REGION-SPECIFIC PLASTIC CHANGES IN THE PREFRONTAL CORTEX DRIVE COMPULSIVE TASK PERFORMANCE

Abstract Background Compulsive actions, defined by maladaptive perseveration and excessive repeat, are shown in various psychiatric disorders such as obsessive-compulsive disorder. Recent evidence suggests that habitual decision-making causes such perseveration; however, neural mechanisms for developing compulsive habitual action remain unclear. Aims & Objectives Earlier studies suggest the involvement of abnormal activity of the prefrontal cortex regions, such as the anterior cingulate cortex (ACC) and orbitofrontal cortex (OFC) in compulsive behaviors, while the functional differences between these regions are not clear. We developed an original 2-step task utilizing operant conditioning procedures and investigated the plasticity-based mechanisms for compulsive habitual action. Method First, mice were trained with a variable-ratio task to establish goal-directed lever-pressing behavior (step 1). Then the task was changed to a variable-interval task, where frequent lever presses no longer accelerate reward delivery. This leads to the formation of habitual behavior (step 2). Habit formation was assessed by the devaluation test which was performed by giving a free access to the reward for 30 min and monitoring lever-pressing behavior for 5 min. The task-associated synaptic plasticity was evaluated by AMPA/NMDA ratio recorded by the whole-cell patch-clamp technique. Neural activity during the tasks was monitored by in vivo Ca2+ imaging with GCaMP7s. Erasure of excitatory synaptic potentiation was accomplished by optogenetic inactivation of synaptic cytoskeletal protein, cofilin via chromophore-assisted light inactivation (CALI). Results In a variable ratio task (step 1), the mice formed goal-directed behavior, which is sensitive to devaluation. After switching to a variable interval task (step 2), they developed habitual behavior, as evidenced by reduced sensitivity to the devaluation. Among the mice, about half reduced the frequency of lever presses to adapt with the change in task. In contrast, the rest of mice failed to adapt and instead maintained the frequency, suggesting that these mice showed habitual responses with inappropriate excessiveness. Such compulsive action was observed even when lever presses did not result in reward delivery. As the mice acquired goal-directed responses, an increase in AMPA/NMDA ratio was observed in the lateral OFC (lOFC) and ACC. After habit formation, the increase in the AMPA/NMDA ratio of the lOFC neurons was detected only in mice with compulsive responses, whereas in ACC, it returned to the baseline level regardless of the adaptation. Consistently, in vivo Ca2+-imaging revealed that, the neural activity of lOFC neurons during the step 2 task was higher in mice with compulsive responses than those which adapted. When the task-induced synaptic potentiation in the lOFC during the step 2 was abolished by CALI, the frequency of lever presses was significantly reduced although it was still habitual. On the contrary, eliminating synaptic potentiation in the ACC during the step 1 facilitated habit formation. Discussion & Conclusion Synaptic plasticity in the lOFC during habit formation develops excessive responses whereas lack of synaptic potentiation in the ACC facilitates habit formation. The concurrent induction of such synaptic changes in the lOFC and ACC results in compulsive habitual responses.