The Neuro-Computational Architecture of Value-Based Selection in the Human Brain

Area of Science:

• Neuroscience
• Cognitive Science
• Computational Psychiatry

Background:

• Value-based decision-making models typically involve separate valuation and selection stages.
• The precise neural mechanisms underlying the selection stage remain poorly understood.

Purpose of the Study:

• To elucidate the neuro-computational architecture of the selection stage in value-based decision-making.
• To identify distinct brain regions and computations involved in choice selection.

Main Methods:

• Utilized drift-diffusion models (DDMs) coupled with model-based functional magnetic resonance imaging (mb-fMRI).
• Employed effective connectivity and multivariate pattern analysis (MVPA) to analyze neural data.
• Investigated the integration and readout computations within DDMs.

Main Results:

• Identified distinct neural substrates for integration and choice readout computations at the selection stage.
• Dorsolateral prefrontal cortex (dlPFC) integrates decision value signals originating from the ventromedial prefrontal cortex (vmPFC).
• Posterior parietal cortex (PPC) is involved in the readout of choice outcomes.

Conclusions:

• A prefronto-parietal network implements behavioral selection via a distributed drift-diffusion process.
• Distinct brain regions contribute specific computational roles to the selection stage of decision-making.