Brain Research, 94 (1975) 447-463 © Elsevier Scientific Publishing Company, Amsterdam - Printed in The Netherlands

447

C O R T I C A L RESPONSES TO E L E C T R I C A L A N D G U S T A T O R Y STIMULI IN T H E RABBIT

T A K A S H I Y A M A M O T O AND YOJIRO K A W A M U R A

Department of Oral Physiology, Dental School, Osaka University, 32 Joancho, Kitaku, Osaka 530 (Japan) (Accepted March 17th, 1975)

SUMMARY

The distribution of surface positive cortical potentials evoked by electrical stimulation of the chorda tympani, glossopharyngeal and lingual nerves which innervate the tongue was mapped in rabbits. All projections were bilateral. Judging from the extent of the cortical response area and the amplitude and latency of the responses, the major projection of the chorda tympani was ipsilateral, whereas that of the lingual and the glossopharyngeal nerves was contralateral. Both the chorda tympani and the glossopharyngeal nerve project to a confined area in the insular cortex and the lingual nerve projects to the appropriate part of the somatotopic pattern of somatic sensory area I. Further, a single unit study was undertaken to characterize the response of units in the cerebral cortex which was induced by gustatory stimulation of the anterior tongue. Twenty-four gustatory units were found in the insular cortex and the claustrum. The gustatory units were divided into an early response type (21 units) and a late response type (3 units) based on latency measurements. Gustatory units were also classified according to discharge patterns into excitation type (21 units) and inhibition type (4 units). Eleven units responded to 1 or 2 kinds of conventional taste stimuli, and 13 units responded to more than 3 different taste stimuli. Sensitivities of cortical units to the 4 conventional taste stimuli were found to be mutually independent and randomly distributed among cortical units. The frequency of discharges increased in the excitation type units and decreased in the inhibition type units monotonically with an increase of NaCI concentration except at the highest concentrations.

INTRODUCTION

Recording of cortical gustatory responses has been largely unsuccessful for a long time. Only a limited number of studies have been reported on cortical taste cell

448 activities 11,15,2°,zl and on cortical evoked slow potentialslr,ls,19, 3~ induced by chemical stimuli applied to the tongue. Ruderman et al. 27 mention that, in the cat, the gustatory (parvocellular) region of the nucleus ventralis posteromedialis of the thalamus projects to a restricted region of cortex hidden within the presylvian sulcus and not within the coronal gyrus as previous studies had indicated. The cortical gustatory area explored by earlier investigators, therefore, was not the appropriate one. They also suggested that because of the sparseness of thalamocortical projections in the gustatory system, it might be difficult to record evoked activity in the cortical taste area. These findings might be the possible reasons for difficulty in recording cortical taste cell responses in the cat. Another possible reason is the high susceptibility of the cortical taste cells to systemic anesthesia; they may be easily depressed by anesthesia as is the case in thalamic taste neurons z°. Gerebtzoff is recorded a cortical evoked response to taste stimuli from the surface of the cerebral cortex in the anesthetized rabbit.This finding gave us the idea that the activity of cortical neurons responding to natural chemical stimuli could be recorded in the rabbit without much difficulty. In the present experiment the rabbit was used, and we wished to study the response characteristics of cortical cells activated by taste stimuli delivered to the tongue by the method of single unit analysis and to obtain some basic ideas of the decoding process of taste information at the cortical level. MATERIALS AND METHODS

Subjects

Thirty adult male and female rabbits (1.8-2.3 kg) were used: 9 in the cerebral evoked potential recording experiment and 21 in the cortical taste unit recording experiment. Preparation

Subjects were anesthetized by intravenous injection of pentobarbital sodium (40 mg/kg) or, in some animals, by intravenous injection of alpha-chloralose (70 mg/kg) and urethane (500 mg/kg). The trachea was cannulated. The hypoglossal nerves were cut bilaterally to avoid the tongue movements. The surgical operation and the animal fixation were performed under general anesthesia. During the recording of electrical activities from the cerebral cortex, each animal was paralyzed with gallamine triethiodide (Flaxedil) (50 mg/kg, i.v.) and artificially respirated without additional anesthetic, but locally treated with Xylocaine.HC1 periodically at wound margins and pressure points. Pneumothorax was used to reduce brain pulsation when necessary. The head of the animal was fixed in the stereotaxic apparatus by ear bars to bilateral burr hollows on the temporal bones after the temporal muscles were partially removed, so as not to damage the chorda tympani nerve by pushing the ear bar into the auditory meatus. One eye was usually enucleated to permit free access to cortex adjacent to the rhinal sulcus. After the cortical face area was exposed, the underlying dura was resected. Blood vessels of the cortical surface were not damaged, and left

449 intact to be used as landmarks for selection of the site of the recording electrode. The cortical surface was photographed. The exposed cortical surface was then covered with warm liquid paraffin. Drainage of cerebrospinal fluid through an opening in the dura over the foramen magnum reduced the brain pulsation and prevented swelling of the cerebral hemisphere after the overlying bone and dura were removed. During the course of experiments, the EEG was monitored in order to observe the body condition of the rabbit, and the rectal temperature was kept at about 37 °C.

