Risk prediction error signaling: A two-component response?

Area of Science:

• Neuroscience
• Cognitive Science
• Computational Neuroscience

Background:

• Organisms adapt to uncertain environments using reward-based learning, primarily mediated by the dopaminergic system signaling reward prediction errors.
• The dopamine response to rewards may involve two components: rapid stimulus detection and a later reward prediction error signal.
• Error-based learning signals are also implicated in risk processing, but the neural mechanisms remain unclear.

Purpose of the Study:

• To investigate if risk learning, or surprise-based learning under uncertainty, utilizes a two-component temporal coding scheme similar to the dopaminergic system.
• To identify the neural correlates and brain regions involved in risk prediction error signaling.

Main Methods:

• Human high-density electroencephalography (EEG) was employed during a card game task.
• Analysis focused on identifying distinct temporal components in brain activity related to risk and salience.
• Source localization techniques were used to pinpoint the origin of neural signals.

Main Results:

• Risk prediction error is reflected in the amplitude of the P3b component of the EEG signal.
• An earlier neural component, modulated by stimulus salience, precedes the P3b modulation.
• Source analyses suggest that both salience and risk prediction error signals originate in the insular, frontal, and temporal cortex.
• P3b amplitude modulation correlates with pupil size, suggesting a link to locus coeruleus activity.

Conclusions:

• Risk learning, like reward learning, appears to involve temporally dissociable neural signals.
• The findings support a two-component model for surprise-based learning under uncertainty.
• The insular, frontal, and temporal cortex play a key role in processing risk prediction errors.
• Pupil size may serve as an indirect indicator of locus coeruleus activity during risk processing.