Modeling Consciousness Using Cognitive Maps

Two types of neuro-phenomenology are distinguished: (1) actualist neuro-phenomenology, which studies isomorphisms between actual states of consciousness and patterns of activity in the brain which support them, and (2) possibilist neuro-phenomenology, which studies the relationship between possible conscious experiences and their neural correlates. I describe a method for studying both types of neuro-phenomenology using cognitive maps, which are state histories for a real or virtual agent. The points in a cognitive map simultaneously represent possible brains states and conscious states. I present initial work on a model of reinforcement learning that shows how certain claims the phenomenologist Edmund Husserl made can be linked to testable hypotheses about brain and behavior.