Evoked potential recording The 3 nerves of one side were electrically stimulated; the chorda tympani, the lingual-tonsilar branch of the glossopharyngeal (IXth) and the lingual branch of the trigeminal nerve. The lingual-tonsilar branch of the IXth nerve was always sectioned at its exit at the base of the tongue and stimulated with bipolar platinum wire electrodes. Since the chorda tympani, after leaving the tongue, travels in the lingual nerve before branching to eventually join the facial root, the chorda tympani was stimulated at one of two locations. The chorda-lingual trunk, which was sectioned at the base of the tongue, was stimulated with bipolar electrodes after sectioning the lingual nerve central to the branching of the chorda tympani. Alternatively, the pure chorda tympani was stimulated with bipolar electrodes after being sectioned central to its exit from the chorda-lingual trunk. To stimulate the pure lingual branch of the trigeminal nerve, the whole chorda-lingual trunk was stimulated with bipolar electrodes after sectioning the chorda tympani central to the branching of the lingual nerve. The connection of the chorda-lingual trunk to the tongue was severed. A 0.01-0.1 msec duration, 5-25 V square pulse was delivered to each nerve from an electronic stimulator. The nerves were stimulated supramaximally with about twice the voltage necessary for a maximal cortical response. A single pulse was presented at 1 sec intervals. After bilateral exposure of the cortical surface, evoked responses were recorded with a silver ball electrode (0.1 m m tip diameter). Glass micropipette electrodes (2-4 /~m tip diameter) filled with 1 M NaCI were used for depth recording. The silver plate indifferent electrode was attached to adjacent tissues. The cortex was mapped in 0.3 m m steps. After conventional amplification and display the responses of 15 superimposed traces were photographed. Presence or absence of a response at any site was determined by visual inspection. During the experiment the location of each recording site was marked on an enlarged photograph of the brain. Response latency was measured from the onset of electrical stimulation to the onset of the initial surface positive wave. Response latency values were determined by measuring the latency for at least 10 responses at the recording site giving the largest responses in any nerve map and averaging them. These mean values were then averaged in several animals to obtain the final mean value for each nerve.

Single unit recording The electrode used was a glass micropipette (1-3 # m in tip diameter) filled with 2 M NaC1 or with 2 M NaC1 containing fast-green FCF. In some experiments, a glassinsulated tungsten microelectrode with a tip diameter of 2-4 #m was used. The loca-

450 tion of each penetration was restricted to the chorda tympani projection area. When the electrode touched the surface of the cortex, 3 ~ agar dissolved in 0.9 ~ NaC1 was used to cover the exposed cortex to minimize the movement of the brain surface, and then the electrode was inserted carefully with a micromanipulator. After conventional amplification and display the responses were photographed during the experiment. In some animals, a pair of platinum wire electrodes were attached to the ipsilateral chorda-lingual trunk under the jaw which was left intact and maintained the connection between its receptors and the brain. Electrical pulses were delivered when necessary to examine the latency of the unitary response. During the experiment the location of each recording site was marked on an enlarged photograph of the brain. At the conclusion of the experiment the locations were marked directly on the cortical surface with India ink applied to the tip of the recording electrode. After perfusion (107o formalin) the brain was removed and photographed.

Gustatory stimulation In performing gustatory stimulation, the rabbit's mouth was opened by pressing on the lower incisors and the anterior part of the tongue was covered with a lucite flow-chamber with an inlet and outlet. The inlet was connected with a rubber tube to a funnel fixed about 30 cm above the flow-chamber. Taste solutions and rinsing water were applied to this funnel. As test chemicals, reagent grade sucrose, DL-alanine, HC1, tartaric acid, NaC1, LiCI, NH4C1, CaCI2, KCI, MgC12, J. P. quinine.HCt, sodiumsaccharin and commercial grade monosodium glutamate dissolved in distilled water were used. About 25 ml of a test solution were passed by gravity flow, followed after each stimulation by tap water. In order to avoid the effect of the preceding stimulus, the interval between successive stimulations was at least 1 min. Each chemical stimulation was applied 2 or 3 times and we applied as many stimulants as possible during the time that recordings from the unit could be maintained. The application of the stimulus was indicated by a signal from a foot pedal at the moment of application. The taste solutions and rinsing water were kept at 27-30 °C.

Histology At the end of the experiment the electrodes filled with 2 M NaCI containing fast-green FCF were fixed at the point where the unitary activity was recorded and an electric current (5 #A) was applied for 15 min to mark the electrode tip position, After this procedure the animal was anesthetized with pentobarbital and the brain was fixed by perfusing warm saline through the heart, followed by 10 ~o formalin. Serial frozen sections were cut frontally at 30 #m and stained by cresyl violet. Electrode tip marks and electrode tracks were located by microscopic examination of these serial sections. RESULTS

(1) Cortical evoked potentials Variability in the position and size of the cortical responses evoked by electrical

451 Ipsilateral

....... " Pc ~",

i

5 mm

Contralateral

t

/

t'

"x

I

Lingual nerve

N

Chorda tympani

INI loeeop bar y ngeal

nerve

Fig. 1. Composite cortical areas in the rabbit activated by electrical stimulation of the ipsilateral (upper chart) and contralateral (lower chart) chorda tympani, lingual branch of the Vth and linguotonsillar branch of the IXth nerves. Praecgr: regio precentralis granularis; Pc: regio postcentralis;

