Characteristics of rat lateral hypothalamic neuron responses to smell and taste in emotional behavior

Brain Research, 491 (1989) 15-32

15

Elsevier BRE 14594

Characteristics of rat lateral hypothalamic neuron responses to smell and taste in emotional behavior Kiyomi Nakamura 1, Taketoshi Ono ~, Ryoi Tamura 1, Motoichi Indo 2, Yasuhiro Takashima e and Michiaki Kawasaki 2 1Department of Physiology, Faculty of Medicine, Toyama Medical and Pharmaceutical University, Toyama (Japan) and 2Takasago International Corporation, Kamata, Ohta-ku, Tokyo (Japan) (Accepted 6 December 1988)

Key words: Lateral hypothalamic neuron; Smell; Taste; Potable; Licking; Cue signal

Single unit activity in the lateral hypothalamus (LHA) of the rat was recorded while the animal learned to discriminate cue signals. Normally preferred potables (glucose, orange, or grape solution) or intracranial self-stimulation (ICSS) were used as rewards. Electric shock or tail pinch were used as aversive stimuli. The same behavior, licking, was the response required to either obtain the rewarding stimuli or avoid the aversive ones. For positive reinforcement a rat was rewarded with fluid or ICSS upon licking a spout presented in front of its mouth. In negative reinforcement experiments, an aversive stimulus, electric shock or tail pinch, was applied if the rat did not lick the spout. Solutions having smell only, taste only, or smell-plus-taste, were prepared from oranges or grape extract. Of 392 neurons analyzed, 256 responded differentially to rewarding and aversive stimuli, and 138 of these were tested with the 3 different solutions. Similar LHA neural responses occurred during actual drinking of the 3 kinds of solutions, as well as on recognition of the cue signal. Responses to smell only had shorter latency than responses to taste only. Neural activity in response to solutions that could be both smelled and tasted was the sum of activity in response to taste-only solutions plus that in response to smell-only solutions. Cue signal responses were rapidly acquired, usually within 2-5 trials, for both taste-only and smell-only solutions. The results indicate the integration of both taste and olfactory information by the same LHA neurons, and these neurons are involved in cue signal learning. Present results of LHA neuronal responses to taste and smell suggest that the intensity of gustation and olfaction may add together to enhance instinctive hedonic sensations. These neurons are involved in the formation of stimulus-reinforcement association in learning, and in elicitation of conditioned emotional responses. INTRODUCTION

affective and stimulus reinforcement association) 16' 61 The association c o r t e x - A M - L H A axis is thus

The m a m m a l i a n lateral hypothalamus ( L H A ) and amygdala (AM) interact in the control of emotional and motivated behavior such as feeding, drinking, and aggression 15'44, and are involved in cognitive

important in recognition of the biological significance of external sensory stimuli such as cue signals predicting food or non-food, and consequently in operant responses 8'28'37'38'49'51. Brain functions in-

functions29. A n a t o m i c a l and physiological studies indicate that A M reception of highly processed information from the sensory association cortices xL 58, and A M projections to the L H A , are essential to motivation-related learningz,22,41,45. Lesion studies

volved in olfactory or gustatory recognition of biologically significant stimuli, however, are less well documented.

suggest A M involvement in processes through which sensory stimuli gain motivational and emotional significance by interaction between the higher sensory association cortex and the L H A (i.e. stimulus-

We have shown that about one third of L H A neurons are inhibited or excited by oral sensing of glucose 34"42"43. Only these ingestion-related, rewardresponsive L H A n e u r o n s acquired similar responses to cue tones (CTS) that predicted the availability of a reward (glucose or intracranial self-stimulation

Correspondence: T. Ono, Department of Physiology, Faculty of Medicine, Toyama Medical and Pharmaceutical University, Toyama 930-01, Japan. 0006-8993/89/$03.50 © 1989 Elsevier Science Publishers B.V. (Biomedical Division)

16

(ICSS)). Relations between hypothalamic n e u r o n responses to smell, taste, and to their predicting cue signals, however, have yet to be clarified. The aim of the present study was to clarify how taste and olfactory information are integrated by L H A neu-

small screws were fastened to the skull as anchors. The stimulating electrode was soldered to the pins of an IC socket (TO-5, Toshiba Electric), which was

A

J

ComPuter

T~k

t~nn~'r~l

rons, and how these neurons are related to cue signal learning during discrimination between various rewarding or aversive stimuli. Unit activity in the rat L H A was recorded during ingestion of glucose, orange taste or smell, grape taste or smell, or the application of ICSS as rewarding stimuli, or electric shock or tail pinch as aversive stimuli, or cues predicting any of these stimuli. The same behavior, licking, was used for all responses. For positive reinforcement a rat was rewarded by the odor of a desired solution, by a taste of a desired solution, or by 1CSS when it licked a spout presented in front of its mouth. In negative reinforcement experiments an aversive stimulus, either electric shock or tail pinch was applied if the rat did not lick the spout.

B

Positive Reinforcement

a CTS1+ (1100Hz) ~ - ~ _ _ CL

Animals and surgery The subjects were 41 male albino Wistar rats weighing 250-350 g. Animals were individually housed in clear cages with free access to water and laboratory chow. The housing area was temperaturecontrolled at 23 °C, and maintained on a 24-h l i g h t - d a r k cycle (on at 07.00, off at 19.00 h). Surgical procedures were essentially the same as described previously34'35'42. Briefly, each rat was anesthetized with pentobarbital sodium (40 mg/kg, i.p.) and restrained in a stereotaxic apparatus. The skull was exposed and small holes were opened for insertion of stimulating electrodes (Fig. 1A). A stimulating electrode was implanted in the MFB caudal to the L H A at stereotaxic coordinates A, 2.5-3.8; L, 1.2-1.5; V, - 2 . 4 t o - 3 . 3 (ref. 20), which was posterior to the intended recording site. The bregma at the surface of the skull was the reference point for the coordinates according to K6nig and Klippel's rat brain atlas. The concentric bipolar electrode was a varnished, 0.3-mm diameter stainless steel tube with an inner 0.1 m m enameled wire. Electrodes were insulated except for 0.1 mm of the outer pole and 0.3 mm of the inner pole. After implantation of the stimulating electrode, 6

CTS2 +

[--]

(8000Hz)

Tube

I

r

Tube__j--I__

~_~__ ~ _ _

Lick Fluid

MATERIALS AND METHODS

b

Lick ICSS

C N e g a t i v e Reinforcement a b

CTSl-

~

CTS2 ~ (5400Hz) Tube. Lick

(2800Hz) T u b e _ _ ~ !!

Lick Electric Shock

J ~ 0

2

4

_

_

I--l__ 7~

Tail Pinch $e¢ +

0

2

4

sec

CTS: Cue Tone Stimulus CL: Cue Light Fig. 1. Experiment diagram. Rats were prepared for chronic recording by using dental cement to form receptacles for 4 modified earbars. A concentric bipolar stimulating electrode for ICSS was implanted in the posterior LHA. The rat was trained to lick when the spout was automatically placed close to its mouth (A). In positive reinforcement experiments (B), the rat was rewarded either by 5/A of a potable or by ICSS only if it licked the spout. Test solutions were 5% glucose or one of 3 other solutions - - smell-plus-taste, taste only, or smell only - - made from extracts of the natural fruits (either orange or grape) described in detail in the text. In negative reinforcement experiments (C), an aversive stimulus (electric shock or tail pinch) was applied if the rat did not lick the protruding spout. In training for CTS learning, 5 kinds of CTS were presented for 1.8 s prior to placing spout close to the rat's mouth: 1100 Hz (CTS1~) signaled glucose (Ba), 8000 Hz (CTS2+) signaled ICSS if the spout was licked (Bb), 2800 Hz (CTS1-) signaled electric shock (Ca) or 5400 Hz (CTS2-) signaled tail pinch (Cb) if the spout was not licked within 2 s after presentation and a computer-generated random tone (CTS0) signaled no reinforcement. In the positive reinforcement experiment with cue light (CL) learning to obtain flavored solutions (Ba), two red beams from light-emitting diodes were projected to the rat's eyes for 1.8 s instead of the CTS 1+.

17

then fixed firmly on the skull with acrylic cement. A short 0.2-mm inner diameter stainless steel tube was fixed to the surface of the skull as a coordinate indicator at a position 2.0-2.5 mm lateral to the intended recording site (A, 4.0; L, 4.0). A receptacle for 4 modified earbars was then formed of dental cement built up around the working region to permit the animal's skull to be later fixed painlessly in the correct stereotaxic planes (Fig. 1A). After c/)mpleting the surgery, the wound was cleaned and the scalp sutured. During the experiment, in some subjects, a thermocouple (611T, Nihon Kohden) was placed near the nasal orifice to record respiration, and blood pressure was measured by means of a pressure transducer (P50, Nihon Kohden) connected through a catheter to the carotid artery. The rats were returned to their home cages after surgery, and allowed 10-20 days for recovery. The day before beginning unit recording, a 2-mm diameter hole was drilled, under ketamine anesthesia (15 mg/kg, i.m.), in the skull over the intended recording site (A, 4.0-5.5; L, 1.1-1.9; V, -2.0 to -3.5) ipsilateral to the stimulating site, and the animals were again returned to their home cages. Just before unit recording, the dura mater was incised with a fine needle for electrode insertion under local anesthesia by a drop of 1% lidocaine. Between recording sessions the opening in the receptacle was filled with sterile saline-soaked cotton, covered with a paraffinbased ointment containing an antibiotic, and the animal was returned to its home cage.

