g£“§1,, 1969), after 0.03 N HCl (Yinon and Erickson, 1970), and after 1.8 M.sucrose (Bartoshuk gt 31., 1971). Funakoshi gtflgl, (1972) refer to a cerebral cortical neuron responding with brief "on" responses to the test stimulus and with "off" responses to the water rinse. Bartoshuk gtflal, (1971) feel that the underlying mechanism of the rinse response involves the specific stimulatory action of water. Ogawa ‘g£_al, (1969) suggest that the sodium.saccharin rinse responses arise from the re-stimulation of gustatory receptors which had been inactivated. All three types of responses can be considered legitimate gusta- tory responses since they occur to specific chemical stimuli or in that they vary in magnitude according to the concentration of the stimulus. The mechanical and thermal stimulation that occurs with the test solu- tions and rinses are assumed to be equal for all stimuli. DeterminingTResponse,Mgggitude Given the variety of forms of the taste responses, investigators have differed in establishing criteria for defining and measuring a response. No standard exists. Some determine a criterion level of test activity such as a 502 increase (Frank and Pfaffmann, 1969) or an increase greater than one standard deviation (Sato g£_gl,, 1969) above 27 spontaneous activity. All investigators have used average spontaneous activity levels preceding all tests for comparison with the response to Most experimenters measure the spontaneous activity each single test. This is an essential and test responses over equal intervals of time. requirement for identifying lingering activity from previous stimula- tion. Responses of depressed activity are often only mentioned qualita- tively. Miller (1971) separately calculated response depression and enhancement ratios. With the variable form and frequency of the single unit taste response, a similar variety of magnitude calculations have been employed. Usually the average frequency of impulses over the test period is calcu— lated (as impulses per second or per five seconds, etc.). Both Miller (1971) and Ganchrow and Erickson (1970) avoid the initial phasic response. They consider the later periods to measure the more stable or steady state response. Any reviewer of taste literature should be aware of the differ- ences in response magnitude estimation. The observations using more astringent criteria with respect to increases in activity above spon— taneous levels may not correlate with other observations. Suppressed activity, a more subtle response may be simply ignored. The exclusion 0f Portions of the response by considering only earlier or later events may Over-emphasize the responses to specific chemical stimuli whose greatest activity is concentrated in a particular period. Elimination Of the initial response period seems especially hazardous given the rat's . (demonstrated gustatory discrimination powers which encompass as 28 little as the first 200 msec. Finally, any averaging method obscures patterned discharge information which may be present in the taste response. Patterns and Phases of the Taste Response The various patterns and phases of discharge in the taste response have been briefly described above. Mistretta (1972) found rat chorda tympani fibers that showed periodic bursting with ten different stimuli. Other fibers she described as "differential" in that they responded to certain stimuli with bursting and to others with various patterns of random decay. Ogawa ggual. (1974) similarly discovered that the gusta— tory response pattern could be stimulus dependent or fiber dependent (rat chorda tympani). Several of the response patterns were unidentifi— able. They also found increased pattern stability with increasing stimulus intensity but no shift in pattern with concentration changes. Correlations between patterns and stimulus were low and they could not determine the role of pattern information with respect to taste quality. Ogawa g£_al, (1974) contrasted the initial five seconds of ten- second gustatory tests with the last five seconds and found that the ixnitial response magnitude was greater (macaque chorda tympani). The stzinnulus-response correlations drawn from either period alone could differ. However, with one exception, the differences were slight and not significant. They found that the different phases of the response had different relationships to stimulus concentration. The first five second period had a semi-Log relationship to concentration, while the secc>r1au00mmo m>uma oaona on» no woman H mm n omuqm ms H.o om Hoo.o mm moo.o «a m.o m~nafi .xcmum ma n mousm ms m.o «a ~o.o mm Ho.o em H.o wsma ..4m:mm «game on u monaoa no o.H ma No.o mm Ho.o Hoa m.o Rama .amasmfim mmmwmwm om u manna ma m.o on ~o.o ms Ho.o mm H.o mama .xamum mm a 5 .00 .u. n.. a mooo.o ON moo.o om mo.o mama .mam: mNHu m nomH m m.o ma Hoo.o me Ho.o omH m.o somfl .camaummum s gangs om u maumm ma m.o on ~o.o me Ho.o mm H.o «mesa .awm.wm «ammo cm a mmumm mm o.H mm Ho.o me mo.o mm H.o mmoma .aomxoaum Oman muumqa mN o.H oq ~o.o me Ho.o nqa m.o amma .amaemsm ON a mmamm mm 2 o.H mm : Ho.o mu 2 mo.o mm 2 H.o mama .aamauumMm m¢m sum.o.m.z .amum .uaoo .aeum .uaoo .aumm .oaoo H.=mmm .oaoo muoau=< .aemm mo .uuam Home Hum Homz mwamm uoz N Handshu mouono so muooaauomxm momma aw wow: “Haaqum humaaum Room mo Hmmmcm>wuuww .mcouswc mm 0>Humwox 35 Ogawa £5 31, (1968) tested the same concentrations of stimuli on the rat as well as the hamster. For the latter, the same stimuli extended over 80 units of relative effectiveness. Frank (1973) stated that the stimulus concentrations were chosen to represent the midrange of the SR function for each stimulus. However, the SR maxima of different chemi- cals are not identical, since some show saturation below the maximum response level of others (e.g., sucrose versus HCl). Hence, midrange concentrations could be quite unequal in their electrophysiological effectiveness on peripheral neurons. Moreover, as mentioned previously, human perceptual gustatory intensity is correlated with the level of summated whole nerve activity. The relative magnitude of the summated nerve responses of the human subjects to six equimolar sugars directly matched their perceptual intensity scaling (Borg g£_§l,, 1968). The major difficulty with the previously described methods is that single neurons often respond with SR functions quite unlike those of the whole nerve. Accordingly, in order to study responses in single ‘units with the least error, it is necessary to test them with a range ‘Of concentrations that encompasses the entire range of the whole nerve response. If this is not done, the next best protocol would be to Cflmoose individual concentrations that are equally effective in the whole Ilearve population. The results in all other cases must be considered clic'iutiously. Functions and Relations with Single Concentration Tests. When El single concentration of each stimulus is used, gustatory neurons €33§hibit a wide variety of response specificity. Some fibers respond to 36 a few chemicals. These have been labelled narrowly tuned fibers (e.g., see Pfaffman, 1955 and Frank, 1975). Their proportion of the total population is small for most peripheral nerves. Of 101 frog glosso- pharyngeal fibers, 11 were monogustatory (ll/105, Kusano, 1960) as were 1/15, 9/48, and 5/25 of rat chorda tympani fibers (Wang, 1972, Ogawa g£_§l,, 1968, and Frank and Pfaffmann, 1969, respectively) and 5/28 and 23/79 of hamster chorda fibers (Ogawa gt El}: 1968, and Frank, 1969, respectively) and 8/27 of rat glossopharyngeal fibers (Frank and Pfaffmann, 1969). The majority of fibers show multiple sensitivity with a random distribution among the primary stimuli (among rat and hamster chorda fibers, Frank and Pfaffmann, 1969 and Frank, 1973, and the rat glosso- pharyngeal fibers, Frank and Pfaffman, 1969). The observed number of fibers responding to l, 2, 3, or all of the primary stimuli does not differ from that predicted by random polynomial distribution of the sensitivities at the frequencies found in the sample population. The property of multiple specificity is illustrated in Figure 8. ‘The responses to three gustatory neurons are shown for the rat (Pfaffmann, 1955), hamster (Fishman, 1957) and frog (Kusano, 1960). line general responsiveness of the neurons varied greatly, one hamster neuron responding with 50 impulses/sec to NaCl, another with less than 10 impulses/sec to HCl and sucrose. The multiple sensitivities also veeried between fibers and animals, three units (neurons) among the nine being clearly monogustatory (Figure 8, A and _C_1_ Bag, and A Hamster). I“(Jtice that the testing of additional, non-primary stimuli (KCl for the Figure 8. 37 Multiple gustatory sensitivity of single neurons. The responses of three single neurons are profiled Q5, g, and g) rat and hamster chorda tympani and frog glosso- pharyngeal fibers, after Pfaffmann, 1955, Fishman, 1957, and Kusano, 1960). Most fibers were multiply sensitive (only two rat, éLand §_and one hamster neuron g, are clearly monogustatory). Testing with non-primary stimuli revealed sensitivities greater than those to the primary stimuli (KCl causing the greatest response for one rat neuron, MgClZ for a frog neuron). Maximum re- sponses varied greatly among the units. KEY: N = NaCl H = HCl Q = quinine hydrochloride S = sucrose HAc = acetic acid (frog) K = KCl Ca = CaCl2 Mg = MgCl 2 38 39 rat and CaCl2 and MgCl2 of the neurons, in some cases with responses that were greater than to for the frog) reveals additional sensitivities any primary stimulus. Testing was conducted with unequal stimulus concentrations and concentrations that are of unequal effectiveness. Investigators have repeatedly attempted to classify neurons according to their observed response specificities. These schemes have varied even when applied to the same species, depending on the authors' choice of stimuli. For example, Pfaffmann (1941), testing mainly the four primary stimuli, found "acid", "acid-quinine", and "acid-salt" fibers in the cat's chorda tympani and mentioned their sensitivities to 03012 and KCl only in passing. Kusano (1960) tested frog glosso— pharyngeal taste neurons with ten different monovalent cations, four divalent cations, HCl and acetic acids, and sucrose and sodium saccharin. He found four fiber types: "D-units", most sensitive to divalent salts; "Mrunits", most sensitive to monovalent salts, and "Q" and "A" units, with specificity to quinine hydrochloride and acetic acid respectively. He added that "D—units" were also sensitive to sucrose, quinine, acetic acid, and some monovalent salts. Frank (1973) has developed a classification based on single con- <:entration tests of each of the four primary stimuli. The responses to enach stimulus are rated as a percent of the largest response. The JLatter is used to classify the fibers as "sucrose—best", "salt-best", Gate. The responses of Figure 8 are transformed in Figure 9 according tic the Frank classification. Transformation of the data in Figure 8 "Vas accomplished as follows: 1) Only the responses to the four primary Figure 9. 40 Best taste profiles of nine single gustatory neurons. The response profiles for the single neurons of Figure 8 are transformed according to the procedure of Frank (1973) to determine their relative sensitivity to the four primary stimuli. Only the four primary stimuli are considered, the response to each being expressed as a percent of the maximum response to sucrose (S), NaCl (N), HCl (acetic acid for the frog, H), and quinine hydrochloride (9). -rg'l'IHS‘F . .--.-5.-af I .‘37' 41 Il‘ »>-- p»; . 1 1 l , . y o ‘ l b .l, .4... ‘T I l 1 A __,.._. , “4...... . n I I . 11.0.} . 42 stimuli were considered. 2) The maximum response to one of the four was identified for each unit (e.g., NaCl for rat neuron A). 3) The responses to the other three stimuli were expressed as a percent of the maximum response. 4). The responses were displayed in the order-- sucrose, NaCl, HCl (acetic acid for the frog), and quinine. The neurons are thus displayed in Figure 9 and can be considered with respect to their sensitivities to the primary stimuli. However, the response profiles are now altered. As can be seen from the three hamster fibers, the response magnitudes are lost in transformation, the third unit (§_Hamster) shows no difference in responsiveness from the first two (A and g). Among the three rat fibers, the middle fiber (§_§§£), an H best fiber after transformation, is actually most sensitive to KCl (of those chemicals tested by Pfaffmann). The frog responses are similarly altered, especially the second (g 2323) which is classed a sucrose fiber after transformation but is twice as sensitive to MgClz. Frank (1973) calls attention to the limitations of this method and points out that the stimuli may not be homologous among all animals with respect to quality, dissimilarity, or purity of perceptual sensation. For the hamster chorda tympani Frank found four fiber types using this method Of analysis. The fibers differed only slightly within each class, only the sucrose- and salt-best fibers differing with respect to their Second best primary stimulus. The power of Frank's analysis lies in the clear demonstration of a. fiber's sensitivity toward the four primary stimuli. Thus for the haumster chorda tympani, Frank found that the sucrose-best fibers 43 responded least to the other stimuli. Quinine sensitivity was least clearly discriminated, few quinine-best fibers being found and those responding almost as well to one of the other stimuli. The lack of specificity to quinine was contrasted in a later study of rat glosso- pharyngeal and chorda tympani neurons (Frank, 1975), in which quinine- best chorda neurons were broadly sensitive (to HCL and NaCl), but glossopharyngeal quinine-best fibers were very specific. This is consistent with the whole nerve response of many mammals in which the tongues show regional sensitivity to the four primary taste stimuli. Bitter stimuli are most effective at the posterior lingual area which is innervated by the glossopharyngeal nerve; whereas, sweet stimuli excel at the tip of the tongue which is innervated by VII. Regional specificity was also demonstrated in the tests for independence of the distribution of the specific and combined sensitivities (Frank, 1972, 1973). While the degree of specificity (i.e., the number of effective Primary stimuli) was again random, certain combinations of sensitivities were lacking or overemphasized. More units were sensitive to quinine along withNaCl or HCl and far fewer fibers were sensitive to sucrose along with other stimuli than predicted by a random polynomial distribu- tion- Sucrose-best fibers were more specific, while quinine-best units Were more broadly tuned than predicted by random distribution. Single concentration tests are frequently used to draw correlations betWQen different stimuli using graphical and statistical methods. Moderate correlation coefficients have been found between pairs of primary stimuli (r = +0.30 to +0.58). Sucrose is an exception for 44 chorda neurons with "r" values that were negative or less than +0.20 (Erickson 2391., 1965; Frank, 1973). Correlation coefficients have also been used to compare the effects of non-primary stimuli. Strong positive values were found between NaCl and LiCl for the rat chorda fibers (1' = +0.92, Erickson £5 31., 1965) and between HCl and citric acid for hamster chorda fibers (r = +0.82, Frank, 1973). The use of correlation statistics with single concentration tests is of limited value. Investigators may be tempted to extend correla- tions between fiber sensitivities to physical properties of the stimuli that may underly perceptual taste quality. Yet the correlation statis— tics do not discriminate between properties that are ascribable to the fiber systems themselves (such as a specific insensitivity to a class of stimuli) and the properties of the stimulus molecules that are associated with a particular taste quality. For example, as cited above, rat and hamster chorda neurons lacked specificity to quinine, and the rat chorda neurons showed strong correlations between responses to quinine and organic acids (acetic and citric, r = +0.72, +0.66 respec- tiVely). Erroneous generalizations would be that acetic and citric acids have bitter perceptual properties to the rat or that the physical pa”:altimeters responsible for "bitterness" are common to these three 8tilllllli. Rat glossopharyngeal nerve fibers, in contrast, showed Specificity to quinine. For the glossopharyngeal fibers, the bitter SUbstances, urea and caffeine, were strongly correlated with quinine (r = +0.59 and +0.60) while the acetic and citric acid correlation co— effiCients fell drastically (r = +0.07 and +0.04; Frank, 1973, 1975). 