Should I study the cereal options in the supermarket again or pick this one? Should I commit to
buying this house now, or have another expert make an inspection? Many decisions, from the trivial
to the life changing, require a judgment as to when to stop gathering information and commit to a
decision. They require a balance between rushing into a decision (risking a mistake), and taking
longer than you need (wasting precious time). Being confident enough to decide depends on several
factors, including task difficulty, and the consequences of being right or wrong. If the consequences
of being wrong are particularly dire (e.g. buying a house with significant problems), you may gather
more information. Conversely, with a strict deadline, you might lower the bar and decide more
rapidly. It is unknown how this sense of being confident enough to decide is neurally computed. We
use a computational model to determine when is the optimal time to stop sampling and decide,
depending on task context. With behavioral experiments, we examine whether people decide
optimally, whether they flexibly change their criterion depending on context, and in what specific
aspects people are suboptimal. We also investigate the neural basis of this confidence computation
by recording neural activity from task-relevant brain regions. Finally, we use transcranial magnetic
stimulation to non-invasively perturb activity in these brain regions to assess their role in processing
confidence.