Animal screening and training After recovery from initial surgery, each rat was deprived of food and water for about 12 h before a training and recording session. Screening and training have been described previously 34'42. Rats were tested for ICSS in an operant conditioning chamber. Each depression of a lever produced a 1.0-s train of capacitor-coupled negative square-wave pulses (frequency, 50 Hz; duration, 0.3 ms) through the hypothalamus-stimulating electrode. The threshold level for ICSS was determined and any rat for which the threshold exceeded 450/~A was excluded. In the experiment, a predetermined current above threshold was used. The rat was placed in a stereotaxic apparatus for progressively longer periods each day for several

days and trained to lick a spout to obtain either 5% glucose solution or ICSS. During training and recording sessions, the rat's body was mildly restrained in an adjustable cylinder constructed of parallel steel rods. After a few days accommodation, most rats accepted this restraint for up to 4 h per day without struggling, presumably because they could get rewards. Most rats learned to lick the spout on the first training day. After initial training for free licking, the spout was retracted and the rat was permitted to lick only when the spout was automatically extended close to its mouth for 2 s. A rat was trained to lick in 80-120 trials for a positive reinforcement of 5/A of glucose per trial or for ICSS (Fig. 1A) without CTS. A mildly aversive stimulus, either electric shock or tail pinch, was administered for negative reinforcement if the rat did not lick within 2 s after spout presentation. Aversive stimuli were applied in 70-110 trials. Electric shock was a 1.0-s train of capacitor-coupled square pulses (frequency, 50 Hz; duration, 0.3 ms; 250-400 ktA) applied to the skin of the ear, and tail pinch was a mild 2-s compression between two acrylic plates activated by an electromagnet (Fig. 1A). The strength of electric shock and tail pinch were adjusted to produce an avoidance ratio of about 20-30% averaged over all trials for each rat. The low avoidance ratio indicates that the noxious stimulation was only mildly aversive. The rats discriminated between rewarding and aversive stimuli, since previous experiments showed that they quickly stopped licking, usually within 2-5 trials, when rewarding or aversive stimulation was terminated 34"42. In training for CTS learning, one of 5 discriminative CTSs was presented for 1.8 s prior to placing the spout close to the rat's mouth. Two CTSs predicted positive reinforcement; CTS1 ÷ (1100 Hz) signaled glucose, and CTS2 ÷ (8000 Hz) signaled ICSS, if the rat licked (Fig. 1B). Two other CTSs predicted negative reinforcement; CTS1- (2800 Hz) signaled electric shock, and CTS2- (5400 Hz) signaled tail pinch, if the spout was not licked within 2 s after presentation (Fig. 1C). One CTS (CTS0, random computer generated tone) indicated no reward even if the rat licked the spout. The total number of trials per day was 700-800 in 4 h from 20.00 to 24.00 h. Licking was signaled by a photoelectric sensor triggered by the tongue.

18 Eight kinds of liquid reward were available from a spout in front of the rat's mouth. These liquids were 5% glucose, 3 kind of orange solutions (tasteonly, smell-only and smell-plus-taste), 3 kinds of grape solutions, and water. Among these liquids, 5% glucose, in accordance with a previous experiment 42, was used for the screening of attention or arousal-related neurons from smell or tasterelated neurons. Distilled water was used for rinsing the spout.

Recording procedures, data analysis and unit screening Single unit action potentials were recorded from the hypothalamus using double- or multi-barrelled glass microelectrodes with one tip extending 10-20 ~m beyond the other(s). Since the rat could be maintained quietly in a comfortable condition while receiving available rewards (Fig. 1A), movement artifacts produced by licking were far less than those due to bar press behavior. Differential recording effectively eliminated movement artifact produced during licking, so stable unit activity could be recorded for a long time. Signals of unit activity were amplified with a conventional high input impedance (FET) pre-amplifier and a main amplifier, and converted to standardized pulses in a windowdiscriminator that detected spike signals with duration shorter than 2 ms and discriminated between low and high amplitudes 34'42'43. During brain stimulation and electric shock, unit activity was isolated from stimulus artifacts by this window discriminator. Stimulation was applied through an isolation unit. Four kinds of trials were repeated in blocks of 20 trials: 10 rewarding or aversive stimuli alone and 10 rewarding or aversive stimuli signaled by CTS. lntertrial intervals were random (5-20 s) within a block, and successive blocks were separated by at least 1 min. Rewarding stimulation was always used first, because of the difficulty of testing with aversive stimulation. Although aversive stimulation was mild, it was disturbing to the animal, and the signalto-noise ratio of unit recording sometimes decreased during aversive stimulation. Signals of unit firing and licking behavior were passed through a pulse former and then to a rate meter. Unit firing-rate and cumulative licks were always monitored. The signals were analyzed by two

computers. One computer (ATAC-450, Nihon Kohden) calculated and displayed a peritrigger histogram (12.8 s in 100-ms bins) on line to ascertain the nature of the response. The other computer (PDP11/ 34, DEC) stored the event times of CTS signals, spikes, and licks, and displayed rasters and histograms (usually in 50-ms bins) off line to measure the response latency. Both neural and behavioral data were summed in histograms in successive 50-ms (PDP 11/34) or 100-ms (ATAC-450) bins for 3 periods. The first was the pretest control period (3.2 s, spontaneous activity); the second was the CTS period (1.8 s); and the third was the rewarding or aversive stimulation period (4 s). Response nature, excitation or inhibition of unit firing, was determined by analysis of variance (ANOVA) between the pretest control period and each stimulus period. The criterion for significance was P < 0.05. The relation between number of responsive neurons and different stimuli was tested by ?(2-test.

Analysis of smell and~or taste solution A previous experiment showed that some L H A neurons responded similarly to both rewarding (glucose and/or ICSS) and aversive stimuli (electric shock and/or tail pinch) and were related to attention or arousal 42. Neurons of this type were omitted from the results of tests of different kinds of liquids. In the present study, if a neuron discriminated rewarding and aversive stimulation, one of the 3 types of orange or grape solution was administered for additional trials of CTSl+-solution. Each of 3 kinds of solution of orange or grape were tested in blocks of 20 trials; 10 trials of solution alone and 10 of solution signaled by CTS1 +. The first few trials were omitted from the data analysis. Each block was separated by several trials of water alone for rinsing the spout. Data from the last few of each group of water trials were used for control. Both smell and taste, then taste-only, and then smell-only solutions were tested sequentially. If these 3 blocks were tested successfully, 3 additional blocks were tested for cue light (CL)- solution learning, in which red beams from light-emitting diodes were projected for 1.8 s in place of CTS1 +. Both neural and behavioral data for 3 of the 6 kinds of solutions plus water were analyzed in 3 periods: pretest control (3.2 s), cue signal (CTS or cue light)

19

(1.8 s), and smelling and ingestion after liquid presentation (4 s). Excitation or inhibition of unit firing was analyzed by ANOVA between pretest control and each stimulus period. The differences in response latencies to 3 kinds of solutions were analyzed by Kolmogorov-Smirnov test during the smelling and ingestion period. The criterion for significance was P < 0.05. Solutions with taste-only were 13% sucrose with either 0.3% citric acid added for orange taste (pH, 3.1 + 0.3), or 0.35% tartaric acid added for grape taste (pH, 2.9 + 0.3). There was no significant pH difference among two taste solutions. Since the sucrose, citric acid and tartaric acid used were all pure, and their odors, if any, cannot be identified by humans, it was assumed that the taste-only solutions contained no odor component. Although this might not be true for the rat, the results obtained can only be explained by this assumption. To make the orange smell solution, orange peel oil which contains 0.5% odorous substance (limonene 0.3%, octanol 0.1% and others) was extracted and incorporated in 0.15% concentration in distilled water. To make the grape smell solution, grape extract which contains 6% odorous substance (ethyl propionate 2%, methyl anthranilate 2%, ethyl maltol 0.9%, maltol 0.5%, ethyl butyrate 0.2% and others) was incorporated in 0.1% concentration in distilled water. Smell-only solutions of either orange or grape contained no acid. Solutions with smell-plus-taste were made by combining the components described above for smell and taste. All test materials were presented to rats at room temperature.

