Pierre Le Denmat (KU Leuven) – How do I know that I know ? Towards a mechanistic understanding of confidence.
(with Tom Verguts)
When making a decision, we are able to precisely estimate how confident we are in this decision, but how did we learn this self-evaluation skill in the first place ? Existing research on confidence has mostly focused on its nature, leaving the mechanisms of confidence almost uncharted territory. Using a modeling approach inspired by reinforcement learning, I investigate how humans learn to accurately self-evaluate their decisions, and what consequences these learning processes have for the way humans compute their level of confidence. Predictions from the developed model are tested using both behavioral experimentation and electrophysiological recording.