This chapter reviews the evidence that goal-directed behavior depends on the interactions between two different "styles" of learning mechanisms in different frontal lobe systems, specifically faster dopaminergic based plasticity in the basal ganglia; and slower more Hebbian based plasticity in the frontal cortex. As these two systems interact in recursive processing loop, the brain leans new things quickly, but also takes more time to link in more experiences and build up abstract, big-picture thoughts, and sophisticated actions. Goal-directed thought and actions are learned via dopaminergic-gated plasticity in frontal cortex-basal ganglia loops. Fast-learning mechanisms in the basal ganglia (specifically the striatum) are more specialized for the detection and storage of specific experiences that leads to reward (activation of the midbrain dopamine signals). The output of the basal ganglia trains slower-learning mechanisms in the frontal cortex. The slower cortical learning is not only less error prone; it also allows the frontal cortex to build up abstract, generalized representations that reflect the regularities across many different experiences. Recursive iterations of these loops allow bootstrapping of more complex and predictive rules and greater abstractions. One result of this plasticity are "rulemaps" in the prefrontal cortex (PFC), the representation of the logic of a likely successful course of thought and action in terms of which cortical pathways are needed. The appropriate rulemap can be activated in a given situation. This sets up top-down signals that feed back to most of the rest of the cortex, dynamically establishing those pathways.
All Science Journal Classification (ASJC) codes
- Experimental and Cognitive Psychology
- Cognitive Neuroscience
- Behavioral Neuroscience