Histology Histological verification of the recording and stimulating sites was essentially the same as previously described 34'42. At the end of each recording session, the coordinates of the recording electrode site were precisely measured from the center of the coordinate indicator tube that had been implanted at the time of surgery. A steel electrode, insulated except at the tip (length 30/tm) was then placed precisely at the recording electrode site, under ketamine anesthesia. Because dura mater was incised, each electrode could be inserted vertically with no horizontal deviation. One recording site in each electrode track was marked by an iron deposit

created by passing a 20-/~A positive current for 30 s. Several penetrations and markings were made for each rat. After all experiments were completed, the rat was deeply anesthetized with sodium pentobarbital (80 mg/kg, i.p.) and several additional sites were also marked by iron deposits. The stimulating site was also marked by an iron deposit. Each animal was then perfused transcardially with physiological saline followed by 10% buffered formalin containing 2% potassium ferricyanide. The brain was removed and cut coronally in 75-/tm sections, and stained with Cresyl violet. All marking and stimulating sites were then carefully verified microscopically, and recording and stimulating sites were reconstructed using plates from the atlas of K6nig and Klippel 2°. Other recording sites along the electrode track were calculated from the stereotaxic coordinates of the marks with allowance for brain shrinkage of 6% by the histological procedures. We estimated that the accuracy of identification of the recording site using this method was less than 0.1 mm, and confirmed this by direct dye injection from the recording electrode. The precision of the relations among the marked sites confirmed our confidence in the accuracy of the loci of the unmarked sites. RESULTS

Self-stimulation behavior Of 41 rats tested, 37 licked to obtain ICSS. The data reported here are from these 37 rats. The threshold current for ICSS in the recording phase of the experiment ranged from 20 to 250/~A (120 + 75, mean + S.D., n = 37), and was similar to that used in the operant conditioning chamber and the later training phase. The stimulating sites that supported high rates of ICSS were in the posterior LHA, within the MFB, at stereotaxic coordinates ranging from A, 2.5-3.8; L, 1.2-1.5; V, -2.4 to -3.3. These sites were all posterior to the recording sites (Fig. 5). The tips of stimulating electrodes that did not support selfstimulation, even at currents of up to 450/~A, were not located within the MFB.

Lick performance in operant conditioning Table I summarizes lick performance during conditioning to different CTSs. Since food and water deprived for about 12 h before each training and

20 recording session, rats licked the spout to obtain glucose after CTS1 + with 97% correct lick responses. Correct lick responses were calculated by dividing the n u m b e r of appropriately made licks by the n u m b e r of lick trials (CTS1 or CTS2), or conversely, dividing the number of times licks were not made by the n u m b e r of no-lick trials (CTS0). Correct responses are summarized in Table I. The correct lick ratios for 3 types of orange or grape solutions were also high. A rat was rewarded with 5 /A of either glucose or one of the 3 types of orange or grape solutions in each trial, and the lick rate was steady in each session of 20 trials through a total of 200 or more trials (total amount of solution, about 1 ml) p e r day. Responses to obtain ICSS after CTS2 + were 99% correct. In the negative reinforcement tests, the electric shock and tail pinch were mild to permit analysis of the neural response during the aversive stimulation itself, so the avoidance ratio was not high. In the CST1--electric shock trials, the avoidance ratio was 25% and the CTS2--tail pinch trials it was 30%. All subjects often preferred to stay in the special restraining apparatus 34'42 even after the stereotaxic

TABLE I Lick performance during conditioning to 5 different cue tones (CTS) that predicted 11 different stimuli Lick ratio ( _ _ ) responses xlO0 trials

CTS 1+-glucose CTS2+-ICSS CTS1--electric shock CTS2--tail pinch CTS 1+-orange (smell-plus-tastesolution) CTS1 ÷-orange (taste-only solution) CTS1 +-orange (smell-onlysolution) CTS1 ÷-grape (smell-plus-tastesolution) CTS1 ÷-grape (taste-only solution) CTS1 +-grape (smell-onlysolution) CTS0-nothing (should not lick)

Neurons testedin blocksof 20 trialsfor 37rats

97% 99% 25% 30%

532 564 384 307

98%

94

98%

23

96%

16

98%

44

head clamp was r e m o v e d at the end of each session, presumably because of the comfortable situation and availability of reward. The restrained animals showed no autonomic evidence of stress during the resting state in intertrial intervals: there was no respiratory, cardiac rate, or blood pressure indication of stress. Analyzed neurons

The activity of 564 L H A neurons was recorded throughout regions from the level of the mcst posterior extent of the v e n t r o m e d i a l h y p o t h a l a m u s to the posterior level of the paraventricular nucleus. Spontaneous firing of these neurons ranged from 0.8 to 54 (22.8 + 15.2, mean + S.D.) spikes/s. Activity from 10 to 25 (15.2 + 7) neurons was analyzed in each of 37 rats. Table II summarizes the analysis of the responses. Some neurons were tested with only the rewarding stimuli (glucose and/or ICSS), but none were tested with only aversive stimulation (electric shock and/or tail pinch). W h e n both glucose and ICSS were used as rewards, they usually had similar effects on neural activity (128 inhibited by both, 93 excited by both) (Table I I A , 22-test, P < 0.001). Nevertheless, ICSS was s o m e w h a t m o r e effective than glucose. O n e h u n d r e d and thirty-five neurons were influenced by brain stimulation but not by glucose reward, while 26 neurons were influenced by glucose but not by brain stimulation. Both aversive stimuli (electric shock and tail pinch) also

TABLE II Correlation of LHA neuron responses to different rewarding and aversive stimuli

I, inhibition; E, excitation; N, no response. 1

98%

41

96%

12

92%

54

A. Glucose ICSS

I E N

B. Tailpinch Electric shock I E N

E

N

Total

128 9 9

19 93 17

91 44 134

238 146 160

146

129

269

544

41

!

1 16

43 12

12 14 167

54 58 195

58

56

193

307

21 TABLE III

LHA neuron responses during operant licking I, inhibition; E, excitation; N, no response.

1

E

N

Total

77 141 110 255

92 131 105 151

363 260 349 158

532 532 564 564

33 61 22 60

40 65 41 56

311 258 244 191

384 384 307 307

A. Positive reinforcement CTS 1÷ Glucose CTS2 + ICSS

B. Negative reinforcement CTS1 Electric shock CTS2 Tail pinch

tended to have similar influence on hypothalamic neural activity (41 inhibited by both, 43 excited by both) (Table liB, P < 0.001). Table III summarizes the responsiveness of hypothalamic neurons to various components of the experimental paradigm. Activity was recorded from 532, 564, 384 and 307 L H A neurons during CTSl+-glucose, CTS2+-ICSS, CTS1--electric shock and CTS2--tail pinch trials, respectively (Table IlIA,B). The activity of 392 LHA neurons was recorded during both positive and negative reinforcement (Table IV). Spontaneous firing ranged from 0.9 to 52 (22.5 + 16.1) spikes/s. Of these, 303 responded to either or both rewarding and aversive stimuli. Spontaneous firing of these 303 ranged from 1.0 to 52 (22.1 + 15.5) spikes/s. Inhibited activity was usually 0-70% of the spontaneous firing rate, and excited activity was usually 130-300% of the spontaneous firing rate. However, the statistical significance of inhibition or excitation was determined by

TABLE IV

Correlation of neuron responses to rewarding and aversive stimuli Rewarding stimuli (glucose and~or ICSS) 1 Aversivestimuli (electric shock and/ortailpinch) Total

I E N

E

N

Total

22 39 106

37 22 46

17 14 89

76 75 241

167

105

120

392

ANOVA. Relatively many neurons responded to only rewarding stimuli (152/392), and relatively few responded to only aversive stimuli (31/392), or similarly to both rewarding and aversive stimuli (44/392). About 19% (76/392) of the L H A neurons responded oppositely to rewarding and aversive stimuli. There was no statistically significant difference between the spontaneous firing rates of the different types of responsive neurons. An example is shown in Fig. 2. This neuron's activity decreased during ingestion of glucose (Fig. 2Aa), and increased in response to electric shock (Fig. 2Ba) before CTS1 + or CTS1- were associated with their respective reinforcements. When the cue tones were associated with glucose and electric shock, inhibitory responses to CTS1 ÷ (Fig. 2Ab) and excitatory responses to CTS1- (Fig. 2Bb,c) were acquired within 2-5 trials. When the animal avoided the aversive stimulus by licking, the response to CTS1tended to be diminished (Fig. 2Bb*). After the glucose reward was stopped, the inhibitory response to CTS1 ÷ extinguished in parallel with the cessation of licking (Fig. 2Ac). Of 392 neurons, 259 responded differentially to rewarding and aversive stimuli. One or more of these neurons were found in every one of the 37 rats tested. Spontaneous firing of these neurons ranged from 1.1 to 52 (22.4 + 15.9) spikes/s. About 25% of the neurons that differentiated rewarding and aversive stimuli also acquired corresponding discrimination of the respective CTS (Table IIIA,B). Of 259 LHA neurons that differentiated reward and aversion, 138 were tested with the 3 types of either orange (94 neurons)- or grape (44 neurons)flavored solutions (Table V). The remaining 121 neurons were tested for another purpose (iontophoresis of chemicals from multi-electrodes). The orange smell-plus-taste solution was tested on 94 neurons, the taste-only solution was tested on 23, and the smell-only solution was tested on 16 of these 23. Of 94 neurons tested with the smell-plus-taste solution, 39 were inhibited and 36 were excited. Of the 23 tested with the taste-only solution, 11 were inhibited and 7 were excited, and of the 16 tested with the smell-only solution, 5 were inhibited and 6 were excited. There was little or no crossover in neural response direction, i.e. excitation or inhibition, to the different or combined modalities. The

