Abstract
The prefrontal cortex plays a crucial role in procedural rule learning; however, the specific neuronal mechanism through which it represents rules is unknown. We hypothesized that sequential neuronal activities in the prefrontal cortex encode these rules. To investigate this, we recorded neuronal activities in the medial prefrontal cortex of mice during rule learning using Ca2+ imaging. We utilized a method based on convolutional negative matrix factorization, iSeq, to automatically detect temporal neuronal sequences in the recorded data. As rule learning advanced, these neuronal sequences began to encode critical information for rule execution. In mice that had mastered the rule, the dynamics of neuronal sequences could predict success and failure of reward acquisition. Furthermore, the composition of cell populations within the neuronal sequences was rearranged throughout the learning process. These findings suggest that as animals learn a rule, the medial prefrontal cortex continually updates its neuronal sequences to assign significance to behavioural actions crucial for reward acquisition.
Keywords:
Neural decoding; Neuronal dynamics; Neuronal sequence; Prefrontal cortex; Representational drift; Reward; Rule learning; Sequential neuronal activity.