Keynotes | K&Ts | GACs | Talks | Posters | Search

Poster E in Poster Session E: Thursday, August 6, 10:30 am – 12:15 pm, Kimmel Center, Shorin & Rosenthal Rooms

Perceptual Consciousness Facilitates Model-based Decision Making

Chang Christina He¹, Benjy Barnett², Ross Otto³, Stephen M. Fleming¹; ¹University College London, ²Birkbeck, University of London, ³McGill University

Presenter: Chang Christina He

Flexible behaviour requires prospective evaluations of actions and their potential outcomes. On some theoretical accounts, perceptual consciousness evolved to support such flexibility by enabling the integration of perceptual information during action planning. In parallel, computational research in reinforcement learning distinguishes between two decision making strategies: model-based control that relies on internal representations of environmental structure, and model-free control that depends on stimulus–reward associations. Although these frameworks have been widely used to characterise decision making, the role of consciousness in enabling model-based control remains untested. We tested this using a modified two-step reinforcement learning paradigm in which cues signalling the transition structure were presented either above or near perceptual threshold. We found that model-based patterns of decision making emerged when the structural cue was consciously accessible, whereas behaviour was more consistent with model-free strategies when the cue was near perceptual threshold. This suggests that conscious perception of environmental structure facilitates model-based decision making, supporting functional accounts that link perceptual consciousness to flexible, goal-directed control of behaviour.

Topic Area: Decision-Making, Cognitive Control & Event Cognition