22

A

B a

Glucose Tri:6

a

ETIre:4t r ,c S h o c k

[~

II

i1

I1

SEc

b

b

CTS1 + -Glucose

Shock

CTS1--Electric

nm,lnll~lnn |lmillmmllloll II|~lllll li||lRml,imnwmmou~nnllllllunmll,,maJ~lmll III

f

I~

I

IIIIIH

IHIII

I

~lllllll~llrll RI 11111111 ii

tll

i~1111 • III

i~llfllll

III t

III

III1~11111 I ~1~111

i i i~111 i ~ ! 1 i

i

II HIIr

i l l

•

i1~

IIIIIIIIII Iiiiill.llHfl

iiiii

PI

I i

i!

i illl

II

IIIII

tl

I I II I

i

i

Illlll I III

II

iiii

IIII111 I I I I I l i l ~ l l l I I111 lit

lllllll I II

III

Ilrll~

I UlIIIIII

IIIII

I IIIIIII

i~l HI

IIIII1~

~111

II Fill Jllll

Pl I I I I I I I I I

IIII

IIi rl

~r

ii i

iIl111

II I I l U ~'11 I I I I I l e l l l I ~

l; I I N I I I

i IIII

i ii

iiii

i iii11

I I ii

iiiI

I Ir I I 1 ~ 1

11 I l f l l

iiiiI

IIIIllll I IIilnl

Illllltllllll

III Illlrlllllllllll

iii

iii

iii

I ii I

iii

IIi

ill

IN i i

I

11 I I I

I

IIII

i

Ill

I

i

i

IIII@111

IIIIIII

I1~

i i iIiiiII I

i

I

I

irl i i i

iiiliiii

III

llll i

I

I l l l I I I M IIIIIIIL

II I h i l l

ii

n t l H n non I ill i l l , a t o l l i i l l l l

~III

I• If•Ill•

I

•lip

I IIII

C

0

CTS1 +-No

2

4

6 SEC

-2

Reward

C

ii

ill

0

CTS1--Electric

IIIII

III

III

I I I H I I I

II

ii II

illm i ill

I IIII

I

IIII

~ 1 1

i i l i i i

HI llll

II I ! I I

Jill Oral

!1

I I I~

iiiii

ii

iiIi

Jlllt

IIIlllllllIIIIItlllll Hill

I I

IIi1~11~

u 1111111

IINIIII

I I m

I l l muiu ~ a l I i l i a l

mIlllllInllIIIllllilIII I Iinllll fill IIIlIIIIIH I

Ill•If ••IN III I~

I I IIII

~lmllllllllJllllllllllllllll

IlL

ni~IUnilamilnllliii

I• IIllll I I I H I I

11 INlIIIil R I I I I I I I I I I I

i

-2

I II

i mllllllll

I I I I l l l l l

imiN•l I

II~mllllll

RIIIIIIIIIIII

III

I lllnI i W l i l r

I ii f I I N I l l

II

ii

i

ii

I I II J I IIlII III

IIIll

~ I R I I N IrI IoI l

IIIllIIIIIIIINIlllIl I l

I •lit•IN

I I IIIIIIllIIIIIIfIIIIIIIIIIIII

t 2

4

6 SEC

Shock

Nn=inmLu.onmnnml • ~t i,Bll| o,. | i ,i . i ~J i•ll |~mnnm|lNmINmlrnl 'm=~'~;' ~n:~,i,',l,',i;',:~ ': q=i'I~ '|n|;r'l'~l"i,,~|'.|,',|llm~ tl~|2'~:2~,',~;r'a'||~, r,I:l:".i~,~nme

....'.'2','.,';'S,'.':,,': "',':;i.~' ;:'","2:';,',.":,,,,--,.',;;;~:'.'72~'.~".~ I I lllll l I l l I l l l l l l II I •fill•HI•Ill

IH

I I l l l

III

IIIII

I lllllIIIII~IIIIilaIII I I I I I l I I

iIiii~I~fiiiiiiiiiIii~iI~ilnll

ill I l l I I

II l l l l I If fill II Ill I • IIIIIltIIII

lltill?lilIIIII

llI

I I I I I I I I l I I I I I I I I II lllll I• IIIll

1

•fill •Ill i l l i i l i l l H i I i i i I i i l f l l l

f ?~;z

o

2

4

6 sec

Fig. 2. Opposite responses of lateral hypothalamic (LHA) neuron to signaled and unsignaled rewards for licking, or aversive stimuli for not licking. All records from the same neuron. A: inhibitory neural responses in glucose trials; a, histogram of activity in 6 trials (100-ms bins) of unsignaled licking for glucose reward; b, raster display of the first 9 trials with CTS1÷ for glucose reward; c, raster display of the first 9 trials of extinction of glucose reward. B: excitatory neural responses in electric shock trials; a, histogram of neural activity in 4 unsignaled no lick, electric shock trials; b, raster display of first 9 trials with CTS1- for electric shock (in trial with star mark, electric shock was avoided by lick); c, histograms of activity in 9 signaled no lick, electric shock trials. Time scale (same in all panels): zero and minus indicate CTS onset and pretrial control, respectively, except in Aa and Ba. Bracket: CTS. Filled triangles: presentation of tube for licking. Solid line: tail pinch or electric shock. Each dot below a raster indicates one lick. In each pair of histograms: upper, neural response; lower, lick. These conventions are the same in this and all following figures.

grape smell-plus-taste solution was tested on 44 neurons, the taste-only solution on 41 of the 44 neurons, and the smell-only solution was tested on 12 of the 41 neurons. Of 44 neurons tested with the smell-plus-taste solution, 16 were inhibited and 16 were excited. Of the 41 tested with the taste-only solution, 14 were inhibited and 14 were excited, and of the 12 tested with the smell-only solution, 5 were inhibited and 3 were excited. There was no significant difference in response ratios among the 3 types of orange and grape test solutions (Table V, X2-test, P > 0.1). Twenty-eight neurons were successively tested with the smell-plus-taste of both flavored solutions, with their taste-only solutions, and then with their smell-only solutions (orange, 16; grape,

12) (Table VI). The spontaneous firing of these neurons ranged from 2.4 to 51 (23.4 + 16.5) spikes/s. Of these, 3 neurons responded to taste only but not to smell only, 19 responded to both smell only and taste only, and none responded to smell only but not to taste. Responses to the smell-only solutions were generally similar to responses to the taste-only solutions (10 inhibited, 9 excited) (Table VI, P < 0.001). Inhibited activity usually decreased to 0 60% of the spontaneous firing rate, and excited activity was usually 150-300% of the spontaneous firing rate. The response latency for the presentation of the smell-only solutions ranged from 100 to 400 ms (220 + 140 ms, m e a n + S.D., n = 19) while latency for the taste-only solutions ranged from 350

23 TABLE V L H A neuron responses to smell-plus-taste, or taste-only or smell-only solutions and their respective predicting CTSs

I, inhibition; E, excitation; N, no response.

Orange (smell-plus-tastesolution) Orange (taste-only solution) Orange (smell-onlysolution) CTS 1÷ (orange, smell-plus-taste solution) CTS 1+ (orange, taste-only solution) CTS 1÷ (orange, smell-onlysolution) Grape (smell-plus-tastesolution) Grape (taste-only solution) Grape (smell-onlysolution) CFS 1÷ (grape. smell-plus-taste solution) CTSI ÷ (grape. taste-only solution) CTSI ÷ (grape. smell-onlysolution)

1

E

N

Total

39 11 5

36 7 6

19 5 5

94 23 16

26 5 2 16 14 5

32 3 2 16 14 3

36 15 12 12 13 4

94 23 16 44 41 12

12 10 4

13 8 3

19 23 5

44 41 12

to 2100 ms (710 + 540 ms, n = 19~. Response latency to smell-only was significantly shorter than that to taste-only (two-tailed t-test, P < 0.01). CTS responses were usually acquired within 2-5 trials for both taste only and smell only (Table VII), and extinguished equally rapidly after the reinforcement was terminated. Responses to CTSs predicting taste only were similar to responses to CTSs that predicted smell only (Table VII, P < 0.01). Typical examples of two different response types are shown in Fig. 3 and Fig. 4. The neuron in Fig. 3A responded with inhibition to CTS1 ÷ and during licking for grape solution with smell-plus-taste (Aa). The inhibitory CTS1 ÷ responses of the neuron were acquired for both the solution with taste only (Fig. 3Ab) and the solution with smell only (Fig. 3Ac).

