Hierarchical learning induces two simultaneous, but separable, prediction errors in human basal ganglia.

Area of Science:

• Neuroscience
• Cognitive Science
• Computational Neuroscience

Background:

• Dopaminergic neurons are traditionally thought to signal a single, global reward prediction error.
• Complex, real-life tasks often involve hierarchical structures and multiple, simultaneous prediction errors.

Purpose of the Study:

• To investigate if dopaminergic neurons can signal multiple, simultaneous reward prediction errors in humans.
• To examine the neural basis of learning in hierarchical tasks using functional neuroimaging.

Main Methods:

• Functional neuroimaging (fMRI) was employed to measure brain activity.
• Participants performed a hierarchical task with simultaneous, uncorrelated prediction errors.
• Analysis focused on signals in the ventral striatum and ventral tegmental area.

Main Results:

• Two distinct and simultaneous reward prediction error signals were identified.
• These signals corresponded to the different hierarchical levels of the task.
• Separable signals were observed in the ventral striatum and ventral tegmental area.

Conclusions:

• Dopaminergic neurons can encode multiple, separable reward prediction errors.
• Hierarchical task learning involves distinct neural signals for different levels.
• Simultaneous learning across many task levels may be constrained by neural processing limitations.