45 Single Unit SR Functions. Several investigators have determined a single neuron's gustatory response to test stimuli through a range of concentrations. While no report shows a systematic investigation of several stimuli using this method of presentation, enough individual examples have accumulated to indicate several properties of single neuron SR functions. SR functions of single neurons often differ from the SR curve of the sumated whole nerve response. With the same stimulus, the thres- hold, slopes and maxima of SR functions vary among fibers (Pfaffmann, 1955; Cohen gig-11., 1955; Fishman, 1957; Ogawa 3511., 1968; Wang and Bernard, 1969; and Ganchrow and Erickson, 1970). Examples are shown in Figure 10. In Figure 10_A_ are the SR functions of four different rat chorda tympani fibers to NaCl (Pfaffmann, 1955). For all neurons, responses increase regularly with greater stimulus concentration, which is the most frequently observed relationship and the form of whole nerve SR functions. Figure 10§ and 10_C_ shows two frog glossopharngeal fibers for which the NaCl SR functions are quite different, one with negative slope and one which is U-shaped with a mid-range minimum, (KUSano, 1960). A bell-shaped function for NaCl is shown in Figure 102, cat chorda tympani neuron (Bernard, 1972). For the same neuron, SR functions can differ with different Stimuli, even among the four primary taste stimuli. In Figure 11 (A-C) are shown SR functions for individual neurons of the rat (Ogawa _e_t_ 31., 1969), hamster (Frank, 1973), and cat (Bernard, 1972). Nearly all the curves are positive and monotonic, showing different threshold and Figure 10. 46 SR functions for single neurons. These are shown for NaCl responses in seven different neurons of the rat, frog and cat (after Pfaffmann, 1955, Kusano, 1960, and Bernard, 1972, respectively). As can be seen, they possess widely different forms. In A_the responses from four rat chorda neurons form regularly increasing functions of concentration, differing in threshold and maximum response. In §_and §_are two frog glosso- pharyngeal fibers with U-shaped and negatively sloping curves. In Q_is a bell—shaped function for a cat chorda neuron's responses to NaCl. The ordinates are not scaled equally but reflect each author's method of magnitude estimation. The abscissae have been standard— ized to Log concentrations to facilitate comparisons. Q_also includes the level of spontaneous activity (mean i.1 standard deviation, shown by arrows). Frequency impulses/2 sec impulses / 2 sec Average spikes per 5 seconds 20 47 - 3 -2 -1 0 log M NaCl Frog f— I l I 1 - 3 -2 -1 0 log M NaCl Frog {7 ‘ 1 - 3 -2 -1 0 log M NaCl Cat IL'] ' I ' I ‘ I - 3 - 2 -1 0 log M NaCl Figure 10. 48 Figure 11. The single neuron's SR functions to different stimuli. These are shown for a chorda neuron of rat (Ogawa g£_§l,, 1969), hamster (Frank, 1973), and cat (Bernard, 1972) for the four primary stimuli as well as sodium saccharin (for the rat, N§§_in A). While most functions have only positive slope, two are complex forms (_N_ for the rat, _l_i_ for the cat) and one is of zero slope (Q for the hamster). With such variety of forms and slopes, classification of the neurons according to the most effective stimulus is arbitrary since the stimuli are ranked differently by threshold or maximum response for each neuron. A11 re— sponses were averaged over five seconds. §_shows the spontaneous activity level (mean 1 one standard devia- tion, by arrows from the ordinate). NaCl HCl quinine hydrochloride sucrose sodium saccharin Key: cncn&3:=:z 49 Rat § 373‘ impulses/ 5 sec 8 Has __|_L._.J.I__I._J PL—L__L -4 -3 -2 -l 0 log Concentration Hamster ss§ impulses / Ssec ’ p L i . -4 -3 -2 -l 0 log Concentration impulses/ 5 sec L l _J 4 L | l I I “'4 -3 -2 -_l 0 log Concentration Figure 11. 50 maxima. Noticeable exceptions are the NaCl SR curve for the rat, the HCl curve for the cat, and the quinine hydrochloride curve for the hamster. In the latter case, the curve for the quinine SR function had zero slope (Frank, 1973 for the hamster, Figure 1113). The responses to quinine were of a consistent, albeit low, level. The author considered the curve to represent the absence of any response, without stating the spontaneous activity level for this neuron. Even with the variety of functions observed, Ganchrow and Erickson (1970) showed that the combined SR curve of a sample of 20-37 rat chorda neurons (averaged for each concentration) closely matched the reported summated whole nerve function (of Pfaffmann, 1955) both with respect to threshold and slopes for the four primary stimuli and KCl. While the individual SR functions were not shown, the authors reported them to have different thresholds and rates of increase. The classification of a fiber from the full range of its SR func- tion can be most complicated. For example, in Figure 115, the rat chorda unit could be a quinine fiber (lowest threshold) or an NaCl f 1ber (with maximum response) and in Figure 119, the cat fiber could be HCl—best or a quinine fiber, without knowing the HCl threshold. Each of the fibers responded in varying degrees to three of the primary 8‘13—‘11".111 and in the absence of further information such as the response la‘iettcies, suppression responses, or rinse activity, can only be arbitrarily classified. Recalling the body of literature relating to correlations and comparisons from single concentration studies, Frank's ca“tions are appropriate (Frank, 1973). She states that if many fibers 51 show drastic response reductions with increasing concentration, severe limitations must be placed on interpretation of data taken at any one moderate concentration and that if stimuli are variably strong or weak, different conclusions with respect to fiber specificity will be made. Two experiments using extended concentration tests have shown that bell-shaped and negatively-sloped SR functions are not derived from arbitrarily strong or weak responses but are consistent functions. Moreover, these can be powerfully related to other SR functions and different response parameters to reveal more information about the taste system and its interaction with several stimuli. Cohen 25 El. (1955) recorded from cat chorda neurons and observed prominent differences between two fiber types. The first showed spon- taneous activity under adapting Ringer's rinse. It was also salt sensitive and responded with a regularly increasing SR function to NaCl with a threshold of 0.1 M (Figure 125, "Salt Fiber"). The second neuron responded to distilled water (after Ringer's adaptation) and remained spontaneously active during water adaptation. To NaCl it responded with a regularly decreasing SR function with its responses declining to zero at 0.03 M NaCl (Figure 12A, "Water Fiber"). The two neurons possessed disparate sensitivities; distilled water did not affect the "salt" fiber and the "water" fiber did not respond to NaCl concentra- ‘tions that were excitatory to the other. Seen together (Figure 12A) the SR functions of the two neurons are related, the responsiveness Passing from one to the other as salt concentration is increased. The allthors observed that discrimination over an entire range of salinity Figure 12. 52 Interrelated SR functions. Negatively sloped and bell- shaped SR functions have been correlated with other SR functions and different response parameters to reveal more information about the taste system. A. B. Two fibers with opposed NaCl SR functions are seen to have maxima at opposite ends of the extended con- centration range. Together the fibers' activity offers a broad sensitivity to salinity. A single neuron's responses to sodium saccharin form a bell-shaped SR curve (N§§_ngr curve). Its rinse responses to sodium saccharin have an SR function whose threshold is at the concentration for the maximum response to sodium saccharin (at 0.03 M, Na§_"gfffl curve). From 0.03 to 0.3 M the responses to the application of sodium saccharin decline while its rinse responses proportionally increase. The authors suggest that this reciprocal relationship shows that "on" and "off" responses are related by a mechanism which can excite the neuron or inhibit it by the action of sodium saccharin molecules. Impulse: per Second N O U ,_. G 1 p O “Water Fiber" 53 “ Salt Fiber” 150 lmpulses / 5 se: p O O .3 0--//—-* x-——- '3 -2 -1 log Concentration NaCl “IS “on" /* x S \ / x iJ—‘x/ - *NaS “off" .2 /_'1 0‘ log Concentration Figure 12. S4 is thus afforded by two distinct fibers with response maxima at opposite ends of the range of concentrations. The second example, from Ogawa g£_al: (1969) correlates different responses of a rat chorda tympani neuron with sodium saccharin. The neuron was classified as a sweet sensitive fiber by its criterion responses to sucrose over an extended range of concentrations forming a positive SR curve (Figure lZB, curve S). However, its SR curve to sodium saccharin was bell-shaped, with a threshold below that of sucrose (Figure 12§,.N§§_"ggv curve). The maximum response to sodium saccharin was at 0.3 M and was equivalent in magnitude to the sucrose maximum at 1.0 M. The neuron also gave a rinse response after sodium saccharin, but only to concentrations that were greater than 0.03 M. The SR func- tion for the rinse response was bell—shaped (Figure 12_I_5_, Nag "_o_f_f_" curve). The positive sIOpe was nearly reciprocal to the declining por- tion of the sodium saccharin "on" function. Thus the negatively sloping portion of the SR function for NaS "on" was correlated with the rinse response SR function by: l) the reciprocal slopes of their SR curves, and 2) the coincidental maximum of the sodium saccharin "on" responses 'with the threshold for the "off" reSponses. The authors suggest that these phenomena represent inhibition (the declining sodium saccharin responses from 0.03 to 1.0 M) and the removal of inhibition (the rinse ‘response over the same concentration range of sodium saccharin). They ‘pmopose a mechanism involving the dual action of sodium saccharin mole- cules at two distinct receptor sites, one which excites and one which inhibits the gustatory neuron. 55 Mechanical and Thermal Sensitivity of Taste Neurons Many gustatory neurons respond to stimulation by two other modali- ties, mechanical and thermal. In mammals, mechanosensitivity appears to be confined to afferent fibers in the lingual branch of the trigeminal nerve. Chords tympani fibers do not readily respond to mechanical stimulation. In the glossopharyngeal fibers of the sheep (Iggo and Leek, 1967), movement of a circumvallate papilla can enhance the neuron's chemical response though this may simply be a result of enhanced diffu- sion of the stimulus. The chemical sensitivity of trigeminal nerve fibers has not been clearly established. Degeneration studies (Beidler, 1969 and Miller, 1974) showed that the trigeminal nerve largely inner- vated the stalk and walls of the fungiform papillae of the rat. While no trigeminal neuron was clearly shown to innervate the taste buds (in the rat) the possibility cannot be completely excluded. In lower vertebrates, e.g., the frog, glossopharyngeal gustatory neurons also respond to mechanical stimulation. Taglietti, Casella, and Ferrari (1969) have demonstrated that the receptive field of a neuron to mechanical stimulation, localized to several fungiform papillae, differed from the chemical receptive field, though the latter ‘was determined only with CaClz. Thermal sensitivity is commonly found in gustatory neurons. (For (examples in the rat, hamster, sheep, and macaque monkey, see Fishman, 31957; Iggo and Leek, 1967; Sato gt_al,, 1969; and Sato EEHEL-a 1975 irespectively). Taste fibers that are sensitive to lingual warming are the exception for most species, while sensitivity to cooling is often 56 found. Sato'gt_al, (1969) found that 48/50 gustatory chorda fibers in the rat were cold sensitive. Ogawa g£_al, (1968) observed a U-shaped thermal response function for a rat gustatory chorda neuron with a minimum at 35° C, ten degrees above lingual surface temperature. The authors concluded that the 250 tongue temperature (excitatory to such thermally sensitive gustatory units) may be the origin of spontaneous activity. Sato 3; a1. (1969) found large correlation coefficients be— tween spontaneous activity and cooling (r = +0.503, +0.781, rat and hamster) with significant differences (t test, p<10.01) between the spontaneous activity of cold sensitive taste fibers and non-thermally sensitive chorda neurons. The macaque monkey is an exception to the relatively large number of warm-sensitive chorda gustatory fibers (30%, versus 40% cold- sensitive). The warm and cold fibers responded to distilled water at temperatures above or below lingual temperature and together offer an extended range of thermal sensitivity (Sato £2 51., 1975). III. STATEMENT OF THE PROBLEM The neurophysiological taste response of the single gustatory neuron of the mud puppy, Necturus maculosus, has not been investigated. The taste response of single neurons of many vertebrates has been found to vary widely. 1. Three forms of the taste response of single neurons have been found: a) increased activity on stimulation, b) decreased activity with stimulus presentation, and c) responses to the water rinse after stimu- lation. Yet only the first form has been consistently recognized and quantified in analyzing the responses of the first order gustatory neurons. One objective of this study was to consider all forms in defining the taste responses of the single neurons of Necturus. 2. The magnitude of the taste response varies with stimulus con- centration, defining the stimulus-response or SR function of a neuron for the taste stimulus. A single gustatory neuron has different SR functions for different stimuli. Among many neurons, dissimilar SR ‘functions have been found for the same taste stimulus. Yet most taste sensory experiments employ single concentrations of stimuli. In part this arises from the limited ability to hold and test a single neuron. A small number of stimuli have been tested at one concentration to allow greater replication. A systematic investigation of SR functions 57 58 among many neurons has not been reported and is the second objective of this study. 