Monkey hippocampal neuron responses to complex sensory stimulation during object discrimination

The purpose of this study was to investigate, during the performance of an object discrimination task, responses of neurons in the monkey hippocampal formation to the sight of several objects that have biological meaning, and compare these responses with those of amygdalar neurons studied previously using the same task. Neuronal activity in the hippocampal formation of conscious monkeys was recorded during performance of a task that led to presentation of familiar rewarding, familiar aversive, or unfamiliar objects. Of 864 neurons recorded in the hippocampal formation and adjacent cortices, 160 (18.5%) responded to the sight of a certain object(s). Responses to the sight of different kinds of objects were analyzed in detail. Nondifferential neurons (n = 73) responded to different objects with no significant difference in response magnitudes, and differential neurons (n = 87) responded to different objects with different response magnitudes. Of the differential neurons, 23 responded more strongly to rewarding objects than to other objects (rewarding-object-dominant neurons), but the magnitude of responses to objects did not necessarily correlates with the order of preferences to the objects as determined from observation of animal behavior. Aversive-object-dominant neurons (n = 13) responded more to aversive objects than to other objects. Unfamiliar-object-dominant neurons (n = 7) responded more to unfamiliar objects than to familiar objects. Selective neurons (n = 10) responded selectively to only one object or one category of objects. Fourteen of the rewarding- or averse-object-dominant neurons were tested in extinction or reversal trials. In 12 of 14 neurons, responses to a rewarding or aversive object did not change, or slightly weakened, in extinction or reversal trials. The results suggest the following. (1) Responses of rewarding- or aversive-object-dominant neurons may be involved in object-reward or object-aversion association. However, responses of many of these neurons might reflect past inputs to reinforcement rather than extant emotional processing. (2) Responses of unfamiliar-object-dominant neurons may be involved in recognition of objects based on their familiar or unfamiliar aspects. These results are further discussed and compared with responsiveness of amygdalar neurons.

Food memory: neuronal involvement in food recognition

Previous studies indicate that ablation of the temporal cortex including the amygdala (AM) and hippocampal formation (HF) induce the Kluver-Bucy syndrome in which animals cannot discriminate food from nonfood. Of 710 AM neurons tested, 129 (18.2%) responded to single sensory stimulation (48 to vision, 32 to audition, 49 to ingestion), 142 (20.0%) to multimodal stimulation and 20 to only one item with affective significance. Eight food related AM neurons were tested in reversal by salting food or introducing saline, and all responses were modulated by reversal. In HF and parahippocampal cortices (PH), 864 neurons were recorded, and 160 (18.5%) responded lo the sight of certain objects. Of these, 23 responded predominantly to food related rewarding objects, 13 to several aversive objects such as a spider model, syringe, objects associated with weak electric shock, ten to one object or one kind of object, seven to unfamiliar objects. Of 14 rewarding or aversive object-related neurons tested, responses of seven to the same test object did not change in extinction or reversal tests. Although responses of the other seven decreased in extinction or reversal tests, the magnitude of response remaining in five of those seven still exceeded that of responses to other categories. Results suggest complementary AM and HF-PH functions. The AM may be important in ongoing recognition of the affective significance of complex stimuli (food-nonfood discrimination) and the HF-PH in sustaining past affective significance.