TABLE VI Correlation o f L H A neuron responses to potables with taste only and smell only Taste solution (orange or grape) I Smell solution (orange or grape)

Total

I E N

E

N

Total

10 0 1

0 9 2

0 0 6

10

11

11

6

28

9 9

The neuron also responded with inhibition during licking for grape solutions with taste only (Fig. 3Ab) and those with smell only (Fig. 3Ac). The response latency after presentation of smell-only solutions was 150 ms and response duration was 700 ms (Ac), while the response latency after presentation of taste-only solutions was 900 ms and response duration was 1900 ms (Ab). The response latency for smell only was shorter than that for taste only (Kolmogorov-Smirnov test, P < 0.01). The neural activity in response to both smell-plus-taste (Fig. 3Aa) was the sum of the activity change in response to taste only (Ab) plus that in response to smell only (Ac). The response latency after presentation of both smell-plus-taste solutions was 150 ms, and duration was 2600 ms. This additivity was observed in 8 of 10 neurons (in 7 rats) that were similarly inhibited by both taste-only solutions and smell-only solutions. The neural responses to the 3 kinds of solutions during CTS learning are schematically designated in Fig. 3B from upper to lower, respectively, for solutions with smell-plus-taste, with taste only, and with smell only. When a cue light replaced CTS1 ÷, the inhibitory response to the cue light in the first few trials was not great (not shown), but shorter latency inhibition in response to the cue light was acquired after training (Fig. 3Ad). This same neuron was also inhibited by rewarding ICSS and CTS2 ÷ (Fig. 3Ae), it did not respond during drinking of water in intervening trials for rinsing the spout (Fig. 3Af). An example of an excitatory response is shown in Fig. 4. This neuron responded with excitation to CTS1 ÷ and during licking of grape solution with smell-plus-taste (Fig. 4Aa). The excitatory CTS1 ÷ response of the neuron occurred for both the solution with taste only (Fig. 4Ab), and the solution with smell only (Fig. 4Ac). When the cue light replaced CTS1 ÷, this neuron did not respond to the cue light that predicted both smell-plus-taste solution (Fig. 4Ba), taste-only solution (Fig. 4Bb) nor smell-only solution (Fig. 4Bc). However, this neuron did respond with excitation during or after licking grape solution with taste only (Fig. 4Ab, Bb), and with smell only (Fig. 4Ac, Bc). The response latency after presentation of smell-only solution was 100 ms and the response duration was 500 ms (Ac, Bc); the response latency after presentation of taste-only

24 solution was 1200 ms and the response duration was 2100 ms ( A b , Bb). The latency of response to smell

only was significantly shorter than that to taste only ( K o l m o g o r o v - S m i r n o v test, P < 0.01). The neural

A a CTS1 + - G r a p e ( S m e l l

d Cue L i g h t - G r a p e ( S m e l l Trained

and Taste)

Tri:3

and Taste)

Tri:3

3 '-2

b

,

,=

~

2

I

4

6

,~,,

-2

SEC

8

e

CTS 1 + - G r a p e ( T a s t e ) Tri:3

J

0

2

, ~

6

;

CTS2 + -ICSS

Tri:9

t3 I

see

3 I

I

I

p

"2

~

2

,

6

"2

sEc

~

f

C CTS1 + - G r a p e ( S m e l l )

Tri:4

2

o

4

6

B

Water

Tri:3

3

J3 I

-2

B

~

I

2

and Taste

8

4

CTS-Smell J

CTS-Taste

CTS ~w.

~2

5EC

CTS-Smell

° CTS

I

°

~

~

Smell and Taste J

CTS

i

Solution

SEC

Taste

I

1

°

/

Smell

I°/ I AVailable

I

4

6

~

scc

25

activity in response to smell-plus-taste (Fig. 4Aa, Ba) was the sum of the activity in response to taste only (Ab, Bb) plus that to smell only (Ac, Bc). The response latency after presentation of smell-plustaste solution was 100 ms and response duration was 3800 ms (Aa, Ba). This was observed in 7 of 9 neurons (in 6 rats) that were similarly excited by both taste-only solution and smell-only solution. Although, not shown, this neuron was excited by rewarding ICSS and predicting CTS2 ÷, it did not respond during drinking of water in intervening trials for rinsing the spout. The differences in neural responses to 3 kinds of solutions during CTS learning are shown schematically in Fig. 4C.

Recording sites Locations of the neuron recording sites, and characteristics of their response patterns corresponding to solutions with smell and taste are shown in Fig. 5. Most were at the middle level of the ventromedial nucleus of the hypothalamus. As described, only neurons that differentiated rewarding (glucose and/or ICSS) and aversive (electric shock and/or tail pinch) stimuli were tested for solutions with smell and/or taste (orange or grape). There was no significant difference in the locations of neurons inhibited by, and those excited by the test solutions. DISCUSSION

General In the present experiment, LHA unit activity in the rat was recorded during discrimination learning of CTS predicting glucose or ICSS as reward, or electric shock or tail pinch as aversion, using the same behavior, licking, for all responses. Previous experiments in this laboratory using the same operant paradigm revealed that glucose and ICSS usually affected LHA neuron activity similarly, and the

effects of electric shock and tail pinch on LHA neuron activity were also similar 42. One difference between glucose and ICSS was the route of administration. The neural responses to glucose were probably not chemical sensory (gustatory) responses to glucose per se, but responses to the perception of reward, because in most cases the same kinds of response were induced by both non-glucose stimuli and glucose. In the present study, about 66% (259/392) of the LHA neurons responded differentially to rewarding and aversive stimuli. Table III summarizes the results, which agreed well with previous applicable results for the LHA 42. Neurons in the LHA that differentiated between rewarding and aversive stimuli also acquired corresponding discrimination of the respective C T S s 42. LHA neural responses to the 3 solutions, those with smell-plus-taste, those with taste only and those with smell only, were tested for neurons that differentiated rewarding and aversive stimuli. In prior investigations with rats, it has been difficult to the point of being impossible to get restrained preparations to function. Some investigators immobilized the lower body by injecting a local anesthetic into the spinal epidural space 5. In the present study, this was unnecessary since the subjects performed an easily shaped response, licking, which they could do without moving the trunk, neck or limbs. We obtained high quality recordings which in itself was prime evidence that the animals were not uncomfortable, since they did not attempt to struggle, or escape, and showed no stereotyped stress reaction such as closed eyes, vocalization, abrupt limb extension or absence of vibrissae movement. Absence of autonomic changes in the intertrial rest periods affords objective evidence of lack of stress in this experimental paradigm. Autonomic responses during CTS learning will be reported in another paper. It was important that the rats were

Fig. 3. Inhibitory response of LHA neuron to 3 kinds of solutions and predicting CTSs. The neuron responded with inhibition to CTS1 ÷ and during licking for grape solution with smell-plus-taste (Aa). Inhibitory CTS1 ÷ responses of neuron were acquired for the solution with taste only (Ab) and for the solution with smell only (Ac). The neuron also responded with inhibition during licking of grape solutions with taste only (Ab), and with smell only (Ac). Response latency for smell only was shorter than for taste only. Neural activity in response to solutions with smell-plus-taste (Aa) was the sum of activity in response to taste only (Ab) plus that to smell only (Ac). When cue light replaced CTS1 +, the inhibitory response to cue light in the first few trials was not so great, but a sharp inhibitory response was acquired after training (Ad). The neuron was inhibited by rewarding ICSS and predicting CTS2 ÷ (Ae). Neuron did not respond during licking water in intervening trials to rinse the spout (Af). Bracket: CTS. Filled triangles: presentation of fluid and tube for licking. B: schematic designation of neural responses to three kinds of solution with smell-plus-taste, taste only and smell only from upper to lower, respectively in CTS learning.

26 rewarded by ICSS for several minutes before and after unit recording and that the rats were acclimated

to humans and tamed by gentle handling before and during the preparation and the experiments. The

A

B a CTS1 +-Grape(Smell

and Taste)

a cue Light-Grape(Smell

Tri:6

and Taste)

Tri:6

3 3 -2

0

2

4

SEC

6

'-z

~

2

4

s

~EC

b cue Light-Grape(Taste)

b CTS 1+ -Grape(Taste) TrI'R

3

'-2

o

2

~

6

8

"2

5EC

o

2

4

6

8

SEC

C Cue Light-Grape(Smell)

C CTS 1 +-Grape(Smell)

Tri:5

3

J3 t

i

• ~ulllmm11, nR • 8

-2

SEC

0

C CTS-SmelII'

J

and T a s t e

CTS-Taste

Solution

I

CTS

I I

CTS-Smell

o

°

\

O

Smell and Taste

*

t

°

I °\

CTS

2

CTS

/

\

-

Taste

Smell

I

Available

[

4

6

8

SEC

27 head holder was specially designed in our laboratory to avoid pain or discomfort to the animal. Many subjects would often, after the first session, voluntarily enter into position in the apparatus, presumably to obtain the reward available during the experiments.