3. Recent experiments have found that the latency of the taste response of single neurons is inversely related to the concentration of the stimulus (T. Sato, 1976). However, this relationship was based upon the average latency of the responses of several neurons. The responses of individual neurons were seen to have a wide range of latencies at a single concentration. The latency functions of the individual neurons were not investigated and is the third objective of this study. Accordingly, the experiments of this study were undertaken to define the taste responses of single gustatory neurons of Necturus according to three different parameters: 1) the form of their taste response, 2) their SR functions, and 3) their latency—concentration functions. The four primary taste stimuli were employed: NaCl, HCl, quinine hydrochloride, and sucrose. .These were not used to classify the neurons according to their sensitivities toward these stimuli, but were used to obtain a broad survey of taste responses from dissimilar taste stimuli. With single tests, the systematic investigation of these parameters sacrificed replication to gain breadth of testing by using the four stimuli over an extended range of concentration, 3.5 Log. IV. METHODS Smaller adult mud puppies, Necturus maculosus, 20-25 cm long were selected. These had been held in 20 gal (76 2) plastic aquaria at 10-150 C in aerated, charcoal-filtered water in an artificial light cycle of 16 hours per day. Animals were held no longer than three weeks and were not fed during confinement. The antibiotic, tetracycline was dissolved in the aquarium water to eliminate possible infection by Saproleggia fungus (see Samanen, 1973). The animals were anesthetized with 5 gm Z (w/v) urethan (aqueous ethyl carbamate solution) in two stages. The first was by whole body immersion for a maximum of five minutes to facilitate handling. The animals were then removed and rinsed with distilled water. Their tongues were extensively rinsed to reduce any depression of the taste system. For the second stage, their gills were placed in cups contain- ing the urethan solution for an additional 10-15 min. Experiments were conducted at 22—240 C ambient temperature. The dissection included first the exposure, transection, and de- Sheathing of the peripheral branches of the glossopharyngeal nerve along the ventral aspect of the ceratohyal cartilage of the hyoid arch. Each branch was further divided into smaller (80 u diam) bundles. Standard differential monopolar recording techniques were employed using two identical, 300 u diam, silver electrodes. The active electrode 59 60 conveyed the electrical potentials from the isolated nerve bundle. The indifferent electrode contacted nearby tissue rendering an optimal signal—to-noise ratio. Both were immersed in paraffin oil. The animal was grounded from its tail to the preamplifier's ground circuit. The Grass P-15, AC, preamplifier (Quincy, Massachusetts) differen— tiated the potentials and then amplified (X100) and filtered the activity through a frequency bandwidth of 300 Hz-BKHz. A Tektronix, Type RM502-A, dual beam oscilloscope (Beaverton, Oregon) displayed the electrical activity which was then recorded by a Magnecord, Model 1043, stereo, AM tape recorder (Midwestern Instruments, Tulsa, Oklahoma)(see Figure 13). ,ngputer Determination of the Isolation of Neurons, FREDSAM In order to determine that the recorded activity was from a single nerve cell, a computer program, called FREDSAM, was used. The only extracellular evidence for single neuron recording was the uniformity of action potential amplitude. Yet variance in spike amplitude is present in all records, arising from the cumulative effects of the recorded electrical noise. Accordingly, a computer program written by Mr. Marc Schneider for the LINC 8 computer (Digital equipment Corpora- tion, Maynard, Massachusetts) was used to quickly and accurately determine the distribution of action potential amplitudes. The come puter program named FREDSAM ("frequency distribution of spike ampli- tude") analyzed all diphasic potentials in the taped records, both the action potentials and the much smaller, irregular baseline or back- ground electrical noise. The FREDSAM analysis was displayed as a 61 .Ammmu mucmaaoo m.umusoawuoaxo on» mo mam .maomo m.uoaHu moasawum can «0 .mflmev usauso m.omoomoaawomo onu mo momma ow mucous uawsmauom < .oooomoaafiomo use Howwwaaam Iona can an omnmwaaaooom ma AccouomwwocH vam.mwmmm¢v mooouuowao uo>HHm 03u aoum huw>wuom oSu mo cowumowwfiaasm u< Hmwuaouomwfin .wswouooou uwas amazon mesfim xuoumumsw How usoaaflsoo oHoouuoon .ma muswsm 62 .2: tou==ao .2 o... nlll.’ :0 «.5 33 «.5 3...; "so 35:. Illa «=0 :- _8_ro> ® 2.33593 .Mu 95w: QC 32.3.58... 25 \ :3 _s 2.23:9... 32.22:: .2. 253 £ 63 population histogram (Figure 14). The bins of the histogram are dis- played as single vertical lines along the abscissa. The width of the bins are based on a relative scale of amplitude which was calibrated for each experiment (Figure 16). Once the data was entered into the LINC 8, several options were available to vary the display, including: 1) scaling the ordinate to show from 25 to 6400 events, 2) displaying a limited number of bins, e.g., a left or right portion of the abscissa, or only the bins which encompass a single population, and 3) varying the bin width from 1 to 999 relative amplitude units per line. The latter option was always used to ensure the persistence of maxima and minima by changing the bin width over several relative amplitude units. (For other options, see Appendix I.) The largest peak, displayed near the origin, was the small (usually 10 UV) background electrical noise. Uniform action potentials, indicating a well-isolated neuron, formed a distinct peak to the right of the baseline population. Multiple unit records had histograms with several peaks. When electrically distinct, they were separate from each other (Figure 15). The peaks of poorly isolated neurons were much smaller and less distinct. The histograms of the neurons of low responsiveness similarly had no clear peaks but only the random display of a few points. Therefore, records from poorly isolated neurons or those with low activity resulted in FREDSAM histograms with only a distinct baseline peak and its positively skewed tail and were rejected from further analysis (see Appendix I). Figure 14. 64 Computer analysis (FREDSAM) of a well—isolated neural unit. Above. The action potentials from a well-isolated lingual neuron from three cold tests (4° C distilled water, only 1 of the 10 second tests is shown). Their diphasic amplitudes are uniform at 35 uV). Below. The display of the FREDSAM analysis of the three tests giving a histogram of the distribution of diphasic amplitudes. Each vertical line represents the action potentials within a single interval of amplitude. The ordinate shows the number of events, the abscissa shows intervals of increasing amplitude. Amplitude intervals are defined on a relative scale (here each equals 0.6 uV). A calibrated scale is shown below the relative scale. Two distinct populations are seen: 1) the population of action potentials (right) and 2) the population of small baseline potentials (left). The legend of the FREDSAM display contains: a) a two-line statement identifying the display, b) a row of numerals indicating-- 1) the maximum ordinate value (lQQ_events), 2) the total number of events displayed (625), 3) the width of each interval (here equals 1 relative amplitude unit), and 4) the lowest interval number displayed (1, at the origin). 65 20 W 100 mSec I... m "r :2; m 1;. .1 NUMBER OF EVENTS 0 n 1 so 100 RELATIVE AMPLITUDE '{ 1'0 2'0 3'0 4'0 s'ofi AMPLITUDE, uV Figure 14. 66 .oomuu some: onu aw omamauaoofi mum «mum .mamaucouoe cowuom wsfioaoemouuoo ona .uwas o>HuHmcmm Hmahosu m Scum momsoemou on use was AH hmaamwvv H oswaommn use: «M .asafixma can um£u caucuses“ was Ammmv ocHHmmmn Hmowuumsahm m wsHHmo>ou .msousoc 08mm mnu Bouw noncommou os oomsmo Ammouosm z m.o was .oofiuoanoouoxn osaaaso z no.0 .Homz z H.o .Hom z Ho.ov coaumaaanum amusemeu Hmsmaasiumuxmmmmmm .onom amazonwuo: mean was on: use xm madmofiuuo> emamwamam ow was Asouum Moods onu mo uanuv H hmammqa mo soauuom unwau mnu ou ucoam>fi=vo mflIIN NmHmwwa .mm was dmmv muso>o mosuwaaam Hanan mo osHHommn Hoowuuosshmm so was mm «m mmv macaumasaom ovsu IHHoam uoswumao oousu nuas sowusnauumflo Hmooslquasa m macaw noncommou Hmauonu memo onu mo mmoauoo Homaoa mo mammamsm zom Scum mamaufiouom soHuoo sues omvsommou :osmun HmswsHH mltoomuu Home: .muao: woumaoma sameness mo mammamsm zuomou woman can msHsaHum onu mo unwfioa an voHHouusoo mm3 Bon mo mama .usoao umou onu mo uso no uoxosn momma can ousH moasoaa oau osu mo moo muoouuv haoumauouam .uoafiu o Scum Hmcwwm so was vommouaaoo up Go>fium m>am> m ouoss.w.0usa muao>uomou Hausa aouw 3oam mwfisvfia nuom .mmsfiu nouns omaafiumfiw magmas «NV was .mm no mm mwv sowusaoo umou mo coauooaom onu sodas afiv ”scan: uaoamfisvo oumuumsaaw U was .m .< maosmm .aoumho muo>HHoo msHaaHum one .NH magmas .2 as»: 74 record of each test (20 second test plus 5 second pre- and post-stimulus rinse) from the magnetic tape. The action potential photographs were compared to the computer data analysis. The sine wave calibration signal was photographed for correlation between computer and photographic records (Figure 16). Post-Experimental Computer Data Analysis, SAM-COUNT. The object of analysis was to determine the total number of action potentials from a single neuron that occurred during a test and to compare this to its pre—stimulus activity during an adapting water rinse. Post-stimulus impulse counts were also enumerated to observe whether activity was maintained into the rinse period or whether rinse responses occurred. Impulse interval analysis was beyond the computer program's ability. A second computer program, written by Mr. Schneider, utilized the amplitude parameters of FREDSAM to define the responses attributable to an individual neuron. The operator analyzed the neuron's activity after the experiment by selecting an amplitude window that defined the particular neuron according to FREDSAM. The enumerator program, SAMPCOUNT (an acronym for "spike amplitude-count"), used the timer pulses to allow the operator to enumerate the activity over partial or whole periods as: a) the number of action potentials in a complete period (between timer pulses), or b) the number that occur immediately before or after a period. The latter partial analyses could be for any predetermined length in seconds up to a complete period. SAMeCOUNT rejected those potentials that were not included in the amplitude window defining the neuron. The SAM-COUNT listing presented the number 75 of potentials from the neuron within the pre-designated test or rinse intervals in their order of occurrence. The taped record of each experiment was played once for every unit under analysis. All counts were compared to the photographic record to l) eliminate discernible artifacts of the same size as the unit under analysis, 2) to reassign potentials which occurred between the initiation of stimulus delivery and the arrival of the test solution at the tongue, and 3) to reassign those action potentials which occurred between the initiation and arrival of water rinse. (See below for the method of measuring the solution delivery time, and Appendix II, SAM-COUNT, Evaluation and Background.) Graphical Analysis 2; Enumerated Response Activity From the SAMACOUNT listing, the 20 second test activity was graphed with the stimulus concentration to determine the SR functions of the neurons. The 5 second pre- and post—stimulus activity was used to indicate the spontaneous activity level of the preparation and the rinse response magnitude, respectively. Longer evaluation of both pre— and post-stimulus periods would have been desirable but was not possible since the magnetic tape record was often only clearly obtained for 5 seconds before and after the tests. For the latter case, the rinse response magnitudes that were calculated from 5 second post-stimulus activity were not seriously altered from 20 second estimates for their response pattern was very phasic, adapting completely within that time. In the former case, the 5 second averages of pre—stflmulus or spontaneous 76 activity levels for most tests with most neurons were 1 or zero impulses/ 5 sec. With this value, most responses of increased activity exceeded criterion levels. The mean spontaneous activity + SE are displayed with each graph of the SR functions along the ordinate by an arrow (for example, Figure 22). For the responses of decreased activity, this criterion value was too small for statistical evaluation. Accordingly, for units with low level spontaneous activity, the photographic records of the tests were used for evaluation of decreased activity responses as in Figure 20. On occasion, longer determinations of pre-stimulus activity were required for a second reason. In this case, the pre- stimulus activity levels were larger and/or the test series succeeded a particularly effective stimulus and the test counts were suspected to reflect only the lingering activity from a preceding large response. In this case, 20 second pre-stimulus counts were obtained on separate runs with SAM-COUNT when the tape record allowed. Each interval of pre- stimulus activity that was used in analysis is noted with the graphs in the Results section. Response Latengy Measuring;DeliverygTime. The timer pulses identified only the onset of events which led to the delivery of solutions to the tongue. Several hundred milliseconds were required for the compressed air to reposition the air valve, for the next solution to flow through the delivery tube aimed at the tongue and for the solution to contact the lingual surface. Although the delivery time varied across experiments, it remained constant during each, since the delivery system and animal's 77 head were securely clamped in position. The circuit in Figure 183 was employed to measure the exact delivery time. A Grass SD9 stimulator (Quincy, Massachusetts) supplied 50 mV, DC current through two leads (one placed in the stimulus reservoir, the other placed directly on the receptive field of the neural unit being examined). The circuit re— mained open as long as distilled water flowed through.the system to the tongue. When an electrolyte replaced distilled water, the circuit was completed and the oscilloscope displayed a change in potential (Figure 1899. The timer pulse that marked the onset of flow of conducting solution (0.1 M NaCl) also triggered the oscilloscope sweep. The time required for the trace to begin its deflection thus represented the stimulus delivery time. The mean of 2-3 determinations was measured from a photograph of the oscilloscope display. Delivery time determina- tions were made after the experimental tests to avoid any effects of current on the gustatory response. Conduction Velocity. The calculated latency values were not corrected for conduction velocity and no conduction velocity determina- tion was made during the experiments. The time of conduction was estimated to be negligible considering the diameter of the fibers found in the glossopharyngeal nerve trunk. The myelinated fiber diameters were found to range from 1-15 u, with a median of 3p (Samanen, Kryda, and Bernard, 1975). This roughly corresponds to conduction velocities of 6-90 m/sec, with a median value of 18 m/sec. A maximum estimate for the distance between the tongue and recording electrode is 2.5 cm which would have required less than 5 msec for the smallest fibers. Figure 18. 78 Measurement of solution delivery time. In A, wire leads La and L2, placed in the stimulus reservoir and on the tongue, act as contact points of a switch in circuit B, The circuit, which includes the oscillo- scope, y, and a 50 my power source, is completed by redirecting the 0.1 M NaCl solution onto the tongue. The oscilloscope trace, 9, triggered with the initia- tion of saline delivery, shows a change in potential (white arrow) at 280 msec identifying the arrival of the solution. 79 Stimulus Reservoir L1 Air Valve (Positioned by Timer) Figure 18. 80 Moreover, this small error was constant for any given unit and could only be a factor in inter-neuronal comparisons, analysis which was not attempted from these experiments. Determiningrfhysiologégal Latency. Response latency was deter- mined from the photographic record by subtracting delivery time from the interval of time between the timer's onset pulse and the first action potential shown on the photograph. This remaining period of time was taken as the true physiological latency. Action potentials that occurred after the timer pulse but within the delivery time were subtracted from the computer counts of the response. Similarly, action potentials occurring between the rinse timer pulse and rinse delivery were reassigned to the test period from the computer's count of the post— stimulus rinse period. V. RESULTS The responses of 34 electrically isolated lingual neurons are described below. Of these, 14 were identified during the experiment by the action potentials that were seen on the oscilloscope occurring in response to chemical and tactile stimulation. These possessed receptive fields (20-40 mmz) within different portions of the gustatory receptive field of the whole glossopharyngeal nerve (see Figure 4). Post- experimental analysis of their activity with FREDSAM revealed that 11 records were actually of multiple neural units (a range of 2-4 individual neurons in each record, with a mean of 2). The activity of the companion neurons of smaller diphasic amplitude had not been clearly observed during the experiment from the oscilloscope display. Nonetheless, their electrical isolation and chemoresponsiveness showed them to be accept— able and their responses were also evaluated with individual analysis with the SAM—COUNT program. Seven other experimental records were deemed acceptable during the experiment and even after testing by the evaluation of their photographic records. Yet the FREDSAM review of their tape recorded activity showed them to be poorly isolated and they were rejected from further analysis. The mean amplitude of the extra- cellular recorded action potentials of the isolated neurons ranged from 14-139.5 uV, with 27 of the 34 neurons having a more restricted range of 14-76 uV. 81 82 Spontaneous Activipy The afferent neurons could be grouped according to three different levels of spontaneous activity. The first and most common group was made up of 29 gustatory units which showed low rates or no spontaneous activity (0 impulses/5 sec for 657 of 844 tests). Only one neuron had a zero spontaneous activity level before each test. The inter-trial spontaneous levels for the other fibers varied with stimulus and/or concentration and ranged from 0-13 impulses/5 sec (0—2.6 Hz). The mean spontaneous activity over all tests of the 29 units in this group was 0.0125 impulses/5 sec. The second and third groups of neurons displayed high levels of spontaneous discharge (1.25-30.00 impulses/5 sec). The second group consisted of fibers from a more posterior branch of the glossopharyngeal nerve. The activity of this branch was of a continuous irregular pattern (see Figure 19A). In one experiment, three neurons fired at rates of 24.3, 10.9 and 4.8 impulses/5 sec. The origin of this dis- charge is unknown and could not be altered by chemical, thermal, or mechanical stimulation of the tongue. Other lingual stimulatiOn or stimulation of other oral or extra-oral regions was not attempted. Nonetheless, these may be considered a class of lingual afferent fibers since the recording was always from the severed distal portion of the branch. Of the 34 units recorded in the experiments, three fibers were of this class. Many more might have been studied but not being respon- sive to gustatory stimulation, they were abandoned when encountered in later experiments. 