Parabrachial gustatory neural activity during licking by rats

1. A total of 51 single neurons was recorded from the pontine parabrachial nuclei of three rats being given sapid stimuli either via intraoral infusions or during spontaneous licking behavior. In 46 neurons, sapid stimuli elicited significant taste responses; of these, 28 responded best to NaCl, 15 to sucrose, 2 to citric acid, and 1 to quinine HCl. The remaining five neurons responded significantly only to water. The mean spontaneous rate of taste neurons during the intraoral infusion and licking sessions was 11.1 +/- 1.1 and 10.8 +/- 1.2 (SE) spikes/s, respectively. 2. Of the 39 neurons tested during both licking and intraoral infusions, four responded significantly only to water via either route. The remaining 35 neurons responded significantly to at least some sapid stimuli. The best-stimulus categories remained the same regardless of the route of fluid delivery (24 NaCl best, 10 sucrose best, 1 citric acid best). When the rats were licking the stimuli, nine taste neurons responded significantly to only one sapid chemical [6 Na specific (Ns) and 3 sucrose specific (Ss)] but were more broadly tuned during intraoral infusions. Conversely, three taste neurons that responded specifically during intraoral infusions (3 Na specific) were not as specific when the animal licked the same fluids. 3. Thirty-five taste neurons were tested via both stimulus routes. These data were compared in three ways. First, for each neuron, the responses elicited during licking and intraoral infusions were compared for each of the four standard sapid stimuli. The Pearson correlation coefficients for the 35 taste neurons ranged from 0.9997 to 0.6785, with a mean at 0.953 +/- 0.012 (SE). The second comparison was between stimulus routes across chemicals. With the use of raw responses, the correlation coefficients for NaCl, sucrose, citric acid, and QHCl ranged from 0.925 to 0.778 (t test, P less than 0.0001). With the activity elicited by water subtracted (corrected responses), the correlation coefficients for NaCl, sucrose, citric acid, and QHCl were 0.900, 0.795, 0.369, and 0.211, respectively. The coefficient for QHCl was not significant (t test, P greater than 0.05). Finally, the mean responses to NaCl, sucrose, and citric acid delivered by both routes were compared and found not to differ (paired t test, P greater than 0.05). 4. In separate hierarchical cluster analyses for the licking and infusion data, the largest cluster in each contained all of the Na-best neurons and the next largest, all of the sucrose-best cells.(ABSTRACT TRUNCATED AT 400 WORDS)

Role of monkey hippocampus in recognition of food and nonfood

To investigate the role of the hippocampal formation (HF) in feeding behavior, single neuron activity in the monkey HF was recorded during performance of an operant task that included food/nonfood discrimination, drinking, and active avoidance. Of 837 neurons recorded in the HF, 155 responded to the sight of one or more objects. Of these, 82 responded to the sight of different objects with different response magnitudes, and some of these 82 responded predominantly to food-related (rewarding) objects or nonfood, aversive objects. The magnitude of response of neurons that responded predominantly to food was not necessarily correlated with the order of animal's preference for those kinds of food. For some neurons that responded predominantly to food or nonfood, effects of extinction or reversal learning on the neuronal responses were tested, and most of the neurons tested maintained their original responsiveness even after behavioral extinction or reversal learning was accomplished. The results suggest that these HF neurons may be involved in preservation of past information concerning food or nonfood.

Parabrachial gustatory neural responses to monosodium glutamate ingested by awake rats

A sample of 41 gustatory neurons isolated in the parabrachial nuclei of awake, behaving rats was tested with sapid solutions of 0.1 M monosodium glutamate (MSG), 0.5 mM of guanosine 5'-monophosphate (GMP), and a mixture of MSG and GMP as well as with 0.3 M sucrose, 0.1 M NaCl, 0.01 M citric acid, and 0.0001 M QHCl. Interneuronal correlation coefficients and factor analysis indicated that both the sodium cation and glutamic anion contributed to the activity elicited by MSG. Guanosine potentiated the responses to MSG, but only in neurons that also responded to sucrose. These results suggest that the gustatory contribution to the flavor denoted by the Japanese word "umami" may be mediated, in part, by neurons that also respond to chemical described by humans as sweet.

Amino acid and NaCl appetite, and LHA neuron responses of lysine-deficient rat

Rats' preferences for amino acids and NaCl in their drinking behavior were noted when they were fed either a control diet of gluten plus 20% purified egg protein, or (same rats, different time) a lysine-deficient diet. In control, the order of the rats' preferences was arginine greater than saline greater than monosodium L-glutamate (MSG, umami) greater than glycine greater than water greater than threonine greater than histidine greater than lysine. When fed a lysine-deficient diet the order of preference was lysine greater than saline greater than MSG greater than glycine greater than threonine greater than water greater than arginine greater than histidine. To relate neural activity with preference for amino acids and NaCl, activity of lateral hypothalamic (LHA) neurons was recorded during ingestion of MSG, lysine, arginine, glycine, saline, glucose, or water by the control and lysine-deprived condition, following a different cue tone. When the diet was lysine deficient, some neurons responded specifically to lysine ingestion. More neurons responded nondifferentially during licking in control, and responded to lysine, but fewer to other amino acid ingestion during lysine deficiency. Responses to cue tone were associated with those during licking. The present results suggest that preference for deficient amino acids might be mediated in the LHA.