Pi: area preinsularis; ail: area insularis agranularis anterior dorsalis; i: area insularis granularis. stimulation of the lingual, chorda tympani and IXth nerves was very small. The data from 3 recording experiments for each nerve were combined and mapped on the lateral view of the cerebral cortex of the rabbit in Fig. 1. Cortical field representation and nomenclatures are from the cytoarchitectonical study of Rose 26. Insular cortex dorsal to the somatic sensory area I (SI) receives bilateral input from each chorda tympani and from the lingual-tonsillar branch of the IXth nerve. The chorda tympani projection area is situated rostral to the IXth nerve area and partially overlaps the anterior IXth nerve area. The lingual nerve projection area is situated bilaterally in the ventral part of SI. The contralateral lingual nerve projection area contains a separate small focus caudal to the dominant one. The lingual area partially overlaps the dorsal chorda tympany focus. All the evoked potentials were recorded from the cortical surface and were initially positive. Judging from the extent of cortex activated and the amplitude and latency, the major input was from the ipsilateral side in case of the chorda tympani and contralateral side in case of the 1Xth and lingual nerves. The latency of the initial positive phase of the evoked potential is compared in each nerve and shown in Table I.

452 TABLE 1 COMPARISON OF THE LATENCY OF

3

NERVES

Values are means :~ S.D. Numerals in the parentheses indicate the number of experiments.

Latency in ntsec

Ner yes

Lingual Glossopharyngeal Chorda tympani

Ipsilateral

Contralateral

12.5 :~: 3.0 (6) 14,2 :i: 2,2 (5) 14,6 - 0,6 (4)

9.7 ~: 3.1 (5) 11.2 ± 1.8 (4) 16.7 ± 0.1 (3)

I n several e x p e r i m e n t s , a r e c o r d i n g m i c r o e l e c t r o d e was a d v a n c e d i n t o the c o r t i cal layers at t h e p o i n t w h e r e t h e m o s t d o m i n a n t r e s p o n s e w a s r e c o r d e d in e a c h n e r v e p r o j e c t i o n area, a n d a r e v e r s a l o f p o l a r i t y , f r o m s u r f a c e p o s i t i v e to d e e p negative, was o b s e r v e d .

(2) Single unit analysis Identification of cortical taste unit Since t h e a n t e r i o r p a r t o f t h e t o n g u e w a s c o v e r e d w i t h a f l o w - c h a m b e r , we c o u l d

A Water

"

O..SM S u c r o s e

B R-ZI-5

Water

C '

'iM NaCl"

Fig. 2. Identification of the cortical taste unit response. Upward deflection of the signal indicates the period of actual flow of the solution on the tongue surface; this is the same in the subsequent figures. In unit A, water application does not affect the spontaneous firing at all, and the sucrose stimulation evokes a long-lasting discharge. This unit is a cortical taste unit sensitive to 0.5 M sucrose. In another unit (B and C) discharges are evoked during the flowing period of both water and taste solutions, but only KC1 stimulation evokes a tong-lasting discharge. Hence, this unit is a cortical taste unit sensitive to 0.5 M KCI, but insensitive to 1 M NaC1. Time in sec.

453 not apply mechanical stimulation to the surface of the tongue with a mechanical stimulator. However, there remains the possibility that the flow of taste solutions over the tongue surface might deform the tongue and stimulate its mechanoreceptors per se. Hence, it is necessary to distinguish between the mechanical response and the taste response when a taste solution was delivered into the flow-chamber. Fig. 2 summarizes the criteria we have used in this study. In the figure, upward deflection of the signal means the period when the solution flowed in the flow-chamber; the rinsing of the tongue was not shown in this figure. Before application of the taste solution, the tongue was well rinsed with water at the same temperature as that of the successively applied taste solution for at least 1 rain to adapt the tongue to water itself and to the temperature. In the example shown in Fig. 2A, water application did not affect the rate of spontaneous discharge but application of the taste solution (0.5 M sucrose) evoked a long-lasting increased discharge. This unit was classified as positive for response to the taste stimulus and negative for the mechanical stimulation. On the other hand, as shown in Fig. 2B and C, when a unit increased its firing rate only during the period of flow of water and every chemical solution the unit was classified as a mechano-sensitive unit. If the unit increased its firing rate after the actual flow period of the sapid solution (0.5 M KCl in the figure), it was classified as a unit sensitive for this taste stimulation as well. If, however, as in Fig. 2C, the same unit did not show any prolonged increased firing rate other than the increment of discharges during the actual period of application of the taste solution, the unit was classified as negative for that stimulus (1 M NaC1 in the figure). Some units showed a decreased activity on application of a taste stimulus and were also classified as responsive to that taste stimulus. A given cortical unit response was classified as excitation or inhibition when the evoked discharge rate for the first 5 sec was larger or smaller than the average spontaneous discharge rate of that unit. Following the above-mentioned criteria, 24 units were identified as taste sensitive among a total of 123 units, of which 46 were recorded; the rest were observed during the experiments. The spontaneous and gustatory evoked discharge rates were generally very low. Although some differences were noted among neurons, almost all the neurons showed average evoked response rates of 4-7 impulses/sec; however, a few units (e.g., R-26-4 in Fig. 3) showed a very high discharge frequency in spontaneous and evoked activity.