83 .Esawxma ou maamsomuw moawsn uncommon on» .musouma osooom oBu m nouw< .oaomo msasaaum was mo uomso onu mafia moofioswou soups one .Amuoosss sauna he oofimausooa www.mw oom.M .muws: umwuma osuv Homz z m.o ou maousoo enoumumsw Hmuo>mm mo uncommon may .Asaonm uosv oom OH ca mHo>oH msassfiumlmum ou wsauomom .oammno ma snouumo uncommon may .omsau nouns ooHHHumHo nouns oawsfiso wcwusomouo .oaomo moanswum ocu mo noose oeu oofiMHuGoofi sound was .Aaomov osauoanoouohn ocwsfiso z H.o ou mm mam .n em .maousos muoumumsw moans mo uncommon one .aowumasawum Hmsmswa he oouoommm no: mums was com m\mooasoaa w.e was m.oa .m.qN mo mouse on oouww omocH am was mm .Aouooou ecu mmouum cacao ma zuw>wuom moonsv.w .omamwusoofi one means mossy .sosmuo Hmswawa uoHuoumoa m aouw momma ewe: um owumeomfio msoosoucoom moossfiusoo no .msousos unassumm Hmowshumnmommoaw mo monsoomou one huw>wuo< .ma magmas 84 .IIIIIIJ gem . .3 as: 52;. so + cafes o .25 2 S t cusses mtg: outs—.33 32.—Ego _ N/N/ __ _ 85 The third group of fibers fired at various pre-stimulus rates, which ranged from 0-14.75 impulses/5 sec. This group consisted of two companion fibers found in one experiment whose spontaneous activity declined throughout testing. The larger fiber fired initially at 12.75/ 5 sec, declined to a steady range of 3.5-8 impulses/5 seconds between tests, and then dropped to zero spontaneous discharge and became unre- sponsive after 18 tests. The smaller neuron showed a similar decline, but started at a lower rate (3.5 impulses/5 sec) and remained responsive through all tests despite its declining spontaneous discharge. Despite their unique occurrence, their responses are reviewed and discussed below, since they epitomize certain forms of gustatory response occurring in more typical taste fibers at drastically lower frequencies. Forms of Gustatory Response The three previously reported forms of gustatory response were observed: 1) increased activity upon stimulation, 2) decreasing activ— ity upon stimulation, and 3) increasing activity with water rinse after stimulation. The most common response was increased activity upon stimulation which occurred with different impulse patterns. Even with- out impulse interval analysis, this could be judged from the photo- graphic records. Regular bursting was not obviously seen. Peak response frequencies usually occurred with response onset but in some cases they increased gradually over many seconds (see Figure 19C). The largest response was 344 impulses/20 sec. A more typical response was 10-30 action potentials during the 20 second test period. 86 For the common gustatory unit with low level of spontaneous activ- ity, the second form of response (depressed activity) was clearly observed in only one of 29 neurons. This resulted in part, because of the usually low level of spontaneous discharge and the limited ability to discern accurate criterion levels of Spontaneous discharge from the taped records. (The five second determinations of pre—stimulus activity disallowed setting more exact response criterion levels.) Depressed activity responses were abundant in two gustatory units with high level spontaneous activity (Figure 20A). The larger unit showed only depressed activity responses. The neuron with smaller ampli- tude action potentials showed both forms of response (increased and decreased activity). Its activity increased on presentation of stronger concentrations of HCl and NaCl. Its activity was depressed with sucrose and low concentrations of HCl. Rinse responses were present in both of these neurons and were most obvious for both units after sucrose tests, both also showing activity depression to sucrose application. On water rinsing after sucrose, both fibers showed rebound activity which increased above pre- stimulus levels (Figure 20A, lower record). Rinse responses were also present for four of other units: for two neurons, only after sucrose; for one neuron, after HCl and NaCl; and for its companion, after HCl, NaCl and quinine hydrochloride. Rinse responses occurred after positive responses to the stimulus for the latter two neurons, and after periods of no activity for the former two. Figure 20, A_and B, compares the rinse responses after depressed activity for a neuron with high level 87 .moumu uncommon woman mam moasafiumlouo umsoa new wcfi30£m.m Sues .Hoaamumo manomxo mH.m osm.¢.:« momdoommu mo auouuma onu mafia .Ho>oa qu>Huom msasawumloum osu o>oem How uncommon m oomsmu omcfiu nouns onu umsu mBonm ououou oossflucou any Auosoqv .sOHumHSBHum omouosm wcwuso >uw>fiuom on noes woocoomou maumHHEHm Aomswu msasawumlopo wswuso Hmfiucmuoa coauom osov mmumnomwo msoosmusoom Hm>oH 30H sows owes o mo uncommon one Anomaav . ml .mHo>oH msasawumlouo o>oom mwumsomfio cows omsfiu nouns onu ou oomdooamu accuses can Aoosswusoo ma.¢.aouw ouooouv soaumassqum omouosm no one can u< nuosoqv .huH>Huom mo scammmuensm ouoaaaoo sows sowumHsaHum omouosm ou woosoomou .omsfiu Hausa ooHHHumHo Hoods owumnomao msoocmusoom mo mHm>oH awe: mao>wumamu no“? «N osm.wv maousoc 038 AnommDV afl .momcommou woman was mufi>wuom mommouooa .om messes 88 e. . . .0N 2:3... wcH3ocm m LuHB gasses « «8.25 a so 3225 .23 t on: .23 0a: 33* 3225 I 000.0 8225 I «00.0 t 01.: um:— _ u . r . . _ . . .mwumu wmeQmmH wmdfiu 05m mdadafiumlmua HGSOH Haw .Hwaamumo >H000X0 mH.M ofim.fl 6H momsoammu mo fiuouuma mfiu 0359 ‘ i i i r 'll I‘SI‘ 89 spontaneous discharge with a common neuron with low level spontaneous discharge that gave a rinse response after no activity during the application of the stimulus. Gustatory Stimulus-Response (SR) Functions Neurons with Low Level Spontaneous Activity, More than one type of SR function was observed for the neurons. Examples of each form appear in Appendix IV and in Figures 22-24. The data were obtained using non-replicated tests in order to accomplish the extended series of tests within the critical life-span of a preparation, usually one hour, before desiccation or other factors altered the neurons spike amplitude or responsiveness. Accordingly, neither the slope nor the regularity of the SR functions could be determined statistically. With lack of rigorous statistical determination, many SR functions were judged to be indeterminate on the basis of their complex and irregular graphs, some of which were seen to fit an overall positive or negative trend (e.g., see Appendix IV, Figure IV-l, quinine hydrochloride test series for neuron 21.1). Others were very complex and were also judged to be of indeterminate form (Appendix IV, Figure 39). 0f 29 units, 28 had at least one definable SR function (not including the two units of high level spontaneous activity). For the 29 common gustatory neurons with low level spontaneous discharge, 155 SR functions were examined, of which 67 were of identifiable form, 59 were indeterminate, and 29 were of no response (see Table II, with "Gustatory Latency Functions" in this section). 90 The gustatory neurons with definable functions most often (39 of 67 definable functions or 39/67) had positively sloped SR curves. Of these, 19/39 had the form of the whole nerve's SR function, having curves with positive slopes. Their responses showed a direct relation- ship to concentration, with larger responses being obtained with greater stimulus concentration. In Figure 21 is the photographic record of this type of response. Figure 22A is the SR function of this form (and Appendix IV, Figure 40). Saturation at the stronger concentration was occasionally seen (e.g., Appendix IV, Figure 40, unit 23.3). Four neurons gave smaller responses with progressively more concentrated stimuli, i.e., their SR functions had negative slope. An example is shown in Figure 23A (open triangles) (and Appendix IV, Figure 41). For one fiber, the decreasing responses to the quinine test series occurred after a previously large response to 0.3 M NaCl (64 impulses/20 second test). To ensure that the declining responses for quinine were not solely due to the decaying activity of this previous test, 20 seconds of pre-stimulus activity was subtracted from each of the responses. The remaining activity still formed a negative function but of somewhat less regular slope (Figure 23A, filled triangles). Positive and negative SR functions were also combined over a concentration series (for 10 of 39 neurons with positive SR functions), giving rise to a U-shaped function. The greatest responses were from the greatest concentrations. Examples are shown in Figure 24 (Aland g) and Appendix IV, Figure 42. Three examples of the reverse combination 91 .umHDwou mm uo: ma wows: ossouwxomo 030 you mnoauossm mosouma 00m Mm ecu mo show one .:m.»: oouuo>dw 0:0 as vowuflusoow 0H0: owuma onu pom hunched youuosm 00m mosuwswma uncommon Houmouw nuoo oomsmo H0: 00 msowumuusmosoo mcfimmouosfi mafia macaumaaafium o>wmmooo:o manna .moauossw kosmuma o>wumwos was Mm 0>Huwmoo m saws monsoomou mucuMuwso .HN magmas 92 _Illlusm . <<<<< Au as»... a .2. 2 So 31:... i .2... 89o « ea: a... .2. 2 59o « on: .s... Figure 22. 93 Regular forms of SR and latency functions. A, Q, E, and Q, are graphs of gustatory SR functions with positive slope (A) and zero sloped SR functions (9, E, and g). The regular latency functions defined for the same neurons were of negative slope (B), zero slope (2), negative slope (E) and positive slope (H). For this and Figures 23, 24, and 39-44 the following symbols are used: Key: -o-o— = responses to HCl. -A-A- = responses to quinine hydrochloride. -X-X- = responses to sucrose. -I-I_ = responses to NaCl. A = the whole nerve's response threshold for the same stimulus, Placed along the abscissa. ----- = latency values for no response (equivalent to a 20 sec latency). -———€> = the criterion level of pre-stimulus activity (mean inpulses/S sec + SE), placed along the ordinate. Responses are given as 20 second averages unless other- wise indicated in the legend. Latency is scaled in seconds except for graphs with thickened ordinate axes or those that are set apart in an inset of the graphs. These will be noted in the legend (e.g., 2, the latency function graph for HCl stimulation of unit 21.1 has ordinate expanded to display only one second. All concentrations are Log functions. 94 RESPONSE (impulses/ 20 sec) A B l-ICI 14.1 20 3:20 -------- \ 3 \ l 510 \q E 0-535ec .— < n a I n _l - a 1 4 a“; n _‘_J -5 -4 A -3 -2 -1 -5 -4 A -3 -2 -1 L06 coucmrnmon (Log in LOG coucmmn'flou (Log in c 0 0 “CI 21" 310 N 10 t .2 ‘a’: [M '7 I #1 n 1 n a 1 n J 1 1 n L—1 t l I n -5 -4 A -3 -2 -1 - -4 a - - .. LogN 5 Log N3 2 1 E F 0 1o . 3 oucu 27-3 1° N I § \ -- a a .1 2 1 Pl 1 l 13 l 2 1 -5 -4 -3 - - - - .. - - Log M 5 410g I 20 8 310 Sucrose 17-l 10 2' . i '- J J l L J 1 J L -5 -4 -3 -2 -1 ' -5 -4 - - + L08! log "3 2 FIGURE 22. 95 A.oom 0N .oofioomou o0 How moaouoa oeu musomouaou ooHH oonmoo onev .NN ouswqm 0H no oaoo one one maooshm .oooam o>HuHmoo mo ma ass: oaom osu now soauossw hosouma 059 am .mfimhaoso sonuwo cues woumwmuom omoao o>Huowoa one .Amoawcowuu soo0v Ho>uousw osooom m Hashes onu now no HHoa mo Amoawsoauu ooHHHmV omaoooo 0N uo>o hua>fiuoo msoosousoom wswuouoasso an mowuom msasawum wsfiuoooua onu aoum hua>auoo wsfiswaooo How oouoouuoo ma soauoasaaum Home ou moosomoou m.sou:os o no sowuosow mm o>Huowos one afl .uHss o mo sowuossw mucouoa o>HuHmoo moo Mm o>fiuowos onH .mN magmas .2 550.... 2 “.3. s 3.. 20.52.2320". I004 2035:2328 03m Fl N... ml ml Fl NI no vt ml 11 .I ii i 4.3.... Us: a . I ,9 (995) AON31V1 1 g 2 (cos oz/sasgndlug) ISNOJSNI Iii .. on to. 5.8 1 m < Figure 24. 97 U-shaped SR functions and bell—shaped latency functions. These are shown for two different units in response to HCl and NaCl stimulation. Notice that the SR minima and latency function maxima fall within 0.5 Log unit of the whole nerve threshold. Symbols are the same as in Figure 22. 8 3 jr—-—v 9? 8 8 RESPONSE (impulses I20 sec) 8 I _s C W i ‘5 NCI 98 174 N O I LATENCY (sec) '5 '5 -4a-3 -2 -1 [.06 CONCENTRATION (LogN) 8 8 impulszss / 20 sec 8 0 8 r v NaCl © T l l I _|__ _J—J -4A-3 -2 -1 L00 couccnrnnnou (Log u) 15-1': 20" ’“s I 089 I see 810- xi. \ In I I l l l l l a '5 -4 - -2 -1 Log M L L n TI _A_l '5 -4 -3 ‘ -2 ‘1 log I FIGURE 24. 99 (positive and negative SR functions) forming a bell-shaped function were found. For two units this occurred with sucrose stimulation, for a third with HCl stimulation (Appendix IV, Figure 42, unit 24.2, sucrose tests). Stimulus-response functions of zero slope were not unusual (12 of 67 functions with defined slopes). These are seen in Figure 22 (g, E, and Q) and Appendix IV, Figure 41. The evidence that these are true gustatory response functions is as follows: 1) their response magnitude is large (2X to 8X above criterion levels) in several cases; 2) these responses were chemo—specific, since other stimuli produced different response functions, and 3) several tests of other chemicals gave re— sponses that were much smaller or zero. This indicates that these were not small responses from mechanical or thermal stimulation which had been minimized by experimental controls. Neurons with HighfiLevel Spontaneous Activi_y. The stimulus“ response functions for the two neurons (units 29.1 and 29.2) with high level spontaneous activity were largely indeterminate. For one test series, the form was clearly recognizable, the responses of unit 29.1 to sucrose forming a U-shaped function (Figure 25) with the minimum representing the greatest depression of activity. In part, this occurred because the pre—stimulus activity level rose and fell through the series. However, the activity during depression was not constant, but decreased with each concentration over the first half of the series and increased over the last half. This is depicted in Figure 25A by the line graphs representing the pre-stimulus activity, 100 Figure 25. Gustatory responses of units with high level spontaneous activity. A- The responses of unit 29.1 to sucrose are with depressed activity over the entire series. A line graph is drawn for each concentration. The left limb of the line graph at each concentration is the pre—stimulus activity level, the center is the test activity, and the right limb is the rinse response. (All activity levels are averaged over 20 sec.) The rinse response is seen to be greater than pre- stimulus activity except for the rinse response following 0.3 M sucrose. The SR function for unit 29.1 ("net response" in the graph) is U-shaped. It arises in part because of changing pre-stimulus activity levels ("V" curve) but also with the U—shaped test activity levels (middle curve of inverted v's). The SR functions to NaCl for unit 29.2 (also with high level spontaneous activity) is complex. It includes both depressed activity (at the weakest concentrations) and increased activity responses at greater concentrations. Squares are the net NaCl stimulus-response function (test activity minus pre-stimulus activity). 101 Sucrose Tests & Rinses A F— 29 -1 7 p p — — n E I M0000 m w m m m m A§0N\a3.=as_v 33:3 «mop .352 NaCl 93. 0N\...8.....E_ A Test Actmty ‘ Net Response ...../ -1 / n—n” —2 /x X\ X\ n \K 3 l 4.! Ba m. .W m a“. V .. . .m\ V\ _ v /v .v [ll-.IIIPp 0 0 3 2 + + 0 -10 -20 -3O amen—2.5. Figure 25. 