Activity and distribution of learning-related neurons in monkey (Macaca fuscata) prefrontal cortex

The involvement of monkey (Macaca fuscata) prefrontal cortex (PFC) neurons in motor output decisions was studied by recording single neuron activity during 3 tasks: (a) go-left, go-right, or no-go decisions, (b) operant bar press feeding based on discrimination of food and nonfood, and (c) either delayed matching-to-sample (DMS) or DMS with response delay. Combinations of single neuron responses during the 3 tasks suggested 7 types of PFC neurons that were related to attention, choice, task-unique memory, reward anticipation, laterality, initiation of movement, and suppression of movement. The authors suggest that responses of PFC neurons do not depend on physical properties of the stimuli, but on their behavioral significance. PFC neurons might be important for appropriate behavior in response to external stimuli and internal or motivational factors.

Action of neurotropin on rat hypothalamic neurons in tissue slices

To clarify some of the actions of Neurotropin (NSP), intra- and extracellular measurements were made of 140 paraventricular (PVN) and 48 ventromedial nucleus (VMH) neurons in rat hypothalamic tissue slices in vitro while perfusing NSP, an extract from the inflamed skin of rabbits inoculated with vaccinia virus, through the recording chamber. NSP mostly decreased activity in PVN neurons (inhibition, 46; excitation, 27; excitation-inhibition, 5; no response, 62), and mostly increased it in VMH neurons (excitation, 12; inhibition, 3; no response, 33). Inhibition of PVN neurons by NSP was due to hyperpolarization with no change in membrane conductance. Since ouabain antagonized the NSP-induced inhibition of PVN neurons, the effect was probably due to activation of the sodium pump. Activity of some NSP-responsive PVN neurons was increased by increase of extracellular osmolarity, and activity of some NSP-responsive VMH neurons was increased by glucose application. The results suggest central modulation of autonomic or neuroendocrinological function by distinct NSP influence on PVN and VMH neural activity.

Responses from parabrachial gustatory neurons in behaving rats

1. The responses of a total of 70 single neurons were recorded from the parabrachial nuclei (PBN) in awake rats. In 59 neurons, sapid stimuli (0.5 ml) elicited significant taste responses. Of these 59 neurons, 10 also had significant responses to water. The mean spontaneous rate of the taste neurons was 13.4 +/- 6.9 (SD) spikes/s. Of the remaining 11 neurons, 9 responded significantly only to water; 2 had no significant responses to the standard fluid stimuli. 2. Based on the magnitude of their response to our four standard stimuli, the taste neurons were classified as follows: 42 NaCl-best, 14 sucrose-best, 2 citric acid-best, and 1 QHCl-best. Of these, 25 responded only to one of four sapid stimuli; 20 of these specific cells responded only to NaCl. All the remaining 34 neurons responded to two or more of the four sapid stimuli, with NaCl and sucrose responsiveness dominant. For the 59 taste neurons, the mean entropy for the absolute value of the responses was 0.68; for the excitatory activity alone, it was 0.58. 3. The mean responses to NaCl and sucrose concentration series increased monotonically. Except at the lowest concentration, responses to citric acid also increased monotonically, but with a lower slope. Mean responses to QHCl, however, remained stable or even decreased with increasing concentration. Thus the power functions for the NaCl and sucrose intensity-response series were higher than those of citric acid and QHCl. 4. A hierarchical cluster analysis of 59 parabrachial neurons suggested four different categories: NaCl-best, sucrose-best, citric acid-best, and QHCl-best. These categories were less evident in the two-dimensional space produced by multidimensional analysis, because the positions of NaCl- and sucrose-best neurons formed a continuum in which neural response profiles change successively from sucrose-specific to NaCl-specific. 5. The results were consistent with previous anatomic and neurophysiological data suggesting convergence in the medulla of sensory input from receptors in the nasoincisor ducts (NID) and on the anterior tongue (AT). Taste buds in the NID respond preferentially to sucrose, whereas those on the AT respond more to NaCl.(ABSTRACT TRUNCATED AT 400 WORDS)

Activity and distribution of learning-related neurons in monkey (Macaca fuscata) prefrontal cortex

The involvement of monkey (Macaca fuscata) prefrontal cortex (PFC) neurons in motor output decisions was studied by recording single neuron activity during 3 tasks: (a) go-left, go-right, or no-go decisions, (b) operant bar press feeding based on discrimination of food and nonfood, and (c) either delayed matching-to-sample (DMS) or DMS with response delay. Combinations of single neuron responses during the 3 tasks suggested 7 types of PFC neurons that were related to attention, choice, task-unique memory, reward anticipation, laterality, initiation of movement, and suppression of movement. The authors suggest that responses of PFC neurons do not depend on physical properties of the stimuli, but on their behavioral significance. PFC neurons might be important for appropriate behavior in response to external stimuli and internal or motivational factors.

