Human insula activation reflects risk prediction errors as well as risk

Area of Science:

• Neuroscience
• Cognitive Science
• Decision Making

Background:

• Reinforcement learning models and primate physiology show the brain encodes reward prediction error signals.
• Organisms must predict and update risk associated with reward forecasts for choice guidance and learning modulation.
• The neural mechanisms of risk prediction remain largely unknown.

Purpose of the Study:

• To investigate whether and how the brain accomplishes risk prediction.
• To identify neural correlates of risk prediction error.
• To understand the brain's mechanisms for updating risk assessments.

Main Methods:

• Functional magnetic resonance imaging (fMRI) was used during a gambling task with dynamic risk levels.
• Participants performed a task where risk varied constantly.
• Analysis focused on brain activation patterns correlating with risk prediction and risk prediction error.

Main Results:

• Early-onset activation in the human insula significantly correlated with risk prediction error.
• The temporal dynamics of insula activation suggest a role in rapid risk updating.
• Delayed activation, linked to general uncertainty, is consistent with risk prediction.
• Neural activity for risk prediction and error mirrors that of reward prediction and error.

Conclusions:

• The human brain actively predicts risk associated with uncertain rewards.
• The insula plays a crucial role in processing risk prediction errors and updating risk assessments.
• Neural mechanisms for reward anticipation under uncertainty must encompass both reward and risk prediction.