102 response, and post-stimulus activity fin: each test. In Figure 258, the SR curve ("net response") the pre-stimulus activity curve, and the curve showing test activity during stimulation for unit 29.1 are plotted to show their relationship (responses of SR curves = activity during stimulation minus pre—stimulus activity). The SR curves of unit 29.2 were more complex and less regular. The sucrose tests all caused activ— ity depression. The HCl tests formed an irregular function with weak depression for the two lowest concentrations, with activity increasing for the greater HCl concentrations. The NaCl SR function (Figure 259) was similar to that of HCl, with no response or very weak depression at the lower concentrations and increases only for the greater concentra- tions, tending to a zero slope over the last 1.5 Log range. Rinse Response SR Functions. As noted above, six neurons exhibited rinse responses after one or more of the four classes of stimuli. The rinse responses were averaged over only five seconds. This does not seriously alter their comparative magnitude, for their response pattern was very phasic, adapting completely within that time. The property was chemo-specific for two fibers and broadly distributed in others (e.g., one unit gave rise responses to NaCl, HCl, and quinine hydrochloride, but not to sucrose). Two types of rinse responses were observed. The first type occurred after activity depression by the test stimulus (see Figure 26). Units 29.1 and 29.2 had high level spontaneous activity and responded with clear activity depression to the sucrose stimulus. Unit 21.1 also shows activity depression but is less obvious since the pre-stimulus activity was low. For all three Figure 26. 103 Rinse response SR functions. These are shown for four units, two with high level, pre-stimulus activity (29.1 and 29.2) and two with the more common low level, pre-stimulus activity (19.1" and 21.1). In some cases, rinse responses followed activity depression by the test stimulus (the solid line, test response, is below the pre-stimulus activity dotted lines, for 21.1, 29.1 and 29.2). The second type of rinse response occurred following positive responses (increased activity) to the test stimulus. (The solid line for 19.1" to NaCl is above the dotted line.) In all cases the rinse response (dashed lines) were of greater magnitude than the pre- stimulus activity or test responses. Impulses/S. 20 sec _s C O 104 NaCl 19-1” KEY In...‘ .... Pre-Stimulus Activity / \ .n-‘\ ,’ ‘\ ---- Rinse Response \/ \ W Response (Test Activity 2:3..2'“ minus Pre-Stimulus Log 5; '1 Activity) Sucrose 29-1 Impulses/ 20 sec l 0 \\ 40 I‘ \ :0 e e u o" \\ 30- \ \ 3. 1. \ \ .. , \ e e . . O . . . . . ”a. 0.". \\‘ 20 b 3.9 1.. \\ J n I l 1....;..": . on”): W 1° ' o l I l 1 I E l I Impulses/ZO sec I a E l to 0 I h l (A) O -40L Figure 26. 105 units, the rinse response not only increased activity to pre-stimulus levels but resulted in a rebound far above initial rates. The rinse response to unit 19.1" represents the second type of response, in which the neuron gave a positive response to the stimulus and was followed by rinse responses greater than test and pre-stimulus activity levels. The SR functions for the rinse responses to sucrose were bell-shaped (Figure 26, units 21.1, 29.1, and 29.2). Gustatory Latency Functions The physiological latency of the taste responses occupied a wide range, from 4 msec to nearly the entire test period, 19.75 sec. Figure 27 shows the relationship between the magnitude of the gustatory response and its latency. Larger responses tended to occur with shorter latencies. However, this relationship is variable (6 responses with a magnitude greater than 10 impulses/20 sec had latencies longer than 5 sec) and at best describes only a limited correlation. The smallest responses had latencies that ranged over the full test period (10 msec to 19.75 sec). The latency functions were of varied form, which helped to clarify the wide variation between the latencies of small and large responses. As with the SR functions, several latencies functions had complex slopes (69 of 155 latency functions, Table II). These varied from those with a generally positive or negative trend to forms that were too complex to define by slope. The most frequently encountered form was the negative latency function (30 of 57 latency functions with definable slopes), in which 106 .moaosouoa mo owsou Hana onu uo>o usooo AmHoHusouoo coauoo NIHV monsoomou uoHHmBo onu .mowosouoa “oppose o>o£ ou oaou oomsoamou Homuoa oau oafina mesa .oosu IHawoa omsoomou was honouoa moosuoo sofiuoaou oouo>sw onu he oouwsfia one use ofismsoauofiou usoumaosoo onsHm om soaaow ouon .soouw onu ca oousomoua one mumou Ham How Aoom 0N\oomasaawv oosuacwma omsoemou was Aoomv monouoa osy .zosouoa msmuo> oosuaawma omdommom .mu ousmfim 107 FIGURE 27. IIIII IIIIITIII IIIT I I a s s m (008) Innate-I 20 25 Response (impulses / 20 sec) 15 IO 108 Table II. Distribution of SR and latency functions. Total SR -SR -SR -SR -SR -L 0L +L IL 0 0 12 4 16 OSR OSR OSR OSR :_L__ SE. :L__ ._I.£ 2 4 2 4 12 +SR +SR +SR +SR :L. 9.1:. fl... .2: 21 l 0 17 39 ISR ISR ISR ISR -L 0L +L IL 7 2 6 44 59 Total L 30 7 20 69 / 126 No Response = 29 155 Above are the numbers of SR and latency (L) functions observed during Ithe experiments on 29 gustatory neurons. Classification of the func— tions was determined for a complete test series (e.g., as +SR) or for partial series when two forms were combined over the whole series (e.g., -SR combined with +SR). The symbols :3 Q, —;, and I_refer to the slopes of the functions as determined graphically (positive, zero, nega— tive, and indeterminate). The two gustatory units with high level spontaneous activity are not included in this distribution and are con- sidered separately in the text. The dashed line separates those pairs of SR=latency functions that are graphically defined from those that have an indeterminate relation to concentration. 109 the latencies declined regularly from several seconds (for responses to the weakest stimulus) to several hundred milliseconds or less as concen— tration increased. Figure 22B_shows one such curve. Notice that graphical extension of latency to its most extreme value does not greatly alter the curve (20 seconds = no response). Other examples of negative latency are seen in Appendix IV, Figure IV—S. The opposite form, a positive latency function, was also observed. As seen in Figure 23B, the latency increased regularly from 50 msec for the most dilute solution, to several seconds, to no response with increasing concentration. Other examples are in Appendix IV, Figure 43. A combined slope function also occurred, in which positive and negative latencies combined to form a bell-shaped function, with the maximum represented by one or more responses of infinite latency (see Figure 24, §_and D_and Appendix IV, Figure 44). Unexpectedly, several examples were found in which latency re— mained unchanged with concentration, i.e., functions with zero slope. For these curves, individual latencies varied slightly about an average value with no discernible trend to positive or negative over the 3.5 Log concentration range. Figures 222_and Appendix IV, Figure IV-S are examples of the zero slope latency function. For many of these, the mean latency value was 500 msec. Exact statistical determination of the slope was not possible. The rinse responses had two regular types of latency function. Most had zero slope. The other type had a positive latency function for the rinse responses to sucrose. 110 Latency functions were not calculated for the suppressed activity responses or for the rinse responses of the units with high level spontaneous activity. In these cases, the exact onset of either response was ill-defined, especially for low concentrations where depression was mildest. The onset of depression was not an immediate drop to the average lower level of activity but rather a gradual decline, which showed in the photographic records as an increase in the duration of the interspike intervals. Without interval analysis, the onset of suppression could not be accurately determined. Combined SR-Latency Functions It was possible to clearly define both SR and latency functions for a test series in many gustatory neurons (42 of 126 functions in Table 11) (see Figures 22, 23, and 24). Figure 22 A and §_shows the most typical combination, a positive SR function with a negative latency function. For these, greatest stimulus concentration resulted in the largest gustatory response occurring with the shortest latency. Weaker concentrations resulted in smaller gustatory responses with latency of seconds, or caused no response (equivalent to a 20 sec latency). The exact opposite SR—latency functions were also observed (Figure 23). The largest responses had the shortest latencies at the weakest concen- trations. Notice that in both combinations, response magnitude and latency are inversely related. U-shaped SR functions were usually found with bell-shaped latency curves. This can be seen in Figure 24. Minimum responses had maximum latencies. 111 Three types of latency function accompanied SR curves of zero slope: 1) a zero slope latency function, in which a response of rela- tively invariable magnitude occurred at the same time after stimulus onset (e.g., 500 milliseconds) over the 3.5 Log concentration range, 2) zero slope SR functions also occurred with negative latency functions, and 3) positive latency functions occurred with zero sloped SR functions. These were the least common, occurring only twice each. Thus a family of taste responses in which each type is defined by combined latency and SR functions was developed for gustatory neurons. These are diagrammed in Figure 28. The three blocks at the top repre- sent three successively greater concentrations of a stimulus, separated by a continuous distilled water rinse. Each type of response is pre- sented in a separate row with vertical lines representing action poten- tials. Figure 28A shows the most typical type of response in which greater responses occurred with shorter latency as concentration increased ("+R, -L"). In B the response is the inverse of A, i.e., the greatest responses occurred with the shortest latencies at the weak- est concentrations ("-R, +L"). In g, D, and §_responses of unchanging magnitude occurred over the range of concentrations with response latencies that shortened ("OR, -L"), remained constant ("OR, 0L"), or lengthened with greater stimulus strength ("OR, +L"). In F and _Ci rinse responses occurred at the same time after water rinse onset ("Rinse, 0L") or at progressively longer latencies as stimulus concentration increased ("Rinse, +L"). Finally, for the units with relatively high level of spontaneous activity, responses of depressed activity occurred (E, I) in which the degree of depression increased ("+R") or decreased Figure 28. 112 Gustatory responses defined by SR and/or latency func— tion. Three tests of stimuli of successively greater concentration, separated by distilled water rinse, are represented by blocks 1:3, A family of nine different gustatory responses was defined by specific combinations of their latency and SR functions. These are illustrated by the rows of vertical lines which represent action potentials. Arg, Responses of increased activity on stimulus pre- sentation. :3, :3, and Qg_are the slopes of the SR functions and :L,{:L, and 9L_the slopes of their latency functions. These occurred in the combinations shown in the figure. (See Table II for their distribution among the tests.) F_and §_are rinse responses which occurred with constant latency (0;) or latency that increased with greater concentration (:9). H and I_are responses of depressed activity. These were observed mainly from the units with high level pre- stimulus activity. These also occurred in combination with a rinse response on application of the post-stimulus distilled water rinse (represented in the figure by the closer spacing of lines after the tests). DECR H A01" Y H'Rl ozcn _ I ACTYl m 113 % 7/ /A Ri W Rise Tt R' It R3 3 3 3 3 3 3 ”ll llllll JlllllLl 3 3 3 3 3 3 _lllllll llllll llll 3 3 3 3 3 3 lll I H l l l l I H I III Jllll » lll I I Jul] ELL 3 3 3 3 3 3 HI] H I “ll Ill l] l lll ll 1 I IN FIGURE 28. 114 ("-R") with greater stimulus strength. These two forms frequently occurred at opposite ends of the concentration range and each was followed by a rinse response. Individual gustatory neurons exhibited a variety of the responses described above. Typical examples are shown in Figure 29. Unit 17.1, for example, was a less versatile unit. It exhibited low spontaneous activity levels, no suppressed activity nor rinse responses. Its strongest response occurred to HCl, for which it formed a U-shaped SR function and bell—shaped latency function. It responded to sucrose with only one or two action potentials with a positive latency function. It was not responsive to NaCl and only weakly so to quinine. Unit 15.2 gave greatest responses to NaCl. For the HCl stimulation series it responded with a negative SR function. The responses to quinine and sucrose do not conform to a regular SR function but are above the spon- taneous activity levels. None of its latency functions is clearly determinate. _Unit 21.1 responded with relatively constant SR and latency functions to HCl at all concentrations. To quinine stimulation its magnitude function was less regular while its latency function stabilized to constancy in the last half series at greatest concentra— tions. To sucrose, its responses resembled complete activity suppres— sion though.its pre-stimulus activity was too low for such decreases to be obvious. Its NaCl responses were relatively weak and indetermin- ate for both latency and SR functions. Finally, unit 29.2, one of the two units with high level spontaneous activity responded positively and strongly, but irregularly, to the HCl series, after initial activity Figure 29. 115 Combined SR functions to the four primary stimuli for several neurons. Both regular and indeterminate SR functions occurred for four different neurons. (§_= HCl, §_= NaCl, Q.= quinine hydrochloride, §_= sucrose, and §£_= rinse responses after sucrose tests for unit 21.1) 17.1, 15.2, and 21.1 were units with low level pre-stimulus activity of 0.268, 1.594, and 0.210 impulses/S sec averaged for the four tests (not shown). The responses for unit 29.2, a neuron with high level spontaneous activity, were determined by subtracting the 20 second pre—stimulus discharge from the 20 second test response. The latency functions (not shown) in many cases served to further define the unit's responses (as described in the text). 116 i a 3 a "’ “é .——r-\\-§—-' 50'- 17-1 § 3 60 \ 403 5; GI! a 350. a \ 3 3° ' :8: £403 20 30 10 -5 —4 '3 -2 -1 Log Concentration § 20 21 1 .§.. 0 .. N \ 3 10 £ 3. s ”N,“ ‘ x r““‘" 1 - ' 7 - -4 ‘3 ‘2 '1 Log Concentration 4.2% 292 H §+1o - ,__-N -5 ’ lmplllSeS/zo o I I J “—-4’ $0 at... a V o'.~o 0.. ..~ 5 “0...“... .. O. . . 0. .° 0.... -10 15'2 :5 :4— —§ .2 Log concentration Figure 29. 117 depression for the two lowest concentrations. To sucrose, its spontane- ous activity was consistently depressed but to varying degrees. To NaCl it showed initially, no response or weak depression until the test concentrations above the whole nerve threshold for which its SR function stabilized to constancy. As mentioned previously, its latency functions were not obvious without interval analysis. Mechanical and Thermal Responses All target neurons (identified during the experiment as responsive to gustatory stimuli) exhibited tactile responses. Nonetheless, in all experiments, gustatory stimuli were tested even when tactile responses were absent. No neurons responded to gustatory and not mechanical stimulation. However, the possibility of non-mechanosensitive gustatory units must remain open, some chemosensitive neurons with smaller ampli— tude action potentials were discovered during computer data processing, for which tactile responsiveness had not been adequately tested during the experiment. Fourteen out of 20 gustatory neurons responded to lingual cooling. A majority of the sensitive fibers responded to cooling both from room temperature to 2-40 C and from hot water (50-600 C) to room temperature (21-230 C). Most units responded best to one of the two cooling modes. Of the thermally sensitive gustatory units tested with both, 6 to 11 responded more to 2-40 C after room temperature water rinse, 4 to 11 responded best to room temperature rinse after hot water, while one showed equal responsiveness to both cooling modes. Six units responded to the application of the hot water itself as well as to either or both 118 modes of cooling. The largest responses to cooling occurred after hot water rinse (38.7 impulses/20 seconds) followed by cold stimulation (23.1 impulses/20 seconds) and then hot responses (4.625 impulses/20 seconds). In most cases higher instantaneous frequencies were observed on the photographic records than were revealed by the 20 second average. Figure 30 (5:9) shows a unit whose thermal responsiveness is greater in the direction of cooling from room temperature. Figure 302_shows a second cold sensitive neuron which responded to the application of hot water. This fiber responded best to cooling to room temperature from hot water rinse. 