Contribution of amygdalar and lateral hypothalamic neurons to visual information processing of food and nonfood in monkey

Visual information processing was investigated in the inferotemporal cortical (ITCx)-amygdalar (AM)-lateral hypothalamic (LHA) axis which contributes to food-nonfood discrimination. Neuronal activity was recorded from monkey AM and LHA during discrimination of sensory stimuli including sight of food or nonfood. The task had four phases: control, visual, bar press, and ingestion. Of 710 AM neurons tested, 220 (31.0%) responded during visual phase: 48 to only visual stimulation, 13 (1.9%) to visual plus oral sensory stimulation, 142 (20.0%) to multimodal stimulation and 17 (2.4%) to one affectively significant item. Of 669 LHA neurons tested, 106 (15.8%) responded in the visual phase. Of 80 visual-related neurons tested systematically, 33 (41.2%) responded selectively to the sight of any object predicting the availability of reward, and 47 (58.8%) responded nondifferentially to both food and nonfood. Many of AM neuron responses were graded according to the degree of affective significance of sensory stimuli (sensory-affective association), but responses of LHA food responsive neurons did not depend on the kind of reward indicated by the sensory stimuli (stimulus-reinforcement association). Some AM and LHA food responses were modulated by extinction or reversal. Dynamic information processing in ITCx-AM-LHA axis was investigated by reversible deficits of bilateral ITCx or AM by cooling. ITCx cooling suppressed discrimination by vision responding AM neurons (8/17). AM cooling suppressed LHA responses to food (9/22). We suggest deep AM-LHA involvement in food-nonfood discrimination based on AM sensory-affective association and LHA stimulus-reinforcement association.

Topographic distribution of modality-specific amygdalar neurons in alert monkey

Neuronal activity in the amygdala (AM) was recorded from alert monkeys during performance of tasks that led to presentation of rewarding or aversive stimuli. The tasks had 3 phases: (1) discrimination (visual, auditory), (2) operant response (bar pressing), and (3) ingestion (reward) or avoidance (aversion). Neuronal activity was analyzed and compared during each of these phases. Of 585 AM neurons tested, 312 (53.3%) responded to at least one stimulus in one or more of 5 major groups: vision related, audition related, ingestion related, multimodal, and selective. Forty neurons (6.8%) in the anterior dorsolateral capsule of the basolateral nuclei responded exclusively to visual stimuli (vision related). Twenty-six neurons (4.4%) further posterior in the basolateral group responded only to auditory stimuli (audition related). During ingestion an additional 41 neurons (7.0%) increased their activity (ingestion related). These were in the corticomedial group and at the boundaries between the nuclei of the basolateral group. Of these, 27 responded only in the ingestion phase, 11 during ingestion and at the sight of food, and 3 during ingestion and to certain sounds. Throughout the AM other neurons (n = 117, 20.0%) responded to visual, auditory, and somesthetic stimuli and, when tested, to involuntary ingestion of liquid (multimodal). Of these, 40 responded transiently (phasic; 36 excited, 4 inhibited). The remaining 77 maintained their altered activity into the subsequent phases of the task (tonic; 69 excited, 8 inhibited). In each of these 4 categories, most cells were activated primarily by novel or unfamiliar stimuli, and their responses habituated during repeated stimulation. A small number of cells in the basolateral and the basomedial nuclei (n = 14, 2.4%) were highly selective in that they responded specifically to one biologically significant object or sound more than to any other stimuli (selective). Some of these neurons responded to both sight and ingestion of a specific food. In summary, most AM neurons responded vigorously to novel stimuli, and some of the neurons had multimodal responsiveness. These results suggest the AM is related to processing of new environmental stimuli and to those cross-modal association.

