A neural code supporting prospective probabilistic reasoning for instrumental information demand in humans

Area of Science:

• Neuroscience
• Cognitive Science
• Decision Science

Background:

• Active decision-making involves seeking information, but mechanisms for estimating information gains are poorly understood.
• The brain must prospectively estimate the value of information from different sources.

Purpose of the Study:

• To investigate how the brain estimates prospective information gains.
• To identify neural mechanisms underlying the valuation of information during active decision-making.

Main Methods:

• Functional magnetic resonance imaging (fMRI) was used to observe brain activity.
• Participants bid for information in a task with varied rewards, uncertainty, and diagnosticity.
• Multivoxel pattern analysis (MVPA) decoded neural representations.

Main Results:

• Bids reflected sensitivity to prior certainty and expected posterior certainty, aligning with information value theory.
• Expected posterior certainty was decoded from activity in posterior parietal and extrastriate cortices.
• Neural representations of expected posterior certainty were independent of instrumental rewards.

Conclusions:

• The posterior parietal and extrastriate cortices are implicated in anticipating information gains.
• These regions may mediate the prospective estimation of posterior probabilities for active information seeking.
• Understanding these mechanisms is key to adaptive decision-making.