TABLE VII Correlation of LHA neuron responses to cue tone stimuli (taste and smell solutions)

I, inhibition; E, excitation; N, no response. Cue tone stimulus (taste solution)

Operant licking behavior Rejection or acceptance of available solutions is one of the most essential behaviors of an organism. In the present study, correct operant licking performance for flavored solutions with smell-plus-taste, or taste only, or smell only was more than 95%, and was quite similar to the correct lick responses for glucose and ICSS. Correct operant licks for aversive electric shock or tail pinch were less than 30%. To maintain performance for the rewarding stimuli, the animals were deprived of food and water for at least 12 h. The aversive electric shock and tail pinch, on the other hand, were kept weak to prevent unnecessary discomfort. The low ratio of correct responses to avoid these aversive stimuli probably reflect their relatively low intensity rather than lack of training. High correct response ratios during positive reinforcement and during the trials with 3 different kinds of solutions indicate that the animals readily responded to the instinctive or hedonic dimensions of the 3 types of orange or grape solutions, glucose and ICSS. A previous experiment 35 showed that taste-associated CTS learning was impaired by microinjection of a local anesthetic into the AM. It is known that taste reactivity based on the hedonic dimension of taste stimulation occurs at the brainstem level without involving the cortex 12'17'55, which is in sharp contrast to the quality-coding mechanism in the cortex 62. In olfaction, it has been shown that the afferent pathway to the lateroposterior part of the orbitofrontal cortex (LPOF) does not involve

Cue tone stimulus (smellsolution)

Total

I

E

N

Total

I E

6 0

0 5

0 0

6 5

N

1

1

15

17

7

6

15

28

non-specific thalamic nuclei, but passes through the hypothalamus 57. In the monkey, when the LPOF was ablated, smell discrimination performance was severely impaired for smells not related to the animals' instinctive needs, but impairment of smell discrimination related to the instincts of food and sex was not so severe 57. It is well-known that the hypothalamus is involved in instinctive behavior. Thus, it is believed that tastes and smells related to the instincts of food and sex can be discriminated in the lower brainstem or limbic system 1'12'47 i.e. L H A or AM, whereas tastes and smells not related to these instincts need further processing and can be differentiated only when the cortex is intact 57"62. In support of this, the present data show that L H A neurons, which have intimate mutual connections with both the AM and the VTA, differentiate rewarding and aversive stimuli and their predicting

CTSs35,42. It is expected that a teleceptive modality like olfaction would be important in operant tasks involving the seeking out of beneficial, and avoiding potentially noxious, stimuli. In contrast, taste may be critical in identifying the consequences of ingestive behavior and thus provide a more potent

"¢r-Fig. 4. Excitatory responses of LHA neuron to 3 kinds of solution and predicting CTSs. A: CTS learning to obtain smell-plus-taste (Aa), taste only (Ab) and smell only solution (Ac). B: cue light learning to obtain smell-plus-taste (Aa), taste-only (Bb) and smell-only solution (Bc). The neuron responded with excitation to CTS1 +, and during licking for grape solution with smell-plus-taste (Aa). The neuron did not respond to cue light, but responded with excitation during licking for grape solution with smell-plus-taste (Ba). The excitatory CTS1 ÷ response of neuron occurred for taste-only (Ab), and for smell-only solution (Ac). But cue light response did not occur for taste-only (Bb), nor for smell-only solution (Bc). Neuron responded with excitation during licking for grape solutions with taste only (Ab, Bb), and with smell only (Ac, Bc). In both CTS and cue light learning, response latency during licking for smell only (Ac, Bc) was shorter than that for taste only (Ab, Bb). In both CTS and cue light learning, neural activity during licking of solution with smell-plus-taste (Aa, Ba) was the sum of activity in response to taste only (Ab, Bb) plus that to smell only (Ac, Bc). Bracket: CTS. Filled triangles: presentation of fluid and tube for licking. C: schematic designation of neural responses to three kinds of solutions with smell-plus-taste, taste only, and smell only from upper to lower, respectively in CTS learning.

28

\

~'~X, x

/ x .,~

)Y>./:/, , .

,"

~

#

t

* #

l"

%', It

I

\

•

',-,~

i'

¢."

tf

, , ,_-., ,"

-~-----._

"~,.~:h," ~,..,., ~

&\ ,\

\

', "\ ~

O,

•

---'_

,

I, ,1'

i

• °

;-....~.__L.../ • "~o "-

"

•

~'~.i

"- .... W ' . - ~ -",

j/'~.-:

....

rl

,, _ , , , , - _

"

~=

= ".

,x~ !,"^ J/'q." '1

..-.-,, j,, - - , - ~ -

,'

, - ~-r'~, -

~

,

/,.,--.-

%.~lui I

\\ ,,.\Y

--.

,, .:

"

•

_

~

,: ',,'

{~/nMH ",ki ~ -'"-

' ', ' 'i

t.)..-,, ,,. .- ", ,, I

\

. .,v..,.,

,.,

A4230um

,'

"' h".<':l

itl,

',, ,,~,\\',/111/://,;! . \,\~111,/;I //7".~//. ', \¢&~14'/,'~'

//~..Y~

" . . l l ~ ~

•

I

I

-"<.

^..'~.o ~. ":'

A4620.um

..

-

.,.

~,--8,,(

"k. \'~ ..~,o"~ ~,. -,!,, .->.~ .' ,' ---',1,.

A4890um

/

--

,

~

, -7 _ . . ., .

.---::

~ ° ' o

\~ .LHA .~ ,-I ) , ,, I ,' \ ' ~ i . = o. " i l l / ",- -~ t I ~',,

"-

-" ',,l:,"'-

,~

,; -

/'.~o "
'

~. / , ; . i' ~ ,

~---'-":-. ~

cA~ /~,~,.o~ :.

--

'/.~-__-7-___ /

,,, - _ . $

/~"

.

.....

\

,

*** . ~

'.

'-, o :Excitation

¢

x~

, .-"

)

" < - ~ 4 ,-" . ' ",' ~ A3990um

, ." - -" " " - ~ - - ~ J

",-

/

"q ~

-

Y./~

~

~.2

7,.:0 7.

~ " - - ~ --'.X --:>'--

• :Inhibition • :No

Response

-A" : I C S S

site

A2970pm

Fig. 5. Recording and stimulation sites and neural response patterns corresponding to solutions with smell-plus-taste (orange or grape). Note; only neurons that differentiated between rewarding (glucose and/or ICSS) and aversive (electric shock and/or tail pinch) stimuli were tested for solutions with smell and/or taste. CAI, internal capsule; DMH, dorsomedial nucleus of hypothalamus; LHA, lateral hypothalamus; ML, medial lemniscus; VMH, ventromedial nucleus of hypothalamus.

stimulus for respondent conditioning in which illness serves as the unconditioned stimulus. Such conclusions are, in fact, predicted by Garcia, Hankins, and Rusiniak's model of the proposed roles of teleceptors and interoceptors in the control of learned behavior 9. In the present experiment, both tasteonly solution and smell-only solution were equally effective in conditioning operant behavior. The high response ratio may indicate that the rats were so well-trained that they readily responded to the

instinctive or hedonic sensations of the taste-only solution and the smell-only solutions. Rats trained in go, no-go discrimination procedures showed slow acquisition in tasks with visual, auditory or taste cues. However, if they were trained with smell cues, their performance was exceptional and they rapidly learned simple smell detection and smell discrimination, and excelled in reversal or multiple problem learning tasks 54. Further information is needed for understanding how exteroceptive

29 and interoceptive systems participate in the regulation of somatovisceral, endocrine, and emotional functions of the hypothalamus and related structures. Functional anatomy of L H A neural response to smell and taste There have been several histological studies of the olfactory projection to the hypothalamus 4s'53. After making a lesion in the prepyriform cortex (PPF) of the rat, Powell et al. 4s traced degeneration of the afferent fibers to the entire anteroposterior extent of the hypothalamus via the medial forebrain bundle. Scott and Leonard 53 obtained a similar result. Electrophysiological studies verified such an olfactory projection to the hypothalamus; responses of single neurons were evoked in the hypothalamus by applying smell to the nostrils or electric pulses to the olfactory bulb (OB) 3"19"53"63. It must be concluded that there is such a connection between the OB and the hypothalamus via the PPF and probably the medial region of the AM. A conduction pathway passes from the OB to the PPF and probably the medial region of the AM and then, sequentially, to the hypothalamus and the orbitofrontal olfactory area (LPOF) 57.

