INTRODUCTION: Patients with amygdala dysfunction generally have behavioral impairment. Temporal lobe surgery might be a model of study of unilateral amygdala resection. The objective of this study was to evaluate behavioral flexibility in epileptic patients who undergo amygdala resection for epilepsy surgery and evaluate its relationship with their neuropsychiatric symptoms. MATERIAL AND METHODS: Ten epileptic patients who underwent amygdala and hippocampal resection (6 left and 4 right) matched by age and educational level with 10 healthy controls were tested with an extensive neuropsychological and neuropsychiatric battery. Psychiatric symptomatology was measured with the positive and negative syndrome scale (PANSS) and the Beck depression inventory. To assess behavioral flexibility the emotion-related visual reversal-learning task (O'Doherty et al., 2001) and the gambling task (Bechara et al., 1994) were used. RESULTS: Patient's mean scores were: Beck: 8 +/- 1.5; PANSS positive: 10 +/- 1.3, and negative: 14.4 +/- 2.2; intellectual quotient (IQ): 101.4 +/- 6.3; category number in Wisconsin card sorting test: 4.6 +/- 2.4. The emotion-related visual reversal-learning task showed significance differences in the number of reversion: healthy controls: 9.3; epileptic patients: 4.23 (p < 0.001); in the number of trials to the first reversion: healthy controls: 5; epileptic patients: 23.42 (p < 0.05). There was no correlation between reversion and depression, PANSS and IQ. CONCLUSIONS: Patients with epilepsy who undergo unilateral hippocampal and amygdala resection appear to have alterations in the reversion capacity with an emotional component that would explain the lack of behavior flexibility that they sometimes have and that are not related with either the isolated presence of executive alterations or low intellectual quotient.