Single neuron responses in amygdala of alert monkey during complex sensory stimulation with affective significance

Among other deficits, amygdalectomy impairs the ability of the animal to recognize the affective significance of a stimulus. In the present study, neuronal activity in the amygdala (AM) was recorded from alert monkeys while they performed tasks leading to the presentation of rewarding or aversive stimuli. Of 585 AM neurons tested, 312 (53.3%) responded to at least one stimulus in one or more of 5 major groups: 40 vision related, 26 audition related, 41 ingestion related, 117 multimodal, and 14 selective. Ingestion-related neurons were subdivided according to their responses to other stimuli: oral sensory, oral sensory plus vision, and oral sensory plus audition. Depending upon their responsiveness to the affective significance of the stimuli, neurons in the vision- and audition-related categories were divided into 2 subclasses: vis-I (26/40), vis-II (14/40), aud-I (8/26), and aud-II (18/26). All 4 subtypes usually responded to unfamiliar stimuli but seldom responded to neutral familiar stimuli. Types vis-I and aud-I responded to both positive and negative familiar stimuli. Types vis-II and aud-II responded to certain familiar negative stimuli but not to familiar positive stimuli. In vis-I neurons, responses were stronger for palatable foods than for less palatable foods. No neurons within vision-related, audition-related, and multimodal categories responded solely to positive or to negative stimuli. Of the 27 oral sensory neurons 9 were tested with saline or salted food, and 8 responded to normally aversive oral sensory stimuli in the same manner as they did to normal food or liquid (water or juice). In contrast to oral sensory neurons, all responses of 4 oral sensory-plus-vision and all of 4 selective neurons tested, as well as bar pressing behavior, were modulated by altering the affective significance of the food. These results suggest that the AM is one of the candidates for stimulus-affective association based on associative learning and memory.

A microcomputer aided control system for visual tracking task of monkey

A microcomputer aided control system for visual tracking tasks is described. A video monitor displays positions of target and pointer. In a handle task, the horizontal position of a pointer is coupled to the angular position of a manipulandum. The task involves alignment and maintenance of a pointer on the target. In a bar press task, the pointer moves incrementally after each preset number of bar presses. A microcomputer and video monitor afford system flexibility, while a panel of manual switches permits the setting up of a variety of experimental conditions without extensive keyboard inputs and has the added advantage of constant display of task details. The program is written in assembly language to meet requirements for program control speed and allow timing measurement in msec. This system may be useful in the behavioral study of physiological and neurophysiological functions.

Neuron activity in and adjacent to the dorsal amygdala of monkey during operant feeding behavior

Neuronal activity of the dorsal amygdala, the substantia innominata and the ventral putamen during bar press operant behavior was analyzed to investigate neuronal responses in various affective situations. Of 1507 neurons recorded, 431 responded to some stimuli and were classified into 6 functional categories: 64 (4.2%) indiscriminately and transiently responded to various stimuli; 98 (6.5%) responded to various objects depending on their significance, whether rewarding or punishing; 44 (2.9%) clearly responded only to certain food or objects associated with potables, but not to both; 16 (1.1%) responded to both food and objects associated with potables; 35 (2.3%) responded primarily at the sight of certain nonfood; 66 (4.4%) responded primarily in the ingestion phase. Relations between function and topography are discussed. The results suggest that the dorsal AM and adjacent areas might be important in recognizing the biological significance of objects and in procuring food.

Hypothalamic neuron involvement in integration of reward, aversion, and cue signals

The lateral hypothalamus (LHA) is involved in integrative functions related to emotion, reward, aversion, and learning. It is, however, unclear whether the medial forebrain bundle (MFB) forms a substrate common to the anterior and posterior hypothalamic areas or whether information regarding rewarding and aversive stimuli converges on and is integrated by the same hypothalamic neuron. In the present study, unit activity in the LHA and lateral preoptic-anterior hypothalamic area (lPOA-AHA) of the rat was recorded during discrimination learning of cue tone stimuli (CTS) that predicted glucose or intracranial self-stimulation (ICSS) as rewarding stimuli, or electric shock or tail pinch as aversive stimuli, using identical behavior, licking. We examined functional differences between the LHA and lPOA-AHA. In positive reinforcement experiments a rat was rewarded by glucose or ICSS only when it licked a spout presented in front of its mouth. The threshold current for ICSS was used. In negative reinforcement experiments an aversive stimulus, either electric shock or tail pinch was applied if the rat did not lick the spout. The strengths of electric shock and tail pinch were selected to produce an avoidance ratio less than 20-30%, averaged in all trials. The activity of 507 LHA and 249 lPOA-AHA neurons was analyzed during both glucose and ICSS trials. The effects of both glucose and ICSS on the same LHA or lPOA-AHA neuron were usually in the same direction, i.e., either both excitatory or both inhibitory. Of 143 LHA and 44 lPOA-AHA neurons that responded to both rewards, the responses of 117 (81.8%) LHA and 35 (79.5%) lPOA-AHA neurons to both stimuli were similar. The activity of 131 LHA and 153 lPOA-AHA neurons was analyzed in both electric shock and tail pinch trials. The effects of both electric shock and tail pinch on the same LHA or lPOA-AHA neuron were usually in the same direction. Of 29 LHA and 27 lPOA-AHA neurons that responded to both aversive stimuli, the responses of 28 (96.6%) LHA and 25 (92.6%) lPOA-AHA neurons to both were similar. The activity of 170 LHA and 195 lPOA-AHA neurons in response to both rewarding glucose and/or ICSS stimulation and aversive electric shock and/or tail pinch was analyzed. About one-third of the neurons in each area were reward specific (57/170 in LHA; 63/195 in lPOA-AHA), whereas relatively few were aversion specific in each area (21/170 in LHA; 15/195 in lPOA-AHA).(ABSTRACT TRUNCATED AT 400 WORDS)