We previously reported that LHA neural responses during ingestion of glucose were modulated by AM procainization but not by VTA procainization in the rat 35, and that LHA neural responses during ingestion of food were modulated by AM cooling in the monkey s . Gustatory pathways from the pontine taste area are traced directly to the LHA and the central nucleus of the A M 39'4°. The AM receives abundant afferent fibers from the insular cortex which is related to gustation 21, and densely project to the LHA via the stria terminalis and the ventral amygdalofugal pathway 4,45.52. The LHA also sends dense projections to the central nucleus of the AM via the same pathways, and the central nucleus of the AM sends axons to the lateral preoptic area, where involvement in attention or arousal is suggested 42. Major projections to the anterior and middle L H A from the posterior LHA, in which the ICSS stimulating electrode was implanted, course through the MFB. Many of these were identified as collaterals of chemically coded fibers that originate in the lower brainstem area and continue rostrally.

Therefore, some neurons in the LHA are activated not only by gustatory (or olfactory) stimulation but also by rewarding ICSS. In the present experiment, although ICSS was more effective than glucose, glucose and ICSS had similar effects on neural activity, when both affected the same neuron. This supported the hypothesis that ICSS in the posterior LHA mimics the gustation (or the olfaction) related reward effect of food for a hungry animal by similarly activating LHA neurons. In the present study, neural responses to taste only were similar to neural responses to smell only during both the CTS and ingestion periods. Neural responses to smell-plus-taste were the sum of the responses to taste only plus those to smell only. These results indicate that both gustatory and olfactory information converge on the same hypothalamic neurons, and these responsive neurons are also related to cue signal learning. Olfactory-gustatoryvisceral connections are also established directly and indirectly between the L H A and the insular gustatory cortex, the central nucleus of the AM, the nucleus of the solitary tract, the parabrachial nucleus of the pons, and the ventromedial hypothalamus 1°. A solution with smell-plus-taste may be considered to be a mixture that involves two modalities. The additive process involves a central (cognitive) mechanism. Murphy and her coworkers demonstrated that overall intensity estimates of smellplus-taste mixtures were equal to the sum of the overall intensity estimates of the odorant and the tastant when they were mixed 31,32. At the other extreme, complete summation was not observed when the estimate of overall intensity of the tasteplus-smell mixture was compared to the sum of the intensity estimates of the smell and taste of the unmixed components 7,9,13. In the present experiment many neurons (68%, 19/28) responded similarly to the smell-only solution and taste-only solution with either inhibition or excitation, although few responded to taste only but to smell only, and none responded to smell only but not to taste only. This suggests that the LHA does not greatly discriminate between taste and odor. In the present study neural responses of many neurons to smellplus-taste were the sum of responses to taste only plus those to smell only (79%, 15/19) when both smell only and taste only affected the same neuron.

30 Therefore, the present results of L H A neural responses to taste and smell suggest that perception of the intensity of oifaction and gustation may add together to enhance the overall intensity of sensation perception, i.e. hedonic sensation TM. These results also indicate the absence of odor in the taste solutions, and of taste in the odor solutions. The neural responses to taste and/or smell solutions were not tactile response to fluid on the tongue, since the neuron did not respond during drinking water in intervening trials. N e u r a l responses to cue signals

The auditory pathway has been traced from the auditory thalamic relay, the medial geniculate body (MG), to the subcortical areas, the central and lateral A M 24'25'46"60, and by a thalamocortical pathway to the neocortical auditory area 23'5°'59. From this auditory cortex, signals return to the subcortical regions, particularly to regions of the limbic system, such as the A M and hippocampus 6'~8'27. The AML H A axis is involved in taste (or olfaction)-associated CTS learning, and L H A or A M procainization stopped licking to obtain glucose 35"37'3s. Animals with A M lesions have been reported to be less responsive than controls to both rewarding and aversive stimulation 33. The lesioned animals had, as a result of altered perceptual processes involved in alerting and attention, general perceptual deficits in their ability to recognize whether stimuli were rewarding or punishing. The L H A , from which neural responses were recorded in the present study, has intimate mutual connections with the A M 26' 30,36,41,45,52,56. Highly processed sensory information REFERENCES 1 Allen, W.E, Effect of ablating the frontal lobes, hippocampi, and occipito-parieto-temporal (excepting pyriform areas) lobes on positive and negative olfactory conditioned reflexes, Am. J. Physiol., 128 (1940) 754-771. 2 Amaral, D.G., Vearey, R.B. and Cowan, W.M., Some observations on hypothalamo-amygdaloid connections in the monkey, Brain Research, 252 (1982) 13-27. 3 Baraclough, C.A. and Gross, B.A., Unit activity in the hypothalamus of the cyclic female rat. Effect of genital stimuli and progesterone, J. Endocrinol., 26 (1963) 339359. 4 Berk, M.L. and Finkelstein, J.A., Efferent connections of the lateral hypothalamic area of the rat: an autoradiographic investigation, Brain Res. Bull., 8 (1982) 511-526. 5 Cassella, J.V. and Davis, M., A technique to restrain

from the A M may converge in the L H A and elicit the learned neural responses to cue signals. Neural responses to cue signals may contribute to the initiation of emotional behavior - .34 .... . . 35 . 42. The present experiment showed that L H A neurons respond similarly to taste only and to smell only of rewarding solutions, and to their predicting CTS1 ÷ or cue light. A previous report showed that the A M - L H A axis is important in taste (or olfaction)- associated CTS learning 35. Taken together, these suggest that L H A neurons might be involved in coding for the instinctive hedonic sensations that are important to the animal's life and are involved in the formation of stimulus-reinforcement association in learning, and in eliciting conditioned emotional responses based on this stimulus-reinforcement association. Clearly, the separate and combined roles of smell and taste in food identification must be investigated further.

ACKNOWLEDGEMENTS We thank Dr. A. Simpson, Showa University, and Dr. R. Norgren, Pennsylvania State University, for advice and reviewing of the manuscript, and Miss M. Furusaki and Miss A. Hatayama for typing. We also thank Takasago International Corporation (Tokyo) for kindly supplying us with the test solutions. This research was partly supported by the Japanese Ministry of Education, Science and Culture Grantsin-Aid for Scientific Research, 63480115 and 62870006. This research was also partly supported by The Takasago Foundation in 1988. awake rats for recording single-unit activity with glass micropipettes and conventional microdrives, J. Neurosci. Meth., 19 (1987) 105-113. 6 Diamond, I.T. and Neff, W.D., Ablation of temporal cortex and discrimination of auditory patterns, J. Neurophysiol., 20 (1957) 300-315. 7 Enns, M.P. and Hornung, D.E., Contributions of smell and taste to overall intensity, Chem. Senses, 10 (1985) 357-366. 8 Fukuda, M., Ono, T. and Nakamura, K., Functional relations among inferotemporal cortex, amygdala, and lateral hypothalamus in monkey operant feeding behavior, J. NeurophysioL, 57 (1987) 1060-1077. 9 Garcia, J., Hankins, W.G. and Rusiniak, K.W., Behavioral regulation of the milieu interne in man and rat, Science, 185 (1974) 824-831. 10 Garcia, J., Lasiter, P.S., Bermuzez-Rattoni, F. and Deems,

