Basolateral amygdala lesions facilitate reward choices after negative feedback in rats

Abstract

The orbitofrontal cortex (OFC) and basolateral amygdala (BLA) constitute part of a neural circuit important for adaptive, goal-directed learning. One task measuring flexibility of response to changes in reward is discrimination reversal learning. Damage to OFC produces well documented impairments on various forms of reversal learning in rodents, monkeys, and humans. Recent reports show that BLA, though highly interconnected with OFC, may be differentially involved in reversal learning. In the present experiment, we compared the effects of bilateral, ibotenic acid lesions of OFC or BLA (or SHAM) on visual discrimination and reversal learning. Specifically, we used pairwise visual discrimination methods, as is commonly administered in non-human primate studies, and analyzed how animals use positive and negative trial-by-trial feedback, domains not previously explored in a rat study. As expected, OFC lesions displayed significantly slower reversal learning than SHAM and BLA rats across sessions. Rats with BLA lesions, conversely, showed facilitated reversal learning relative to SHAM and OFC groups. Furthermore, a trial-by-trial analysis of the errors committed showed the BLA group benefited more from incorrectly performed trials (or negative feedback) on future choices than either SHAM or OFC rats. This provides evidence that BLA and OFC are involved in updating responses to changes in reward contingency and that the roles are distinct. Our results are discussed in relation to a competitive framework model for OFC and BLA in reward processing.

Figure 1.

Representative photomicrographs of lesioned and intact OFC and BLA. A, OFC lesion, left hemisphere, +3.70 mm relative to bregma. Black arrows outlining the extent of the lesion most specific to ventral orbital cortex. B, BLA lesion and central amygdala (CeA) nucleus, left hemisphere, −2.80 mm relative to bregma. Black arrows outlining the extent of the lesion and sparing of cells in CeA. C, Intact OFC, left hemisphere, +3.70 mm relative to bregma. D, Intact BLA, left hemisphere, −2.80 mm relative to bregma.

Figure 2.

Discrimination learning is intact following OFC, BLA, or SHAM surgeries. A, Mean percentage correct by session for the first 20 sessions of initial visual discrimination learning by lesion group. B, Mean number of sessions to criterion by lesion group. Bars represent group means ± SEM.

Figure 3.

Reversal learning is facilitated by BLA lesion and impaired by OFC lesion relative to SHAM. A, Mean percentage correct by session for first 20 sessions of reversal learning by lesion group. B, Mean number of sessions to criterion by lesion group. Bars represent group means ± SEM **p < 0.01.

Figure 4.

Trial-by-trial performance for BLA, OFC, and SHAM groups during reversal learning. A, Mean percentage correct following either negative (Error+1) or positive (Correct+1) feedback by lesion group. B, Mean percentage correct following multiple trials where positive feedback (reward/correct trials) was delivered (EC analysis: E + 1 − EC(2) + 1) by lesion group. C, Mean percentage correct following multiple trials where negative feedback (unrewarded/error trials) was delivered (EE analysis: E + 1 − EE + 1) by lesion group. Bars represent group means ± SEM *p ≤ 0.05, **p < 0.01.