Neural activity in either the amygdala (AM) or lateral hypothalamus (LHA) was examined while monkeys obtained food as a reward for operant bar pressing. Neurons in the AM were tested before, during, and after reversible cooling of the inferotemporal cortex (ITCx). LHA neurons were tested similarly except that the cooling probe was located in the AM. Cooling probes were chronically implanted bilaterally over the dura of the anterior ITCx in one monkey and in the lateral part of the AM in two monkeys. The activity of 43 AM neurons was analyzed. Before ITCx cooling, each AM neuron was classified into one of four groups based on its response pattern: 6 neurons responded primarily to the sight of food, 11 responded primarily to the sight of nonfood objects, 21 responded to the sight of both food and nonfood (arousal related), and 5 neurons did not respond. ITCx cooling changed the spontaneous firing rate of 15 AM neurons (2 increased, 13 decreased). Responses to the sight of food and/or nonfood of two food-related, four nonfood-related, and seven arousal-related neurons were depressed, and responses of two nonfood-related and three arousal-related neurons were enhanced. Of 17 neurons that were normally food or nonfood specific, 8 became nondiscriminative during ITCx cooling. The activity of 55 LHA neurons was tested. Of these, 22 were food related, 6 responded primarily during ingestion, 22 were arousal related, and 5 did not respond. AM cooling changed the spontaneous firing rates of 21 LHA neurons (12 increased, 9 decreased). Visual responses of nine food-related neurons and two arousal-related neurons were depressed by AM cooling. Ingestion-related responses in the AM (3 of 6 food-related neurons) were not affected by ITCx cooling, but responses of three ingestion-related LHA neurons (3/6) were depressed by AM cooling. The data suggest that dynamic interactions among the ITCx, the AM, and the LHA are important in discriminating between food and nonfood, and, consequently, in the stimulus-reinforcement process, i.e., recognition of reward or no reward. The results are discussed in terms of known anatomical data and behavioral evidence from earlier lesion experiments.