Depth distribution of units Twenty taste responsive units were situated in the relatively deep layers between 500 # m and 1300 ,urn beneath the cortical surface, and 4 units were found between 400 p m and 500/~m. As the recording electrode could not be inserted into the cortex perpendicularly to the cortical surface, measurement of the depth at which recordings were made was obtained by the study of serial sections of the experimental brains. Laminal analysis derived from the cytoarchitectonical study of the rabbit cortex by Rose 26 showed that 7 units responding to gustatory stimulation were situated in claustrum.

454

A ,-,,-4

.

k O.5M

Nat1

'

'

Water

'"

lOmse¢ 120O)4V

B ~-19-z

m

0.514 Nat1

I

I

Water

i

Fig. 3. Examples of the units with a short latency response (A) and with a long latency response (B), Suprathreshold electrical shock applied to the tinguo-chorda trunk at the base of the mandible evokes repetitive spikes in unit A whereas it evokes none or a single spike in unit B. A~: single sweep; A~ i 5 superimposed Sweeps; Ba: 5 superimposed sweeps. Unit A responds immediately after onset of gustatory stimulation applied to the tongue and returns to the spontaneous firing level after onset of water rinsing, while in unit B evoked discharges appear about 5.5 sec after onset of stimulation and continue even after onset of water rinsing. Time in sec.

Classification of two unit types From latency measurements of units responsive to electrical stimulation:and natural chemical stimulation, two types of units could be distinguished; units with a short latency response (early response) and units with a long latency response (late response). Actual recording of typical examples of early response units and late response units are presented in Fig. 3. When the responsiveness of the 2 units was com-

i:1 A

R-19-2

0J U

n

&o ..q

oO

0

oJ

.¢.,

'iI

R-23-4

0.~

6 I'o 2'o 3'o to

/o sec

Fig. 4. Diagrams showing the impulse frequency of 2 late response cortical taste units (A and D) and 1 early response unit (C) responding to taste solutions on the tongue. Upward arrow indicates the onset of stimulation and downward arrow the onset of water rinsing.

455

A,.,,., 11..lli _LIII -

.lilHI Itl lltH[ JLJzgtIlt

-

FI ~I[IT~-'I-III-1

"

"11"T1111

I

l~Nlill

"

I If

I lll]II~llll--~

I

O.01M

"

/

~ll

I

II

II

. . . .

I

_ _

HC1

Water

R i i O.~M Sucrose

i

i

Water

C J D

0.SM

NaC1

i

I

l ~

Water

R-21-7

a

1M NaC1

i

I

Water

Fig. 5. Classification of the response patterns of the cortical taste units. A, B and C: excitation type response; D: inhibition type response. Time in sec.

pared, the following relative differences could be characterized. The unit showing early response has moderate spontaneous firing, a short latency to electrical stimulation of the chorda-lingual trunk (about 15 msec, in Fig. 3A~), a short latency to gustatory stimulation, and responds repetitively to each suprathreshold electrical stimulus (Fig. 3Ab). On the other hand, the unit of late response has a very low or no spontaneous firing, a long latency to electrical stimulation (about 45 msec, in Fig. 3B~), a long latency to gustatory stimulation and produces only one spike or no response to the electrical stimulation (Fig. 3Ba). Within a total of 24 units responsive to gustatory stimulation, 21 units were classified as the early response type and 3 units as the late response type. Sample records of 2 late response units and 1 early response unit are shown in the full course of the actual recording (Fig. 4). The late response units continued firing for 15 sec in unit A and for 24 sec in unit B after the onset of the tongue rinse. The late response units evoked a gradually increasing response (A) or a steady state response (B). On the other hand, early response units stopped their firing almost immediately after the onset of the tongue rinse. Most of the early response units evoked a steady state response to gustatory stimulation of the tongue (C) whereas some units evoked a phasic (initial burst) response (an example is seen in Fig. 3A). We could not find any characteristic difference in the location or depth of responsive units of these two types. Since we could not find any direct evidence to separate the two types according to the functional significance of their taste information processing, we include both types as cortical taste units sensitive to gustatory stimulation in the following description.

456 ABCDE

FGH

I JKLMNOPQRS

TUVWX

0.SM Sucrose

o.sMD,,-alaninel 0.3~ ~a-saccharinI

~

m~

~

r n [] []

~

l

rT'i ~ []

O.OIM HCI

0.1~ .ci•

~

O.1M T a r t a r i c a c i d •

[]

~m

0.sM Ms01

[]

[]

m

[] []

r-~

[]

[]

rl-I [] [ ]

0.5M NaC1

0.sM L i c l •

r'r'!

0.sM N.,cl I

[]

~

~

[]

O.02M Quinine-HCl 0.5M C a C 1 2 ~ 0.5M K C I ~

mB[]

0.SM M g C 1 2 • Increased response

[]

NO response

B

Decreased response

Fig. 6. Diagram of response profile of 24 cortical taste units (A-X) to 14 different chemical stimuli. Not all the stimuli were tested on each unit.

