Fredericks J Megan, Elorette Catherine, Boyce Jared E, Rudebeck Peter H
Nash Family Department of Neuroscience, Lipschultz Center for Cognitive Neuroscience, and Friedman Brain Institute, Icahn School of Medicine at Mount Sinai, One Gustave L. Levy Place, New York, NY, 10029, USA.
bioRxiv. 2025 Jun 27:2025.06.26.661815. doi: 10.1101/2025.06.26.661815.
Contingent learning, the process by which specific courses of action become associated with subsequent outcomes, is dependent on the amygdala and ventrolateral prefrontal cortex (vlPFC). The amygdala and vlPFC are bidirectionally connected but it is unclear what the contribution of individual feedforward and feedback pathways is to contingent learning. Here we tested the role of amygdala projections to vlPFC in mediating two key components of contingent learning: signaling the outcome (reward/no reward) that follows a choice and maintaining representation of the choice that was made prior to outcome delivery. To test for these two aspects of contingent learning, we trained macaques to perform a probabilistic reward learning task where for separate stimulus pairs reward was either delivered immediately or after a trace interval. Inhibiting vlPFC-projecting amygdala neurons impacted contingent learning irrespective of whether there was a trace interval or not, and this effect was primarily driven by maladaptive learning on unrewarded trials. Notably, deficits in contingent learning caused by manipulating activity in the amygdala-vlPFC pathway were distinct from impairments in motivation and the ability to update the value of specific rewards in a reinforcer devaluation task. Thus, vlPFC-projecting amygdala neurons appear to play a specific role in contingent learning through signaling the outcomes of a choice, but not in maintaining a memory of the prior choice.
偶发学习是指特定行动过程与后续结果建立关联的过程,它依赖于杏仁核和腹外侧前额叶皮层(vlPFC)。杏仁核和vlPFC双向连接,但尚不清楚各个前馈和反馈通路对偶发学习的贡献是什么。在这里,我们测试了杏仁核向vlPFC的投射在介导偶发学习的两个关键组成部分中的作用:发出选择后出现的结果(奖励/无奖励)的信号,以及在结果传递之前保持所做选择的表征。为了测试偶发学习的这两个方面,我们训练猕猴执行概率奖励学习任务,其中对于不同的刺激对,奖励要么立即给予,要么在一个痕迹间隔后给予。抑制向vlPFC投射的杏仁核神经元会影响偶发学习,无论是否存在痕迹间隔,这种影响主要是由无奖励试验中的适应不良学习驱动的。值得注意的是,通过操纵杏仁核 - vlPFC通路中的活动导致的偶发学习缺陷,与动机受损以及在强化物贬值任务中更新特定奖励价值的能力受损不同。因此,向vlPFC投射的杏仁核神经元似乎通过发出选择结果的信号在偶发学习中发挥特定作用,但在保持先前选择的记忆方面并非如此。