119 .oaomo msasawum osu mo uomao onu nuH3 mvwocfioo m3ouu< .aousoa vcoomm o How naonm ma mm AU ommv wsasaflum umumz uoa onu cu wovaommou acousoa maom am a 5 cmsu zoaouma umchH m nuHB moduwawma nommoa mo omaoamou m momsmo nouns uos mo cowumowadmm m50H>oum osu umumm omswu ouaumuomaou aoou .u .omaau unnumuoaaou Boon noumn Au omqv Honda uo: cu oncommou on o>mw condos oSu «m CH .Ao oNN .uoums voHHfiumHv ousu Imuoaaou Boon Hound o oav Houm3 vaoo ou uncommon any mfi.«. .aoudoc ago How mumou m>Hmmooo=m sonny aoum noncommou Hmauonu on» www.mfifl .mcouaoc mucuMumsw mo noncommou HmaumSH .om ouswwm 120 00m .— u .8 mzaafi VI. DISCUSSION Mechanical and Thermal Sensitivity of the Gustatory Neurons The mechanical and thermal sensitivity of the mud puppy glosso- pharyngeal neurons paralleled that of the frog, both classes of stimulus being effective for most of the fibers. The thermal sensitivity was not identical among all fibers. The gustatory neurons may possess varying thermal response functions since many responded more greatly to one of the modes of cooling. The form of the thermal functions could not be determined using only the two broadly defined thermal stimuli. The Expanded Gustatory Response Considerations of Stimulus-Response (SR) Functions. The systema- tic investigation of the SR functions of the gustatory neurons of Necturus has revealed the type that have been observed in the past: 1) the classical SR function with a direct relation between stimulus concentration and response magnitude, 2) SR functions with an inverse relation between concentration and response, 3) SR functions with con- stant response magnitude with increasing stimulus concentration, and 4) SR functions with curves of both positive and negative slopes, combined as U-shapes and bell-shapes. The less common SR functions, such as those with zero or negatively sloped curves and U-shaped curves, 121 122 were more greatly represented among Necturus gustatory neurons than has been reported for other species. The U-shaped form may have been more frequently observed because of the testing of an extended range of stimulus concentrations, 1.5-2.0 Log M below the whole nerve threshold. The minimum responses of the U-shaped curves frequently occurred near the whole nerve threshold. The negatively sloping portion of the curve was often confined to the subthreshold range of concentrations. The complete definition of the SR function for any of the four stimuli was not attained for the neurons. This would have involved the testing of several greater dilutions of the stimulus, approaching pure distilled water. The minimum dilutions were at least two Log units above this concentration (e.g., -5 Log N was the least concentrated solution of HCl tested and -7 Log N is equivalent to neutral pH). Moreover, the U-shaped functions often had responses from the most dilute stimuli that were much greater than the criterion water-adapted activity levels. For unit 17.1 (Figure 24A) the response to -5 Log N HCl was 17 impulses/20 sec. If solutions between -7 and -5 Log N HCl were tested, it could be expected that they would have risen from the pre-stimulus activity level of l impulse/5 sec to attain that observed magnitude. The role of the SR functions was found to be incomplete in defin- ing the taste response when considered alone. Other aspects of the taste response were revealed in specific relation to the SR functions. Considerations of Latency Functions. Alone, the latency functions were as varied and complex as the SR functions. Many types were 123 observed across the sample population of neurons. Latency—concentration curves were found with zero slope and positive slope, as well as with the inverse relation between stimulus concentration and latency found by T. Sato (1976) and Ogawa 25.31, (1974). However, the latency func- tions were found in specific relation to the SR functions of the neurons. Examination of Table II for the units with defined latency and SR functions (within the dashed line) shows missing combinations of latency and SR functions. For example, the +SR function was never observed with a zero or positively sloped latency function. This might have reflected sampling limitations only. On the other hand it served to emphasize that the latency function and SR function of the neurons were joined in specific combinations. Similarly, the reverse combination of -SR, + latency function was found to be nearly the sole combination for neurons which responded to stimuli of greater concentration with smaller magnitude responses. The most revealing combination of latency and SR function was found for responses that were previously considered to represent the absence of a taste response, the zero sloped SR functions. Combined with.constant latency, the neuron fired with the same.magnitude after the same delay to all concentrations of the stimulus. This function potentially conveys information about stimulus onset. Moreover, this particular combination of functions in any neuron was seen for only one stimulus among those tested. This therefore represented informa- tion about the onset of a particular stimulus for that neuron. In a similar manner, the latency and SR functions were found to combine specifically, to form five disparate gustatory responses. 124 The zero sloped SR function was also found with positive latency func— tions for some units, and with negative latency functions for others. Both of these combinations could convey information about the concen- tration of the stimulus in a novel manner, by varying latency as concen- tration increases. This is functional redundancy in the sense that many neurons relay intensity information by response magnitude (the + or - SR function). However, it is novel in that the information can be relayed with constant activity and with activity of a lesser amount. Varied Response Forms. The various forms of the taste response, i.e., increased activity or depressed activity on stimulus presentation, or rinse responses, have also been demonstrated previously in other species as considered in the Literature Review. However, in the mud puppy taste neurons, the several response forms were clearly seen in the same neuron. One of the units with high level spontaneous activity epitomized this variety of response form. The unit's responses included increasing activity to HCl and NaCl as well as depressed activity to sucrose tests which were followed by rinse responses. These then, when common to a variety of neurons, and all arising from the same unit could be seen as distinct classes of responses. The depressed activity responses formed U-shaped SR functions with positive and negative slope. These were considered separately to define two additional types of gustatory response: decreased activity with a positive SR function, and decreased activity with a negative SR function. The rinse responses were seen to combine with either positive or negative latency functions and defined two additional types of gustatory response. 125 Taken together, SR functions, latency functions, and forms of the gustatory responses defined nine different types of response (Figure 28). The individual parameters joined in specific combinations and thus more completely defined the gustatory responses of the glossopharyngeal neurons of Necturus. Multiple Chemosensitivity_of the Taste Neurons Many units exhibited multiple sensitivity within any one type of the nine responses, e.g., unit 28.2 gave increased activity responses with positive SR and negative latency functions to both quinine hydro- chloride and HCl. On the other hand, few fibers were truly monogusta- tory, i.e., by responding with only one form of response and one type of SR and latency function to one chemical and being non-responsive to the other stimuli. For several neurons both multiple sensitivity and chemo-specificity were found. In this case the neurons displayed one of the nine response types to two or three of the stimuli but gave a distinctly different type of response to one other stimulus. The com- plete specificity to all four stimuli by distinct types of response to the four stimuli was not often observed (for only one of the 31 gusta- tory neurons).' The relation more commonly observed was for the neuron to respond with less complete response chemospecificity. The unique feature observed with these fibers though, is that any degree of specificity can be conveyed by the different forms of response of a neuron. In part the lack of more complete definition may have arisen with the complex or indeterminate forms of the latency and SR functions. 126 These forms are an enigma. They were found for 59 SR functions and 69 latency functions (of 126 SR functions and 126 latency functions). They might be simply considered to be complex functions occurring under the regular schedule of stimulation. However, they might also reflect an error in establishing the particular stimulation protocol, that too little inter-trial rinsing may have disallowed the system's return to baseline conditions. In this case it must follow that single neurons have quite different adaptation properties, for the schedule resulted in the complete definition of SR and latency functions for some neurons, and that the adaptation properties must vary among stimuli, since some neurons displayed a regular function to one stimulus and not to another. Considerations of the Origins of the Taste Response Considering the physical parameter of stimulus diffusion, only the positive SR function and negative latency function should be expected. The greatest stimulus concentrations should diffuse most rapidly through surface barriers such as mucus or saliva and arrive in greatest concen- tration at the receptor interface to most rapidly initiate the response. A very different explanation is required for the reverse combina- tion of functions, -SR, +latency, in which greater stimulus concentra- tions cause smaller, more delayed responses. This is even more apparent for the non-monotonic, U-shaped and bell-shaped latency func- tions. For these, the neuron responded equally and with nearly the same latency to extreme stimulus concentrations which were often more than two Log units apart. The classical model assumes that the action 127 of the stimulus molecule on the taste system is invariably excitatory, a consideration belied by the demonstration in several taste fibers of the depression of activity by chemical stimulation. It was also shown in these experiments that depression is dependent on stimulus concentra- tion. These neuronal functions can be explained by postulating a concentration-dependent inhibitory action of the stimulus molecules. The exact form of inhibition is not suggested by these experiments. It may be by the diverse action of stimulus molecules at disparate receptor sites of the receptor cells of the taste buds as suggested by Ogawa g£_§l, (1969). On the other hand it could involve more complex relations between many elements of the taste system, actions from other cells of the taste buds or from interactions between the neurons. Single Unit Activity and the Whole Nerve Response Given the varied forms of the SR and latency functions, it is not obvious that all of those sampled in the experiments could contribute to the responses of the whole nerve. For example, the SR functions of the whole glossopharyngeal nerve of Necturus are all regularly increas- ing functions for the four primary stimuli. To test this concept, the responses of the single neurons were summed for each test and then expressed as relative response SR functions as described in the Literature Review. All the individual responses sum to the SR functions shown in Figure 31. The curves for the four stimuli are comparably related to those from the whole nerve's responses (Figure 6). This parallels the findings of Ganchrow and Erickson (1970) who summed the Figure 31. 128 Summed SR functions of the four primary stimuli for 29 gustatory neurons of Necturus. These can be compared with the summated whole nerve SR functions (Figure 6). Responses to each stimulus concentration are expressed as a percent of the response to 0.1 M NaCl. Baseline activity was determined from the 5 sec pre-stimulus activity for all tests (increased 4X). The inter- quartile range of the water baseline is shown from the ordinate (arrows). The summed responses form regularly increasing SR functions despite their widely varying individual forms. The summed HCl (H) and sucrose (S) SR functions are in close parallel to the whole nerve functions. NaCl (N) and quinine hydrochloride (Q) differ in their slopes and thresholds from the whole nerve functions but are in the same relation to the HCl and sucrose functions. § Relative Response 3 o 100 129 Log Concentration (Log ll) Figure 31. Oh 130 individual responses of rat chorda tympani neurons forming the regularly increasing functions of the whole rat chorda. Thus, though an SR func— tion with positive slope is observed for the whole nerve, quite varied activity from the many neurons can be hidden within the population's responses. Similarly, various response latencies underlie the response of the whole nerve, which for Necturus is a consistently phasic response (see Figure 5). To compare the varied onset and activity patterns of the units with the phasic response of the whole nerve, the activity of each neuron in the sample (of 29 neurons) was first averaged over succeeding one second intervals for their twenty second taste response to 0.1 M NaCl. The bar graphs showing the individual response patterns and latencies of the neurons are shown in Figure 32. The activity of all neurons was then added for each second of the response. The net activity occurring in the sample population over succeeding one second intervals is shown in the graph to the right. This summed activity is seen to form a large initial phasic response paralleling the whole nerve response. The varied latency functions found in the experiment are not inconsistent with the population response of the whole nerve. These observations also reveal the severe limitations of whole nerve studies in defining the mechanisms underlying the taste system. Similar limitations can now be easily seen for studies of the taste system using only single concentrations of stimulus. The response magnitude of the single neuron at any one concentration may represent the maximum, minimum or any intermediate locus of the neuron's SR Figure 32. 131 Summed response patterns for 29 gustatory neurons of Necturus. The responses of the individual neurons to a twenty second test of 0.1 M NaCl are shown to the left (lines §_through g_and §:_through 2:). Despite their varied latencies and response patterns, the summed response pattern (to the right) parallels the whole nerve response pattern, both having a large initial phasic response. (Compare with Figure 5.) 132 403- 2 com >832... 83 32.5.... V WXV. Figure 32 . L Illblc ON Time (sec) 133 function for that particular chemical. Furthermore, without latency information, the response detected with one concentration may seem to be of insignificant magnitude and all the while its precise delay may have carried the information that sucrose is on the tongue. VII. SUMMARY The individual gustatory neurons of the mud puppy responded with a variety of stimulus-response (SR) and latency functions to the four primary stimuli, HCl, NaCl, quinine hydrochloride, and sucrose: 1. SR functions with curves of positive, negative, or zero slope. 2. SR functions with combined positive and negative slopes to form U-shaped or bell-shaped curves. 3. Latency functions with curves of positive, negative, and zero slopes. 4. Latency functions with.com— bined positive and negative slopes to form bell-shaped curves. 5. SR and latency functions with complex slopes. Several different response forms were seen: 1) increasing impulse discharge with stimulus pre- sentation, 2) decreasing activity on stimulation, and 3) rinse responses. The parameters of the form of the response, the SR function, and the latency function were not randomly associated but were seen in specific combinations that defined nine types of gustatory response: 1. Responses of increased activity with positive SR and negative latency functions. 2. Responses of increased activity with negative SR and positive latency functions. 