Topographic organization of projections from the amygdala to the hypothalamus of the rat

Afferent fibers from the amygdala to subdivisions of lateral, ventromedial and dorsomedial hypothalamic nuclei were investigated in rat by retrograde transport of horseradish peroxidase. Small (intranuclear size) peroxidase deposits were placed in hypothalamic nuclei by iontophoresis of a tracer solution containing poly-L-alpha-ornithine which greatly limited diffusion. The medial, central and amygdalo-hippocampal nuclei of the amygdala were found to be the major donors of amygdaloid afferent fibers to the hypothalamus, but there was also substantial labeling of somata in cortical, basomedial, basolateral and lateral amygdaloid nuclei and the intra-amygdaloid bed nucleus of the stria terminalis. No fibers projected from the posterior cortical nucleus of the amygdala to the hypothalamus. Most amygdaloid projections to the lateral hypothalamic area originated in the anterior half of the amygdala, while projections to the ventromedial hypothalamic nucleus arose along the entire length of the amygdala except the posterior cortical nucleus. The amygdalo-hippocampal area projects to the medial hypothalamus. Other amygdaloid nuclei project to both the medial and lateral hypothalamic nuclei. These topographic organizations of amygdaloid afferent fibers to various subdivisions of the hypothalamic nuclei are discussed and compared with other anatomical studies on these connections.

Single neuron activity in dorsolateral prefrontal cortex of monkey during operant behavior sustained by food reward

The activity of 190 neurons was recorded from the dorsolateral prefrontal cortex of monkeys during an operant task that consisted of 3 phases: visual discrimination of food and non-food, bar pressing to gain access to the food and ingestion. In area 8, a fairly large proportion of the 49 recorded neurons responded in both the visual discrimination (37%) and motor initiation (35%) phases. Some functional heterogeneity seems evident within area 8 since visual discrimination responses were rostral, visuokinesis was central and motor initiation was in the caudal bank of the arcuate sulcus. Neurons in area 9 responded primarily (37%) during the bar pressing phase and less during the visual discrimination phase. Neurons in area 10 responded variously during most phases of the task--food discrimination, bar pressing, and ingestion. Neurons in the periprincipal sulcal area usually responded in the visual discrimination phase, but some which did not respond to food presented in front of the subject responded to meaningful visual or auditory cues that were related to food reward. The data suggest that neurons in the dorsolateral prefrontal cortex have multiple functions related to all phases of complex, learned feeding behavior. Functional roles of the prefrontal cortex and the lateral hypothalamus in development of feeding behavior are discussed.

Differential input from the amygdaloid body to the ventromedial hypothalamic nucleus in the rat

Differential amygdaloid afferents to anterior dorsal, anterior ventral, posterior dorsal and posterior ventral subdivisions of the ventromedial hypothalamic nucleus (VMH) were studied by means of retrograde transport of horseradish peroxidase (HRP). Injections of tracer confined to the VMH subdivisions mentioned, and enhancement of tracer uptake and transport were achieved by iontophoretic delivery of an HRP solution containing poly-L-alpha-ornithine. It was shown that the medial, central, basolateral, basomedial, lateroposterior and intercalated nuclei of the amygdala constitute afferent input sources to the ventromedial nucleus in a topographic pattern related to the various subdivisions of the VMH. This topographically organized amygdala-VMH projection is discussed against the background of the functional role that both amygdala and VMH play in the control of feeding, apart from various other autonomous functions that both brain centers are known to be concerned with.