Reinforcement, Punishment and Risk in the Basolateral Amygdala

The ability to learn behaviors that lead to reward is as fundamental for survival as is the ability to react to punishment and adapt behavior accordingly. At the neuronal level, changes that affect the circuits that support the latter ability are thought to underlie many maladaptive behaviors, such as addiction. The basolateral amygdala (BLA) has been proposed to play a central role within such circuits, given its importance in aversive stimuli-driven associative learning and in encoding values of specific reinforcers in instrumental learning. Research in the field of valence encoding in the BLA has uncovered the existence of positive and negative valence cell groups, which learn to respond to cues that predict appetitive or aversive outcomes, respectively. These studies have shown that valence cells can update the value of a cue when it changes from predicting one valence to predicting the other, and that the BLA is indeed needed to orchestrate the appropriate adaptive behavior. However, it is not known whether there exist BLA valence cells that learn to respond to actions with appetitive or aversive outcomes, nor if these cells update the value of an action whose outcome changes from being a reward to being a punishment. Because integrating new information on the outcome of actions is of paramount importance to adaptive behavior, an investigation on this matter at the cellular level is fundamental to enhance our understanding of maladaptive behaviors. Yet, the study of changes in the value of learned actions differs from that of changes in the value of learned cues in that cues can be presented to the study subjects regardless of their state, while actions need to be executed by the subjects themselves. To address this methodological challenge, we successfully established a behavioral paradigm where we suddenly introduce a probabilistic risk of punishment to a previously only-rewarded action. We use a head-mounted calcium imaging microscope to present the first account on action-encoding BLA projection- neurons in freely moving animals.