3. Responses of increased activity with both SR and latency functions of constant slope, perhaps signalling stimulus onset. 4. and 5. Responses of increased activity with constant SR functions but with positive or negative latency functions, representing a unique form of response representation of stimulus concentration-by the latency parameter. 134 135 6. Rinse responses with constant latency. 7. Rinse responses with positive latency functions. 8. Depressed activity responses with positive SR functibns. 9. Depressed activity responses with negative SR functions. The neurons were seen to respond with more than one of the nine types of response and in some cases with specific types of response for different stimuli. Thus while most neurons exhibited multiple sensitiv— ity to the four primary taste stimuli, they also showed some degree of chemo-specificity. The varied forms of the SR and latency functions were seen to require an underlying mechanism involving inhibition by the stimulus molecules. Despite the variety of SR and latency functions, the responses of the sample neuronal population summed to a net regularly increasing SR function for each.of the stimuli, the form of the whole nerve SR function. Similarly, the various latencies and subsequent activity patterns added over time to give a net phasic response, the form of the summated taste response of the whole glossopharyngeal nerve of Necturus. This showed that a wide variety of SR functions and response latencies can underlie the whole nerve SR function and response and suggests that the func- tions and responses seen in the sample population of Necturus gustatory neurons may not be unrepresentative. VIII. RECOMMENDATIONS FOR FUTURE STUDY 1. It is not known whether the central nervous system of Necturus integrates the latency of the taste response. That must be decided by experiments that seek to define the responses and interrelationships of higher order neurons of the taste system. 2. The experiments performed in this study should be repeated to ascertain the persistence of the variety of nine types of taste response found for the mud puppy. Protocol should be altered to record pre- stimulus activity over longer periods so that a more accurate criterion of background spontaneous activity can be used to evaluate the responses. 3. The experiments might be best conducted on a different species since the responses and levels of spontaneous activity were far lower for Necturus than other animals. The largest response observed for any neuron was 344 impulses/20 sec or 17.2 Hz while to the same stimulus, 0.3 M NaCl, rat chorda tympani neurons gave responses of 70, 30, and 35 Hz (Figure 105, from Pfaffmann, 1955). 4. In its present form, the enumerator program, SAM-COUNT requires separate analysis runs if different lengths of time are desired for test counts (e.g., for the first 5 sec vs the full 20 sec test count). The program should be altered to analyze and store the activity occurring 136 137 in each successive second of the rinse and test periods. These should be stored on magnetic tape so that the operator can easily determine the activity in periods of any duration by simple recall and summation of the activity in each second. BIBLIOGRAPHY BIBLIOGRAPHY Bartoshuk, L. M., M. A. Earned, and L. H. Parks. 1971. Taste of water in the cat: effects on sucrose preference. Science. 171: 699- 701. Beidler, L. M. 1953. Properties of chemoreceptors of tongue of rat. J. Neurophysiol. 16: 595-607. Beidler, Lloyd M. 1969. Innervation of rat fungiform papilla. In Olfaction and Taste III. Ed. C. Pfaffmann. The Rockefeller University Press. Beidler, L. M., I. Y. Fishman, and C. W. Hardiman. 1955. Species differences in taste responses. Amer. J. Physiol. 181: 235-239. Bernard, R. A. 1972. Antidromic inhibition: a new model of taste receptor function. In Olfaction and Taste IV. Ed. D. Schneider. Wissenschaftliche Verlagsgesellschaft MBH, Stuttgart. Borg, (L, H. Diamant, L. Stram, and Y. Zotterman. 1968. Neural and psychophysical responses to gustatory stimuli. In The Skin Senses. Ed. D. R. Kenshalo. Thomas, Springfield, Illinois. Boudreau, James C., Brunce E. Bradley, Peggy R. Bierer, Stefan Kruger, and Chiyeko Tsuchitani. 1971. Single unit recordings from the geniculate ganglion of the facial nerve of the cat. Exp. Brain Reg, 13: 461-488. Bronk, D. W. 1935. The mechanism of sensory end organs. In Sensation: Its Mechanisms and Disturbances. Eds. Clarence A. Patten, Angus M. Franz, Clarence C. Here. The Williams and Wilkins Company, Baltimore. Cohen, Melvin J., Susumu Hagiwara, and Yngve Zotterman. 1955. The response spectrum of taste fibers in the cat: a single fiber analysis. Acta Physiol. Scand. 33: 316-332. Drfiner, L. 1901. Studien zur Anatomie der Zungenbein—, Kiemenbogen—, und Kehlkopfmuskeln der Urodelen. I Theil. Zool. Jahrb., 138 139 Erickson, Robert P. 1963. Sensory neural patterns and gustation. In Olfaction and Taste. Ed. Y. Zotterman. The Macmillan Company, New York. Erickson, Robert P. 1967. Neural coding of taste quality. In The Chemical Senses and Nutrition. Eds. Morley R. Kare and Owen Miller. The Johns Hopkins Press, Baltimore. Erickson, Robert P., Gernot S. Deutsch, and David A. Marshall. 1965. The gustatory neural response function. J. Gen. Physiol. 49: 247-263. Farbman, Albert 1. 1965. Fine structure of the taste bud. J. Ultrastructure Res. 12: 328-350. Farbman, Albert 1. and Jenny D. Yonkers. 1971. Fine structure of the taste bud in the mud puppy, Necturus maculosus. Am. J. Anat. 131: 353-370. Faull, John R. and Bruce P. Halpern. 1972. Taste stimuli: time course of peripheral nerve response and theoretical models. Science. 178: 73-75. Fishman, Irving Y. 1957. Single fiber gustatory impulses in rat and hamster. J. Cell. and Comp. Physiol. 49: 319-334. Francis, E. T. B. 1934. The Anatomy of the Salamander. The Clarendon Press, Oxford. Frank, Marion. 1972. Taste responses of single hamster chorda tympani nerve fibers. In Olfaction and Taste IV. Ed. D. Schneider. Wissenschaftliche Verlagsgesellschaft MBH, Stuttgart. Frank, Marion. 1973. An analysis of hamster afferent taste nerve response functions. J. Gen. Physiol. 61: 588-618. Frank, Marion. 1975. Response patterns of rat glossopharyngeal taste neurons. In Olfaction and Taste V. Eds. D. A. Denton and J. P. Coghlan. Academic Press, Inc., New York. Frank, M. and C. Pfaffmann. 1969. Taste nerve fibers: a random.dis- tribution of sensitivities to four tastes. Science. 164: 1183- 1185. Funakoshi, M., Y. Kasahara, T. Yamamoto, and Y. Kawamura. 1972. Taste coding and central perception. In Olfaction and Taste IV. Ed. D. Schneider. Wissenschaftliche Verlagsgesellschaft MBH, Stuttgart. 140 Ganchrow, Judith A. and Robert P. Erickson. 1970. Neural correlates of gustatory intensity and quality. J. Neurophysiol. 33: 768-783. Halpern, Bruce P. and Leonard A. Marowitz. 1973. Taste responses to lick duration stimuli. Brain Res. 57: 478-482. Halpern, Bruce P. and Daniel N. Tapper. 1971. Taste stimuli: quality coding time. Science. 171: 1256-1258. Iggo, A. and B. F. Leek. 1967. The afferent innervation of the tongue of the sheep. In Olfaction and Taste 11. Ed. T. Hayashi. Pergamon Press, Oxford. Kingsbury, B. F. 1895. On the brain of Necturus maculatus. J. Comp. Neurol. 5: 139-205. Kusano, Kiyoshi. 1960. Analysis of the single unit activity of gusta- tory receptors in the frog tongue. Jap. J. Physiol. 10: 620-633. Miller, Inglis J. Jr. 1971. Peripheral interactions among single papilla inputs to gustatory nerve fibers. J. Gen. Physiol. 57: 1-250 Miller, Inglis J. Jr. 1974. Branched chorda tympani neurons and interactions among taste receptors. J. Comp. Neurol. 158: 155-166. Mistretta, Charlotte M. 1972. A quantitative analysis of rat chorda tympani fiber discharge patterns. In Olfaction and Taste IV. Ed. D. Schneider. Wissenshaftliche Verlagsgesellschaft MBH, Stuttgart. Murray, Raymond G. and Assia Murray. 1967. Fine structure of taste buds of rabbit foliate papillae. J. Untrastructure Res. 19: 327-353. Ogawa, H., M. Sato, and S. Yamashita. 1968. Multiple sensitivity of chorda tympani fibers of the rat and hamster to gustatory and thermal stimuli. J. Physiol. 199: 223-240. Ogawa, H., M. Sato and S. Yamashita. 1969. Gustatory impulse discharge in response to saccharine in rats and hamsters. J. Physiol. 204: 311-329. Ogawa, Hisashi, Satoru Yamashita, and Masayasu Sato. 1974. Variation in gustatory nerve fiber discharge pattern with change in stimu- lus concentration and quality. J. Neurophysiol. 37: 443-457. 141 Pfaffmann, Carl. 1941. Gustatory afferent impulses. J. Cell and Comp. Physiol. 17: 243-258. Pfaffmann, Carl. 1955. Gustatory nerve impulses in rat, cat, and rabbit. J. Neurophysiol. 18: 429-440. Samanen, David W. 1973. The taste bud structure and whole nerve response of the mud puppy (Necturus maculosus). Master of Science Thesis. Michigan State University, East Lansing, Michigan. Samanen, David W. and Rudy A. Bernard. 1975. Scanning electron microscopy of the taste bud of the mud puppy, Necturus maculosus. In 33rd Annual Proceedipgg, Electron Microscopy Society of America. Ed. G. W. Bailey. Claitor's Publishing Division, Baton Rouge, Louisiana. Samanen, David W., Winifred J. Kryda, and Rudy A. Bernard. 1975. Electron microscopic observation of the taste system of the mud puppy (Necturus maculosus). In Abstracts: Society for Neuro- science, Michigan Chapter. Sato, Masayasu, Hisashi Ogawa, and Satoru Yamashita. 1975. Response properties of macaque monkey chorda tympani fibers. J. Gen. Physiol. 66: 781-810. Sato, Masayasu, Satoru Yamashita, and Hisashi Ogawa. 1969. Afferent specificity in taste. In Olfaction and Taste III. Ed. C. Pfaffmann. The Rockefeller University Press, New York. Sato, T. 1976. Latency of taste nerve signals in frog. Experentia. 32: 877-879. Smith, David V. 1975. Time course of the rat chorda tympani response to linearly rising current. In Olfaction and Taste IV. Ed. D. Schneider. Wissenschaftliche Verlagsgesellschaft MBH, Stuttgart. Sokal, R. R. and F. J. Rohlf. 1969. Biometry. W. H. Freeman and Company, San Francisco. Strong, 0. S. 1895. The cranial nerves of amphbia. J. Morph. Taglietti, V., C. Casella, and E. Ferrari. 1969. Interactions between taste receptors in the frog tongue. Pflfigers Arch. 312: 139-148. Tateda, H. 1961. Response of catfish barbels to taste stimuli. Nature. 192: 343-344. 142 wang, Michael B. 1973. Analysis of taste receptor properties derived Wang . West, from chorda tympani nerve firing patterns. Brain Res. 54: 314- 317. Michael B. and Rudy A. Bernard. 1969. Characterization and interaction of taste responses in chorda tympani fibers of the cat. Brain Res. 15: 567-570. Charles H. K. 1977. Intracellular characteristics and responses to gustatory stimulation of cells in the mudpuppy tongue. Doctoral Dissertation. Michigan State University, East Lansing, Michigan. Yinon, Uri and Robert P. Erickson. 1970. Adaptation and the neural code for taste. Brain Res. 23: 428-432. APPENDICES APPENDIX I OPERATION OF FREDSAM 143 OPERATION OF FREDSAM FREDSAM, an acronym for "frequency distribution of spike ampli- tude", is a computer program that identifies the number and distribution of action potentials according to their diphasic amplitude. The major indication of a neuron's isolation with extracellular recording is the uniformity of action potential amplitude. Computer analysis was required since most experimental records had several active neurons with superimposed electrical noise from recording and playback equipment to add variance to the spikes' amplitudes (Figure 33). Furthermore, the neurons of the mudpuppy's taste system usually responded with a low average frequency (< 1 Hz), too few spikes to contrast visually on the oscilloscope. FREDSAM measures all recorded potentials-action potentials, baseline fluctuations, and any artifacts (e.g., filtered EKGs, EMCs, static discharge) and displays these as a histogram showing the distributions of the diphasic amplitudes of the waveform. The diphasic amplitude was measured in contrast to the usual analyses which determine the size of only one phase of the action potentials since the cumulative effects of electrical noise altered both phases of diphasic action potentials but not necessarily equally nor coincidentally. Potentials were measured above and below a balanced null potential, established between the playback equipment and 144 145 Figure 33. Action potential alterations by background electrical noise. A, 14 action potentials with variable amplitude are displayed at 200 msec per division. B, Action potential #12 of Ar: viewed at 2 msec per division. The cumulative effects of the irregular baseline can be seen as: l) the initial rise of the negative (upward) phase starts from a point above the zero potential centerline (white arrow) and 2) the altered appearance of the positive (downward) phase of the action potential. C. Action potential #6 of A, at 2 msec per mm. Neither positive nor negative phase of the action potential's waveform are greatly altered. This action potential rises from the centerline in contrast to #12. 147 computer, by the following procedure. On the signal to start (an elec- tric pulse delivered by the operator or the timer pulse recorded on the second tape channel) the computer tracked the signal, rapidly sampling its height above (or below) baseline. Each sample of succes- sively greater amplitude was saved as the present maximum. When a new maximum was found, the previous maximal value was discarded. When the signal encountered zero baseline, the computer paused to detect whether the signal was passing to the region of opposite sign. If the signal did proceed across, the previous maximum was temporarily saved. In the same manner, the signal's greatest deflection of the opposite sign was also retained. When the computer detected a new movement across the baseline, the previous maximum positive and negative values were summed (with respect to their absolute value) and the total was recorded as the value of the diphasic amplitude of the signal (see Figure 34). The temporary maxima were discarded and sampling continued. Sampling proceeded at one of two different rates: 1) when tracking a new maximum and comparing it to a previous maximum the rate was slower (20 KHz) than when, 2) encountering no new maximum.and sampling was at 25 KHz. Therefore, for a diphasic action potential of 2 msec duration (both phases) 45 samples were evaluated to find its total amplitude. All signals, whether nerve impulses, the irregular background electrical noise, or artifacts were evaluated in this same manner, i.e., were assigned a diphasic amplitude value regardless of their waveform. Data collection was terminated when a second electrical pulse was delivered by the operator or the data tape's command channel. Figure 34. 148 Diphasic amplitude measurement and display. Above. Center. £91.01.- The computer follows the signal trace (left to right) rapidly sampling its amplitude above and below zero potential centerline (dashed line is zero amplitude). Maxima are temporarily saved (9) or recorded for computation (:§, :§) if no greater are encountered before the signal passes the centerline. The amplitudes of succeeding pairs of positive and negative maxima are added together to determine the absolute amplitude (YE) of the diphasic waveform. The FREDSAM display of the diphasic amplitudes of the above waveforms. Baseline potentials (Yl‘Y4’ Y5) are nearest the origin. The signals of equal amplitude (11 and 14) are entered in the same bin (12). Notice that all signals, regardless of waveform are evaluated by their greatest positive and negative maximum. 149 «+40 “s ”'30 4+ZO o+x6 ‘ 0+X1 3*: 413+“ o 0 0 +10 _. - , _ __._ __.. ._._.. .. ._ o _ o '5 o o .. X1 2x2 'X4 _xe .3 10 a -20 " -' 0 Computations: l+x..