Classification of the response patterns Twenty-four taste units were classified largely into 2 types according to the discharge patterns in response to taste stimulation (Fig, 5): excitation type (A, B and C) and inhibition type (D). The excitation type unit increased its firing rate in response to gustatory stimulation. Among the excitation type unit, we found one unit increasing its firing at the water rinse (Fig. 5B and C). The inhibition type unit ceased or reduced its firing with gustatory stimulation. The number of units classified into excitation and inhibition types was 21 and 4 units, respectively. We observed a unit (R-26-5) which showed either excitation type or inhibition type responses according to the stimulus quality. The inhibition type responses were divided into short-term (2 units, e.g., Fig. 5D) and long-term (2 units) responses according to the duration of the reduction in firing rate from the spontaneous level. Response profile to various kinds of chemical stimuli Fig. 6 summarizes the response profile of 24 taste units in response to 14 different chemical stimuli. This figure indicates where a unit showed increased or decreased response or no response to gustatory stimulation. The test chemicals with the concen-

457 TABLE II CLASSIFICATION OF C O R T I C A L TASTE CELL RESPONSES

Symbols and abbreviations used: + : increased response; O : no response; - - : decreased response; S: 0.5 M sucrose; H: 0.01 M HC|; N" 0.5 M NaCI; Q: 0.02 M quinine.HCl; W: water.

Type

S

H

N

Q

W

Total

I

+ 0 0 0

0 0 O 0

O + 0 0

0 0 + -F

0 + 0 +

2 1 1 1

II

111

IV V

+

+

0

0

O

2

+

0 0

+ q-

0

0

O +

0

1 1

+ + O O

+ O + +

O + -I+

g+ q4-

O + O +

2 2 2 1

+ +

+ -F

+ +

-F q-

O +

2 2

O

1

--

.

. --

.

. --

0

0

1

--

0

--

0

0

1

-F

0

--

O

O

1

Total

TABLE ]II D I S T R I B U T I O N OF SENSITIVITIES TO

4

TASTE STIMULI

A

Number o f responses 1 2 3 4

Number o f units Predicted Observed 2.92 7.80 9.05 3.84

5 6 8 5

B

Response combinations

Number of" units Predicted Observed

Sucrose-HC1 Sucrose-NaC1 Sucrose-Quinine HC1-NaC1 HC1-Quinine NaC1-Quinine

9.21 10.62 11.33 8.13 8.68 10.00

10 10 10 9 10 11

5

4

7

4

4

24

458

A

_E 2 l"*' R.19-z o.6~

. R.23-2 ,.-"~_.Gd :1

.5

i

81 B 7" -"

.o,i

/

it)

E 2~ "],

0.01

. 0 5 ,1

:5

i

Concentration of NaCl (M) Fig. 7. Concentration-response curves for 4 excitation type units (A) and 3 i~ibition type units (B). The arrows indicate tbe mean rate of spontaneous firing of each unit. Ordinate: average humber of discharges per see over the first 10 see of evoked response. Abscissa: molar cot~ntration of NaCI.

tration used in the present study are shown on the left. Each of 24 taste units is arranged alphabetically (A-X) from left to right. The chemicals and the units are arranged arbitrarily. A total of 24 cortical units were examined for responsiveness to each of the 4 conventional taste stimuli (0.5 M sucrose, 0.01 M HC1, 0.5 M NaCI and 0.02 M quinine.HC1). They were classified into 5 types shown in Table I I according to the response to taste quality. In the table, W denotes water application to the tongue, and this implies as described earlier, a mechanical rather than a water taste stimulation. Out of 24 units, 5 belonged to Type I and showed a specific response to a single taste quality. Four units belonged to Type II, responding to 2 taste qualifies, 7 to Type HI, responding to 3 taste qualities and 4 units, responding to all 4 taste stimuli, were classified as Type IV. Those units which decreased their firing rate to taste stimulation were classified as Type V. One unit out o f 4 in Type V was stimulated by one specific taste quality (0.5 M sucrose), and inhibited by another taste quality (0.5 M NaC1).

Random distribution of responsiveness of cortical units I f the stimulating effectiveness o f t b e 4 conventional taste solutions on the cortical units are independent and random, the probability o f obtaining responses to any

459 pair of the 4 stimuli would be given by the product of the probabilities of obtaining responses to each stimulus 14. It must be noted here that the following data were obtained from all 24 cortical taste units, those showing a decreased as well as those showing an increased firing rate. The probability of occurrence of responses to individual stimuli can be estimated from the ratio of responsive units to the total number of units tested (Ps = 17/24 for 0.5 M sucrose, P a ~ 13/24 for 0.01 M HCI, P~- ~ 15/24 for 0.5 M NaC1, and PQ ~- 16/24 for 0.02 M quinine-HC1). Table I I I A shows the predicted and observed numbers of units responding to 1, 2, 3 or 4 of the 4 taste solutions. There is no significant difference between the 2 sets of numbers (chi-square test, P ~ 0.05). The predicted and observed numbers of the cortical taste units responding to the possible 6 pairs of the 4 stimuli are given in Table IIIB. It is noted that the observed values are quite close to the predicted ones (chi-square test, P ~- 0.05).

