Neurobiology and neuropsychology interpret the Prefrontal Cortex as a key region, in connection with others, for behavioural planning, i.e. for the generation and evaluation of goal oriented complex courses of actions, without actually executing them. We illustrate an abstract reward-based computational model of planning and its possible neural counterparts. We then focus on how the Dorsolateral PFC and the Striatum learn incrementally to chunk sequences of actions, thus allowing fast and hierarchically structured planning. We test the abstract model by simulating internal model formation and action chunking in motor and cognitive planning problems.
|Number of pages||7|
|Publication status||Published - Jun 2003|