| + l-an = Y" X1 *8 " 4 12 X2 +5 “'9 15 X3 ' +1 "‘ 2 3 X4 +6 “’6 12 X5 +32 --29 61 X6 +10 “'10 20 FREDSAM Display .5 Y1Y4 3% fE Y3 'V2 '6 Y5 -= o E .5 2 1o 20 so 40 so so Relative Amplitude FIGURE 34. Relative Amplitude 150 Data values were organized by frequency of occurrence (the number of events) within each amplitude category and were displayed as a popula- tion histogram (see Figure 14, "Methods" section). Each vertical line (or bin) of the histogram represents the number of events of a particu- lar amplitude. The width of each amplitude bin is set by the operator at l to 999 relative amplitude units. The value of the amplitude unit is arbitrary and depends on the amplification of the input signal. To calibrate this relative value in uV, a signal (sine-wave) of known amplitude (20 uV) was processed and displayed with a histogram using bins of the smallest possible amplitude unit (1 relative unit per bin) (see Figure 16, Methods section). The peak in the amplitude histogram of the calibration signal then identified the calibration factor, e.g., 20 RV = 63 computer amplitude units. The maximum resolution of FREDSAM depended on the playback amplification but averaged 100 divisions for a 50 uV diphasic potential or 0.5 uV. When a record had more than one active neuron with distinct or well-isolated action potentials, the histogram contained discrete peaks corresponding to each neuron to the right of the initial peak, which represented the baseline potentials. The histogram of poorly isolated neurons did not have well-defined peaks and appeared instead as a large initial baseline peak with a positively skewed tail. Neurons of low activity had histograms with an irregular array of amplitudes (Figure 35). Both cases were rejected from further analysis since no active neuron could be clearly defined. Therefore, FREDSAM was employed to judge an experiment's acceptability when completed or to direct further dissection during the experiment. 151 Figure 35. FREDSAM displays of units with poor isolation, low activity. Uppe . To the right of the baseline population, no distinct maximum is resolved among this large number of potentials (1214) indicating that the record contains poorly isolated neurons. Lower. For this record, only 88 action potentials (above the baseline population) occurred in 26 of 31 responses analyzed by FREDSAM. No maximum was resolvable and the experiment was eliminated from further analysis. 152 Number of Events Relative Amplitude Number of Events in... ..,..|.Jl.|.|.|...l.ll homo ; Relative Amplitude FIGURE 35. 153 Our current version of FREDSAM offers several options to the operator: 1. The amplitude option, already mentioned, was used to check the persistence of maxima and minima defining the various populations by displaying their histograms with bins of different amplitude (1, 2, and then 3 relative amplitude units). 2. The total range of amplitudes displayed could be limited (from the minimal display of a single amplitude to the maximum of all amplitudes showing). This option was used continually to eliminate the omnipresent large baseline population from the display. A second useful application of this option was to display only the range of amplitudes defining a single neuron. The corresponding total number of events could then be read from the legend (see Figure 14, 2nd numerical value). 3. The ordinate could be scaled to various maximal number of events, from 25 to 6400. 4. Any single histogram could be recorded on magnetic tape. 5. Similarly, the data of any FREDSAM analysis could be permanently recorded and any histogram then reconstructed on recall. This option was used with all experiments, even those rejected for poor isolation or low responsiveness. APPENDIX II OPERATION AND EVALUATION OF SAM—COUNT 154 OPERATION AND EVALUATION OF SAM-COUNT SAM-COUNT is a computer program that counts the action potentials occurring in test and rinse periods, accepting only those impulses within the amplitude window established by FREDSAM as the limits defin- ing a single nerve cell. FREDSAM and SAM-COUNT computed the diphasic amplitude of the recorded signals according to the same algorithm (Appendix I, Figure 34). SAM-COUNT enumerated impulses in three different ways depending on the period to be analyzed. 1. For full test counts, SAM-Count started track- ing and counting at the recorded onset of the timer pulse and continued until the arrival of the second timer pulse. 2. For post-stimulus rinses, tracking and counting started similarly but continued beyond the second timer pulse until a pre-set interval had elapsed. 3. Pre-test counts proceeded in a more complex manner. (The desired count was from a period of pre-set length occurring before a stimulus onset signal.) The method of evaluation and rejection used in this procedure required more analysis time than the direct count procedures (1 and 2 above). To evaluate SAM-COUNT, two procedures were used. The first analyzed the program's reproducibility. Repeated counts were taken of the same test and rinses using the three modes of operation (full count, pre-test count, and post-test count). The standard errors of the mean counts were calculated (Table III). 155 156 Table III. SAM-COUNT reproducibility. Mean A.P. Number Mode Count Range Runs S.E. Full period count (20 sec) 30.821 30-31 28 0.075 Partial count, pre-test (20 sec) 47.773 47-49 21 0.115 Partial count, post-test (20 sec) 48.107 48-49 28 0.061 The variability can be ascribed to: 1) the computer's sampling rate (22-25 KHz) in determining diphasic amplitude, or 2) to variability in the triggering of enumeration by the command pulse (which would in turn effect the reading of each amplitude value), or 3) to noise in the play- back equipment. The larger standard error of the pre-test mode was expected given its more complex operation. The second procedure was to compare action potential counts determined by photographic analysis and by SAM-COUNT. Twenty-one tests were chosen randomly and the counts were taken for 5 sec pre-test, 20 sec test, and 5 sec post-test (using a 5X hand magnifier with scale in 10 mil divisions from the photograph of the responses). These were compared to the SAM-COUNT evaluation for the same periods. (In Table IV, 15 of the 63 counts can be seen to differ.) The product-moment correlation coefficient was calculated for each mode (Sokal and Rohlf, 1969). Correlation between the photograph and computer analysis is high (r = 0.969 for full test counts, 0.931 for pre—test counts, 0.989 for post-test counts). Furthermore, the product-moment correlation coefficients were all found to be highly significant (p<10.0005) using 157 Table IV. Enumeration product-moment correlation. Number A.P's Number A.P's Computer Photggraph Neural Unit Test Pre- Full Post- Pre- Full Post- 14-1 0.0003 M NaCl 0 6 2 0 6 2 14-1 0.0003 M sucrose 0 2 3 O 2 3 15-1 0.0003 M quinine 0 4 l 0 Z_ 1 15-1 0.0003 M sucrose 0 0 0 O 0 0 15-2 0.001 M NaCl 6 l3 l 19_ 11_ 1 15-2 0.03 N HCl 0 3 0 3. 2. 0 15-2 0.03 M NaCl 1 5 l 1 .19 1 15-2 0.1 M quinine l 8 0 g_ .19 0 19-1' 0.03 N HCl 1 8 6 l 2_ 6 19-2 0.0003 N HCl 0 0 1 0 1_ .0 23-1 0.3 M NaCl 0 l4 0 O 14 0 23-2 0.0003 M quinine O 3 0 0 3 0 23-2 0.03 M NaCl 1 3 0 l 3 0 23-3 0.1 M NaCl 1 24 1 l 24 1 23-3 0.0001 M NaCl 0 6 O 0 6 0 24-3 0.00003 M quinine l 2 0 l 2 0 27-1 0.003 N HCl 1 2 1 1 2 1 27-3 0.01 M sucrose 0 0 0 l_ 1. 0 28-1 0.003 M quinine O 2 O 0 .3 0 28-2 0.003 N HCl 0 l 0 0 Q_ 0 30-1 0.03 M sucrose 1 3 2 l 3 2 (Discrepant photographic counts are underlined.) Pre- Full Post- r 0.931 0.969 0.989 ttest 11.118 17.096 29.145 tcritical (p=0.000S,\l=19) 3.883 3.883 3.883 158 the t-test for the significance of correlation coefficients (Sokal and Rohlf, 1969). All experimental enumerations determined by SAM-COUNT were re- viewed and altered l) to eliminate artifact potentials identified in the photographs, 2) to adjust test count action potentials occurring before stimulus delivery, and 3) to adjust post-test counts occurring before rinse delivery. Other discrepancies in the two enumerations usually were with greater counts resolved by photographic analysis. This is ascribed to the less precise determination of amplitude allowed with the hand lens and scale and the computer counts were usually accepted. APPENDIX III EVALUATION OF DELIVERY TIME MEASUREMENTS 159 EVALUATION OF DELIVERY TIME MEASUREMENTS The delivery time was measured using a fluid switch circuit described in the "Methods" section. The flow of 0.1 M NaCl served to close the circuit between leads placed in the stimulus reservoir of the stimulus delivery system and on the tongue (Figure 18, "Methods"). The oscilloscope trace, triggered on the initiation of stimulus delivery registered the arrival of solution at the tongue by a steady deflection from the change of potential when the circuit was completed. To evaluate this method of detecting delivery time, we compared the fluid switch to an independent measure using the flow of colored saline detectable by a phototransistor. A fiber-optic light source illuminated a small region of a flexible plastic tube which was con- nected to the output of the stimulus delivery system (see Figure 36 and Figure 17 of the "Methods" section). Distilled water from the rinse reservoir preceded the flow of green saline (25 ml green food color/ liter 0.1 M NaCl, U. S. Certified Food Color, 2.5% in propylene glycol and water, The Kroger Company, Cincinnati, Ohio). A silicone photo- transistor detected the passage of green saline through the illuminated tube. The non-insulated tip of the "lingual" lead of the fluid circuit was inserted through the wall of the tubing into the illuminated region and also detected the saline flow. The phototransistor (Archer, No. 276-130, Radio Shack, Fort Worth, Texas) was placed above the ocular of 160 Figure 36. 161 Dual determination of delivery time. The flow of green saline following the flow of distilled water through a flexible, transparent tube is detected by two methods: 1) the standard measurement using the saline to close a circuit to the lead, L2, and 2) detection of the green color by a phototransistor placed above the ocular of a microscope which views the illuminated region about L2. 162 / Phototransistor \ ’/ Microscope l m/ with Solutions “a 3‘ \M 3s 3.. \ \ \ \\\\‘ 0‘ //WW lll/ILA 1 ml/J Fiber-optic Light Source FIGURE 36. 163 a dissecting microscope which viewed the tip of the fluid switch Circuit's lead in the illuminated region of the tubing at 40X magnifi- cation. The flexible tubing was flattened between two glass slides to decrease the dispersal of light from the fiber-optic light source. As can be seen in Figure 37 (bottom traces, above and below) both systems detected the arrival of green saline at the illuminated region. Detection times for the phototransistor (800 msec) and fluid switch (770 msec) were nearly equal. The controls, the flow of distilled water and stopping the water flow affected neither system. Repeated measures of the delivery time using the fluid switch alone revealed its variability (mean delivery time of 650 msec : 2.622 for 12 measurements). This small standard error occurred when measured without an animal present. lfl.§$£2: only two measurements of delivery time were taken with each experiment, insufficient replication to calculate any variance statistics. The ip_§i£p_measurements, showed a range of differences between the two determinations (0-26 msec). This indicates that the variance in the experimental measurements may be less than 26 msec. The effects of concentration of the saline used in the circuit to detect delivery time are shown in Figure 38. The slope of the deve10p- ing potential varied with concentration but the onset remained the same. Because of this lack of effect on concentration on the delivery time determinations, only 0.1 M NaCl was used for the experimental measure- ments . Figure 37. 164 Comparison of phototransistor and fluid switch determin- ations of delivery time. Traces 1 (above and below). lst Control. The initial flow of distilled water is terminated (no green saline in stimulus reservoir) affecting neither determination. Traces 2 (above and below). 2nd Control. The flow of distilled water followed by distilled water from the stimulus reservoir cannot be detected by phototransistor nor is the fluid switch circuit completed. Traces 3 (above and below). When green saline (food color in 0.1 M NaCl) flows from the stimulus reservoir, the phototransistor detects the change in illumination (downsweep, above) and the fluid switch circuit's com- pletion is registered by the observed change in potential (upsweep, below). 165 PHOTOTRANSI STOR Water — Air Water — Water Water — Green Saline ¢_.a 200 msec FLUID SWITCH CIRCUIT Water— Air Water— Water Water — Green Saline FIGURE 37. Figure 38. 166 Effects of saline concentration on delivery time measurement. The delivery time determination is made with different concentrations of NaCl and distilled water. The slope of the developing potential is altered but the delivery time measurement is unaffected. Only 0.1 M NaCl was used in the experiments. Water 167 .____. 200 mSec FIGURE 38. 0'3 M NaCl 0'1 M 0'03 M 0'01 M 0003 M Water APPENDIX IV SR AND LATENCY FUNCTIONS 168 Figure 39. 169 Complex SR and latency functions. Several of the SR and latency functions were of such complex form that graphical analysis resolved no regular slope for part or all of the concentration series. These are shown for the SR functions (above) and the latency functions (below). The symbols of Figure 22 apply (1 8 impulses). Notice the expanded ordinate for the quinine latency of unit 24.4 (to 0.5 seconds). RESPONSE 170 g NaCl 232 “20 QIICI 23-1 $10 3 - 8 3. \10 .E .. " -5 LOG CONCENTRATION (LogM) - ' ' ‘ ' l l l J -5 -4 -3 -1 Loan __ 1 l a l l l J -- , l l I 1 l '5 ‘4 '3 Lt2 '1 '5 -4 -3 -2 -1 Log It log It §2o- A l\ 20 NaCl 194” " , , NCl 27-1 5 \ \ 210 \ 31o “‘ \ .- < —l _L J l j - Ll R J -1 -2 -1 -5 -4 -3 5-2 -1 OG CONCENTRATION (Log N) to: N 20r- ‘ ONCI 15-2 §1°" QIICI 244 L l I L I .I _J ‘5- Lngflz -1 girlie..|.. -5 -4 3 -2 -1 logl 2° Sucrose 15-2 A I\I\ S10 A7152: 5 -4 Log I FIGURE 39. 171 Figure 40. Positive SR functions. These are shown for seven neurons (with two functions for unit 13.2). The same symbols as in Figure 22 apply here (i = impulses). Notice the expanded ordinate for the NaCl test series of unit 13.2. 172 350 r 40 .. 340 P 830 . O 3 g; 8 NCI 13-2 8 NaCl 13-2 :20 g 50 - s 10 E z: 40r- Iu “J J 3 c 30 r- L tn 2: 20 - 1o - ‘ - ‘9 log N .L—I—I -5 -4 -3 A -2 -1 LOG coucmmmou (Log u) §10 30 ~ 6 NaCl 27-1 N '\ 3 IICI 14°1 '- W a log M3 2 1 \ i/ZO sec ° 8 IICI 134 10 - §1O g Sucrose 24-1 / N \ o- '_ I j l ‘ I -5 -4 - .. - Log "3 2 1 I I l. p l l. -5 -4 A-S '2 -1 LogN FIGURE 40. Figure 41. 173 SR functions with zero and negative slope. These shown above and below in the figure. Symbols are those of Figure 22 (i = impulses). The NaCl test series for unit 1.2 was for only ten second tests shown on the ordinate. The pre-stimulus activity the quinine series of unit 15.1 was averaged for twenty seconds (Q at the origin). are as for 174 go “a“ "2 §10 NaCl 30-1 3 N a .3 l 1 l l I l a A l h l l a l J .- ‘5 '4 '31-2 -1 -5 -4 -3 A-2 -1 2: Log I log IA 3 Sucrose 17 1 gg‘o w §1° Sucrose 28°1 a. WW :3 ' 3) n L l I 1 J 1 1 L § 1 n . . . g -5 -4 -3 -2 -1 -5 -4 -3 -2 -1 3 Log M Log I o. m IICI 21‘1 m IICI - g 10 §10 Q 27 3 § 0 S S «W h 151 14:14 It I l -5 --.4,$-»A_Il-‘.21..‘1l - Log-3 -2 '1 Log I LOG CONCENTRATION 20- ; IICI 152 c N h 320 " Sucrose 191” O 8 :10- FIGURE 41. 175 Figure 42. U-shaped and bell-shaped SR functions. These are shown for six units. Symbols are those of Figure 22 (i = impulses). §_at the origin of the graph for unit 29.1 refers to its 20 second average pre-stimulus activity determinations. 603- § 50 '- 3 NaCl 15-1 \ ii 40 L E z 303 m z 2 2o - m Ill 8 10 '7 I J I L ___'_| -5 -4 -3.-2 -1' LOG coucmmmou (Leg u) 70- T 603- IICI 17-1 50!- ii 2.40- \ 8 83°” .§ 20- “'5 4 I I I I I -4 ‘-3 '2 '1 LogN 176 20 NaCl 24'4 .5 O i/ZOsec ,1 J I I I i I I_I I -5 -4 -3 - -1 Log" 2 10 IICI 19'1’ 3 8 h L n 1 I I -5 -4 .-3 -2 -1 LogN L IA .. .. °‘-3 -2 -1 c r 1 r l 1 r l IT -103. p 3 8-20- \ 8 Enor- e " Sucrose 29-1 -40.. Sucrose24-2 8 8 >- V -5 -4 -3 -2 -1 Log" FIGURE 42. 177 .~N ouowwm mo muonu one maonskm .uomsH new on nonsense soon was magnum Hum m.m~ mass How ouoswuuo one .muHa: mews now osoam one omega .omoam o>HuHmom use .ouou .o>«umwos sues mcoauossw Assumed .me masses z u...— 2 we.— T a. «.4 T mu 7 N- and .7 m- u u q - u u n - u d n - f #90” a fl 3 l — 0 0.? Né a: / Or I “IN“ _o—. M? 552..— s. u... z u... w! NI MI Qt m... _.I NI ml‘ Vt m: . . . .- IJIJI 1|4| lid . .la . . 1&4 was 0 ’ raw .2... 9. x .9m 8 r U S we.— ritrlulll. om sue. NI MI Q... ml . . . . . . . EmmezoEEzmozoo 2.....- m- 4 .3. 3x .. as. , to“ 323m Emu: . T 2»... - 2. T N- mic..- m- Tole!" Ténnowmwll....... m... I... .2... elele\elelo\e mom .7. new 5.. 3.x om . 0.. $2 (335) AONJLVT O N 179 .nomo seas mucosa mm N.n~ use N.mu mugs: now woumnfiuuo couscous one moauoz .NN ouowwm mo mHooahm pudendum one dues wows: o>Hm How macaw one mouse .oosaeaoo moaon o>Huowms use o>HuHm0d mafia msoauossm zonouma .ee museum 180 N.hN NéN .1? 552... z a... d l m 1|.loIOIOIJ 1 anew/A “I! N. el 1 1 a J a q d E a \0 *m6” /-I\ I ON 2.. on» .0: p r ’ I or ’ ‘l\r\ m méw 32....” . (\ LON E»...— T. «3‘ n: v... n... q u d a W T 1 \ . x .29 8oz /\ P ON A: no... 22.—12.530260 Go.— ..I «I at C v: on ]I I1 . a . . . 1| §$o\ \. u ocean... \ nowfl s m It... Loam I... x. l4 ”'ililllillli 3 1293 lllrlllllllljljlllllll“