Concentration-response function of the cortical units Although the main part of the data is a comparison of cortical unit responses to a variety of substances at a given concentration, 11 units were tested over a range of concentrations of NaCI solution. Graphical representations of concentration-response curves for NaCI are shown in Fig. 7 for 4 excitation type units (A) and for 3 inhibition type units (B). The rate at which response magnitude increases or decreases with increases or decreases in concentration varied for different units. The concentration-response function is not always monotonic, some units showing increasing firing rate (R-13-1 and R-23-2) and some showing decreasing firing rate (R-21-7), respectively, reduced or increased their response magnitude at higher concentration as shown in the graphs. DISCUSSION

Evoked potential Cortical responses evoked by electrical stimulation of the chorda tympani nerve have been recorded in the rat v,lv, the cat1°,11, z3, the squirrel monkeyS, 6, and the marmoset Iv. In the squirrel monkey, Benjamin and Burton 5 and Benjamin et al. 6 studied the cortical responses evoked by electrical stimulation of the chorda tympani in detail and stated that two cortical taste nerve projection sites existed ; one response region lay within the tongue portion of SI and a second area was located within the sylvian fissure on the most anterior opercular-insular cortex. They suggested the opercular-insular cortex may be a pure cortical taste area. The rest of the species studied have been shown to possess only one cortical taste nerve projection site. In our present study in the rabbit we show that the single chorda tympani projection area was situated in the insular cortex. Though it has not clearly been settled where any pure cortical taste area may be, a few electrophysiotogical investigations'5, ~1 and several degeneration experiments1,3,4, 22 have suggested that opercular-insular cortex, insular cortex, and/or claustrum may be possible locations for cortical taste area. The implication of insular cortex and claustrum in gustation is supported by our present experiments in the

460 rabbit; the chorda tympani projection area and the IXth nerve projection area are situated exclusively within the insular cortex and not in Sl, and unit activity in response to taste stimulation was recorded from the insular cortex and claustrum. The cortical chorda tympani zone revealed in our study coincides well with the taste area in which electrical activities were recorded in response to natural chemical stimulation as reported by Gerebtzoff is in the rabbit. Rose and Woolsey z5 have already suggested that the primary taste area of the rabbit may lie in or near the insular cortex and not in neocortex SI. Though the chorda tympani projection area was observed bilaterally in the rabbit, the ipsilateral projection was more prominent than the contralateral one. This ipsilaterality of chorda tympani projection has already been reported in the cortex of the squirrel monkeyS, 6 and the cat 11, and in the thalamus of the squirrel monkey s and the cat 12. In the rat thalamus lz,17 and cortex 7, the chorda tympani projection has been reported to be equally distributed bilaterally. In the rabbit, the lingual nerve projection area and the IXth nerve projection area were also observed bilaterally, but the contralateral projection was more dominant in each nerve. In the squirrel monkeyS, 6, the cortical I X t h nerve area was prominent in the ipsilateral side contrary to the present result, but Benjamin and Pthffmann 7 reported that only a contralateral cortical representation was found for the IXth nerve in the rat. This may simply result from the species difference, or may reflect the fact that the I X t h nerve of the rabbit and rat contains more non-gustatory afferent fibers than gustatory fibers. Benjamin 3 has proposed that taste input projects ipsilaterally only and the contralateral projection of the chorda tympani and the IXth nerve represents non-gustatory pathways. Spatial relationships among the 2 or 3 nerve projection areas were essentially the same as those reported in the squirrel monkey 6, the cat11, 2a and the rat 7. The characteristic differences in the rabbit are that the chorda tympani and the IXth nerve projection areas are situated exclusively in the insular cortex and that there are not coextensive cortical loci for the 3 nerve inputs, that is, the 3 projection areas on the cortex are essentially non-overlapping.

Single unit analysis The evoked response rate in the cortical units recorded in this experiment were very low except in a few units. Some possible explanations m a y be given. First, the animal had not completely recovered from the anesthesia. Scott and Erickson 3° demonstrated that the evoked response rate in thalamic neurons was severely depressed by systemic anesthesia (pentobarbital sodium or chloratose) in the rat. This seems to be the most likely explanation, but another possibility is that the thalamic neuron activity might normally be deamplified at the cortical level. This idea is derived from results in the rat z0 and the squirrel monkey 3 where the average discharge frequency o f thalamic neurons was less than that of the second-order neurons in the nucleus of the tractus solitarius. It has been shown that the rabbit responds to distilled water and the spontaneous neural activity in the rabbit is reduced when low concentrations of NaCI were applied to the surface of the tongueZ,24, az. In our present study the solvent of the taste