31

11

12

13

14 15 16

17

18

19

20

21

22

23

24

25

26

27 28 29

30

D.A., A general theory of aversion learning, Ann. N.Y. Acad. Sci., 443 (1985) 8-21. Gloor, P., Amygdala. In J. Field (Ed.), Handbook of Physiology, Neurophysiology, Vol. 2, Am. Physiol. Soc., Washington, DC, 1960, pp. 1395-1420. Grill, H.J. and Norgren, R., The taste reactivity test. II. Mimetic responses to gustatory stimuli in chronic thalamic and chronic decerebrate rats, Brain Research, 143 (1978) 281-297. Hornung, D.E. and Enns, M.P., The independence and integration of olfaction and taste, Chem. Senses, 9 (1984) 97-106. Hornung, D.E. and Enns, M.P., Odor-taste mixtures, Ann. N.Y. Acad. Sci., 510 (1987) 86-90. Isaacson, R.L., The Limbic System, Plenum, New York, 1982. Jones, B. and Mishkin, M., Limbic lesions and the problem of stimulus-reinforcement associations, Exp. Neurol., 36 (1972) 362-377. Kawamura, Y. and Yamamoto, T., Studies on neural mechanisms of the gustatory salivary reflex in rabbits, J. Physiol. (Lond.), 285 (1978) 35-47. Kelly, J.B. and Glazier, S.J., Auditory cortex lesions and discrimination of spatial location by the rat, Brain Research, 145 (1978) 315-321. Komisaruk, B.R. and Beyer, C., Responses of diencephalic neurons to olfactory bulb stimulation, odor, and arousal, Brain Research, 36 (1972) 153-170. K6nig, J.ER. and Klippel, R.A., The Rat Brain: A Stereotaxic Atlas of the Forebrain and Lower Parts of the Brain Stem, Krieger, New York, 1974. Kosar, E., Grill, H.J. and Norgren, R., Gustatory cortex in the rat. II. Thalamocortical projections, Brain Research, 379 (1986) 342-352. Krettek, J.E. and Price, J.L., Amygdaloid projections to subcortical structures within the basal forebrain and brainstem in the rat and cat, J. Comp. Neurol., 178 (1978) 225-254. Krieg, W.J.S., Connections of the cerebral cortex. I. The albino rat. A. Topography of the cortical areas, J. Comp. Neurol., 84 (1946) 221-277. LeDoux, J.E., Ruggiero, D.A. and Reis, D.J., Projections to the subcortical forebrain from anatomically defined regions of the medial geniculate body in the rat, J. Comp. Neurol., 242 (1985) 182-213. LeDoux, J.E., Sakaguchi, A., Iwata, J. and Reis, D.J., Interruption of projections from the medial geniculate body to an archi-neostriatal field disrupts the classical conditioning of emotional responses to acoustic stimuli, Neuroscience, 17 (1986) 615-627. Maeda, H. and Mogenson, G.J., Electrophysiological responses of neurons of the ventral tegmental area to electrical stimulation of amygdala and lateral septum, Neuroscience, 6 (1981) 367-376. Meyer, D.R. and Woolsey, C.N., Effects of localized cortical destruction on auditory discriminative conditioning in cat, J. Neurophysiol., 15 (1952) 149-162. Mishkin, M., A memory system in the monkey, Phil. Trans. R. Soc. Lond. Set. B, 298 (1982) 85-95. Mishkin, M. and Aggleton, J., Multiple functional contributions of the amygdala in the monkey. In Y. Ben-Ari (Ed.), The Amygdaloid Complex~ Elsevier/North Holland Biomedical Press, Amsterdam, 1981, pp. 409-420. Mogenson, G.J. and Wu, M., Neuropharmacological and

31 32 33

34

35

36

37

38

39 40 41

42

43

44

45

46

47 48 49

electrophysiological evidence implicating the mesolimbic dopamine system in feeding responses elicited by electrical stimulation of the medial forebrain bundle, Brain Research, 253 (1982) 243-251. Murphy, C. and Cain, W.S., Taste and olfaction: independence vs. interaction, Physiol. Behav., 24 (1980) 601-605. Murphy, C., Cain, W.S. and Bartoshuk, L.M., Mutual action of taste and olfaction, Sens. Proc., 1 (1977) 204-211. Nachman, M. and Ashe, J.H., Effects of basolateral amygdala lesions on neophobia, learned taste aversions, and sodium appetite in rats, J. Comp. Physiol. Psychol., 87 (1974) 622-643. Nakamura, K. and Ono, T., Lateral hypothalamus neuron involvement in integration of natural and artificial rewards and cue signals, J. Neurophysiol., 55 (1986) 163-181. Nakamura, K., Ono, T. and Tamura, R., Central sites involved in lateral hypothalamus conditioned neural responses to acoustic cues in the rat, J. Neurophysiol., 58 (1987) 1123-1148. Nauta, W.J.H., Fiber degeneration following lesions of the amygdala complex in the monkey, J. Anat., 95 (1961) 515-531. Nishijo, H., Ono, T. and Nishino, H., Topographic distribution of modality-specific amygdalar neurons in alert monkey, J. Neurosci., 5 (1989) 3556-3569. Nishijo, H., Ono, T. and Nishino, H., Single neuron responses in amygdala of alert monkey during complex sensory stimulation with affective significance, J. Neurosci., 8 (1988) 3570-3583. Norgren, R., Gustatory responses in the hypothalamus, Brain Research, 21 (1970) 63-77. Norgren, R., Taste pathways to hypothalamus and amygdala, J. Comp. Neurol., 166 (1976) 17-30. Ono, T., Luiten, P.G.M., Nishijo, H., Fukuda, M. and Nishino, H., Topographic organization of projections from the amygdala to the hypothalamus of the rat, Neurosci. Res., 2 (1985) 221-239. Ono, T., Nakamura, K., Nishijo, H. and Fukuda, M., Hypothalamic neuron involvement in integration of reward, aversion, and cue signals, J. Neurophysiol., 56 (1986) 63-79. Ono, T., Sasaki, K., Nakamura, K. and Norgren, R., Integrated lateral hypothalamic neural responses to natural and artificial rewards and cue signals in the rat, Brain Research, 327 (1985) 303-306. Oomura, Y., Input-output organization in the hypothalamus relating to food intake behavior. In EJ. Morgane and J. Panksepp (Eds.), Physiology of the Hypothalamus, Handbook of the Hypothalamus, Vol. 2, Dekker, New York, 1980, pp. 557-620. Oomura, Y., Ono, T. and Ooyama, H., Inhibitory action of the amygdala on the lateral hypothalamic area in rats, Nature (Lond.), 228 (1970) 1108-1110. Ottersen, O.E and Ben-Ari, Y., Afferent connections to the amygdaloid complex of the rat and cat, J. Comp. Neurol., 187 (1979) 401-424. Pfaff, D.W. and Gregory, E., Olfactory coding in olfactory bulb and medial forebrain bundle of normal and castrated male rats, J. Neurophysiol., 34 (1971) 208-216. Powell, T.ES., Cowan, W.M. and Raisman, G., The central olfactory connexions, J. Anat., 99 (1965) 791-813. Rolls, E.T., Neuronal activity related to the control of feeding. In R.C. Ritter, S. Ritter and C.D. Barnes (Eds.), Feeding Behavior: Neural and Humoral Controls, Aca-

32 demic, New York, 1986, pp. 163-190. 50 Ryugo, D.K. and Killackey, H.P., Differential telencephalic projections of the medial and ventral divisions of the medial geniculate body of the rat, Brain Research, 82 (1974) 173-177. 51 Sanghera, M.K., Rolls, E.T. and Roper-Hall, A., Visual responses of neurons in the dorsolateral amygdala of the alert monkey, Exp. Neurol., 63 (1979) 610-626. 52 Saper, C.B., Swanson, L.W. and Cowan, W.M., An autoradiographic study of the efferent connections of the lateral hypothalamic area of the rat, J. Comp. Neurol., 183 (1979) 689-706. 53 Scott, J.W. and Leonard, C.M., The olfactory connections of the lateral hypothalamus in the rat, mouse and hamster, J. Comp. Neurol., 141 (1971) 331-344. 54 Slotnick, B.M., Brosvic, G.M. and Parker, S.R., Failure of rats to acquire a reversal learning set when trained with taste cues, Ann. N . Y Acad. Sci., 510 (1987) 627-629. 55 Steiner, J.E., The gustofacial response: observation on normal and anencephalic newborn infants. In J.F. Bosma (Ed.), Fourth Symposium on Oral Sensation and Perception, U.S. Government Printing Office, Washington, DC, 1973, pp. 254-278. 56 Takagi, H., Shiosaka, S., Tohyama, M., Senba, E. and Sakanaka, M., Ascending components of the medial forebrain bundle from the lower brain stem in the rat, with special reference to raphe and catecholamine cell groups.

57

58

59 60

61

62

63

A study by the HRP method, Brain Research, 193 (1980) 315-337. Tanabe, T., Iino, M. and Takagi, S.F., Discrimination of odors in olfactory bulb, pyriform-amygdaloid areas, and orbitofrontal cortex of the monkey, J. Neurophysiol., 38 (1975) 1284-1296. Turner, B.H., Mishkin, M. and Knapp, M., Organization of the amygdalopetal projections from modality-specific cortical association areas in the monkey, J. Comp. Neurol., 191 (1980) 515-543. Vaughan, D.W., Thalamic and callosal connections of the rat auditory cortex, Brain Research, 260 (1983) 181-189. Veening, J.G., Subcortical afferents of the amygdaloid complex in the rat: an HRP study, Neurosci. Lett.. 8 (1978) 197-202. Weiskrantz, L., Behavioral changes associated with ablation of the amygdaloid complex in monkeys, J. Comp. Physiol. Psychol., 49 (1956) 381-391. Yamamoto, T., Yuyama, N., Kato, T. and Kawamura, Y., Gustatory responses of cortical neurons in rats. III. Neural and behavioral measures compared, J. Neurophysiol., 53 (1985) 1370-1386. Yarita, H., Iino, M., Tanabe, T., Kogure, S. and Takagi, S.F., A transthalamic olfactory pathway to the orbitofrontal cortex in the monkey, J. Neurophysiol., 43 (1980) 69-85.