461 solutions was water and the tongue was rinsed with water. Hence, the cortical unit activity sensitive to gustatory stimulation may partly be due to the solvent of the stimulating solution. However, in our pilot study, it was found that after prolonged adaptation of the tongue surface with water, successive water application did not cause any significant change in spontaneous activity (i.e., activity under water adaptation of the chorda tympani). Moreover, the concentration of each of the solutions except the NaCI solutions used for testing the stimulus intensity-response relationship was selected to be strong enough to evoke only an increasing positive response of the chorda tympani. For these reasons, it may be certain that the cortical unit activity in response to taste solutions was derived from the excitation of peripheral taste nerves caused by the stimulatory action of the chemicals contained in the solutions. F r o m the measurement of latency of the taste sensitive units, taste responses were divided into 2 groups: early response (91.7 ~ of the taste units) and late response (8.3 ~). The early repetitive response of a cortical neuron to an electric shock applied to the peripheral nerve is a normal event. Neurons showing similar early repetition to electric shock to the peripheral nerve are observed in the thalamic gustatory relay nucleus iv. It is reasonable to suggest that the early response is the direct and primary cortical response. On the other hand, it seems to be somewhat difficult to interpret the late response. It may be suggested that the late response results from activation of the cell via afferent inflow other than the direct thalamo-cortical projection, possibly involving a greater number of pre-synaptic interneurons. How can the cortical taste neurons operate to discriminate taste quality? This problem seems very difficult to solve from our present results, obtained from a limited number of single unit recordings. We can, at best, point out some basic properties of the cortical taste cells which possibly contribute to taste information processing at the cortical level in the rabbit. (1) Sensitivities of cortical neurons to the 4 conventional taste stimuli are mutually independent and are randomly distributed among cortical neurons. It should be noted that such a random distribution of sensitivities has already been observed at the gustatory cell level2S, 29 of the rat and frog, and the peripheral taste nerve level 14 of the rat. (2) The number of units showing a specific or relatively specific response type (Type I and II in Table II) in the rabbit was unexpectedly small in contrast to results 1~ in dog and rat which showed that the number of cortical units responding to 1 or 2 kinds of stimuli was greater than that responding to 3 or more stimuli. It seems that cortical taste cells in the rabbit are not more narrowly tuned than those in the dog and rat. (3) The patterns of the cortical taste unit responses were divided roughly into 2 categories: excitation type response (83.3~,,) and inhibition type response (16.7~). We think it possible that the information transfer involves excitation rather than inhibition. If so, inhibition may serve to distinguish the excitation by restraining the activity of surrounding units. H o w could taste intensity be discriminated at the cortical level? We examined the concentration-response functions using NaC1 of varying concentrations. Out of 4 excitation type units, 3 reduced their spontaneous activity when low concentrations (below about 0.1 M) of NaC1 were applied to the tongue. This finding agrees with observations2,24, 33 from the whole chorda tympani response in the rabbit that the

462 s p o n t a n e o u s nerve activity was reduced when low c o n c e n t r a t i o n s o f NaCI flowed over the t o n g u e surface. Generally, the frequency o f discharge increases a n d decreases with an increase in NaC1 c o n c e n t r a t i o n over the whole range o f c o n c e n t r a t i o n s in excitation a n d inhibition type units, respectively. The very s t r o n g 2.0 M NaCI p r o d u c e d a lower frequency t h a n the next w e a k e r stimulus in 2 o u t o f 4 excitation units, and ~l higher frequency in 1 out o f 3 inhibition units. Pfaffmann 24 observed a similar effect which he called the ' o v e r l o a d ' p h e n o m e n o n in a n u m b e r o f p r e p a r a t i o n s in the rat c h o r d a t y m p a n i , a n d such a p h e n o m e n o n was also r e p o r t e d by M a k o u s et al. 21 in the s e c o n d - o r d e r taste units o f the rat. Thus, it m a y be c o n c l u d e d t h a t the stimulus int e n s i t y - r e s p o n s e functions observed at the p e r i p h e r a l nerve level are still fairly faithfully m a i n t a i n e d at the cortical taste cell level. Borg et al. 9 have a l r e a d y r e p o r t e d in their study on h u m a n subjects t h a t p s y c h o p h y s i c a l estimates o f taste intensity corres p o n d e d closely with the m a g n i t u d e o f the c h o r d a t y m p a n i r e s p o n s e ACKNOWLEDGEMENT The a u t h o r s are grateful to Dr, M o r l e y R. K a r e , M o n e l l C h e m i c a l Senses Center, University o f Pennsylvania, for critical r e a d i n g o f drafts o f this manuscript.

REFERENCES 1 BAGSHAW, M, H., AND PRIBRAM, K. H., Cortical organization in gustation (Macaca mulatta), J. NeurophysioL, 16 (1953) 499-508.

2 BEIDLER,L. M., FISHMAN,I. Y., AND HARDIMAN,C. W., Species differences m taste responses, Amer. J. PhysioL, 181 (1955)235-239. 3 BENJAMIN, R. M., Some thalamic apd cortical mechanisms of taste. In Y. ZOTTERMAN(Ed.), Olfaction and Taste, VoL 1, Pergamon, Oxford, 1963, pp. 309-329. 4 BENJAMIN,R. M., AND AleRT, K.. Cortical and thalamic areas involved in taste discrimination in the albino rat, d. comp. NeuroL, 111 (1959) 231-259. 5 BENJAMIN, R. M., AND BURTON, H., Projection of taste nerve afferents to anterior opercutarinsular cortex in squirrel monkey (Saimiri sciureus), Brain Research, 7 (1968) 22t-231. 6 BENJAMIN, R. M., EMMERS, R., AND BLOMQUIST, A. J., Projection of tongue nerve afferents to somatic sensory area I in squirrel monkey (Saimiri sciureus), Brain Research, 7 (t968) 208-220, 7 BENJAMIN.R. M., AND PFAFFMANN, C., Cortical localization of taste in albino rat, J. NeurophysioL. 18 (1955) 56-64. 8 BLOMQUrST,A. J., BENJAMIN,R. M., AND EMMERS,R., Thalamic localization of afferents from the tongue in squirrel monkey (Saimirisciureus), J. comp. NeuroL, 118 (1962) 77-88. 9 BORG, G., DIAMANT, H., OAKLEY,B., STROM,L., AND ZOTTERMAN,Y.. A comparative study of neural and psychophysical responses to gustatory stimuli. In T. HAYASHI(Ed.), Olfaction and Taste, Vol. 2, Pergamon, Oxford, 1967, pp. 253-264. 10 BURTON, H., AND EARLS, F., Cortical representation of the ipsilateral chorda tympani nerve in the cat, Brain Research, 16 (1969) 520-523. 11 COHEN, M. J., LANDGREN,S., STROM, L., AND ZOTTERMAN, Y., Cortical reception of touch and taste in the cat, Acta physiol, scand., 40, suppl. 135 (1957) 1-50. 12 EMMERS,R.. I.x)calization of thalamicprojectionofafferents from the tongue in thecat, Anat. Rec., 148 (1964) 67-74. 13 EMMERS,R., BENJAMIN,R. M., AND BLOMQUIST,A. J., Thalamic localization of afferents from the tongue in albino rat, J. comp. NeuroL, 118 (1962) 43-48. 14 FRANK,M., ANn PFAFFM~NN,C., Taste nerve fibers: a random distribution of sensitivities to four tastes, Science, 164 (1969) 1183-1185.

463 15 FUNAKOSHI,M., KASAHARA,Y., YAMAMOTO,T., AND KAWAMURA,Y., Taste coding and central perception. In D. SCHNEIDER(Ed.), Olfaction and Taste, Vol. 4, Wissenschaftliche Verlagsgesellschaft MBH, Stuttgart, 1972, pp. 336-342, 16 FUNAKOSHI,M., AND KAWAMURA,Y., Summated cerebral evoked responses to taste stimuli in man, Electroenceph. clin. Neurophysiol., 30 (1971) 205-209. 17 GANCHROW,D., AND ERICKSON,R. P., Thalamocortical relations in gustation, Brain Research, 36 (1972) 289-305. 18 GEREBTZOFF, M. A., Recherches oscillographiques et anatomophysiologiques sur les centres cortical et thalamique du gofit, Arch. int. Physiol., 51 (1941) 199-210. 19 INANAGA, K., NAKAO, H., AND FUCHIWAKI, H., Electrical excitation on the cerebral cortex by gustatory stimulation. Report 1. Wide spread evocation type of potentials of the cerebral cortex by gustatory stimulation. Report I1. Localized evocation type of potentials of the cerebral cortex by gustatory stimulation. Folia Psychiat. Neurol. Jap., 7 (1953) 1-10. 20 LANDGREN,S., Convergence of tactile, thermal, and gustatory impulses on single cortical cells, Acta physiol, scand., 40 (1957) 210-221. 21 MAKOUS,W., NORD, S., OA~:LE¥,B., AND PFAFEMANN,C., The gustatory relay in the medulla. In Y. ZOTTERMAN(Ed.), Olfaction and Taste, Vol. 1, Pergamon, Oxford, 1963, pp. 381-393. 22 PATTON,H. D,, Physiology of smell and taste, Ann. Rev. Physiol., 12 (1950) 469-484. 23 PATTON,H. D., AND AMASSIAN,V. E., Cortical projection zone of chorda tympani nerve in cat, J. Neurophysiol., 15 (1952)245-250. 24 PFAFFMANN,C., Gustatory nerve impulses in rat, cat and rabbit, J. Neurophysiol., 18 (1955) 429-44O. 25 RosE, J, E., AND WOOLSEY,C. N., Organization of the mammalian thalamus and its relationships to the cerebral cortex, Electroenceph. clin. Neurophysiol., 1 (1949) 391-404. 26 ROSE, M., Cytoarchitektonischer Atlas der Grosshirnrinde des Kaninchens, J. Psychol. Neurol. (Lpz.), 43 (1931) 355-440. 27 RUDERMAN,M. 1., MORRISON, A. R., AND HAND, P. J., A solution to the problem of cerebral cortical localization of taste in the cat, Exp. Neurol., 37 (1972) 522-537. 28 SATO, M., AND OZEKI, M., Transduction of stimuli into electrical events at the gustatory cell membrane in the rat fungiform papillae. In D. SCHNEIDER (Ed.), Olfaction and Taste, Vol. 4, Wissenschaftliche Verlagsgesellschaft MBH, Stuttgart, 1972, pp. 252-258. 29 SATO, T., Multiple sensitivity of single taste cells of the frog tongue to four basic taste stimuli, J. cell. Physiol., 80 (1972) 207-218. 30 SCOTT, T. R., JR., AND ERICKSON,R. P., Synaptic processing of taste-quality information in thalamus of the rat, J. Neurophysiol., 34 0971) 868-884. 31 SUDAKOV,K., MACLEAN, P. D., REEVES,A., AND MARINO, R., Unit study of exteroceptive inputs to claustrocortex in awake, sitting, squirrel monkey, Brain Research, 28 (1971) 19-34. 32 YAMAMOTO,T., AND KAWAMURA,Y., Summated cerebral responses to taste stimuli in rat, Physiol, Behav., 9 (1972) 789-793. 33 ZOTTERMAN,Y., Species differences in the water taste, Aeta physiol, scand., 37 (1956) 60-70.