Schultz · Dialogues in clinical neuroscience 2016 · narrative review · n=?

Dopamine reward prediction error coding.

Level 5 - mechanism / opinion, no new human data

Narrative review synthesizing mechanistic animal and human neuroscience without original data or systematic methodology

PubMed 27069377 · doi:10.31887/DCNS.2016.18.1/wschultz · record verified 2026-08-26

What was done

This review synthesizes neurophysiological and neuroimaging findings from rodents, monkeys, and humans on how midbrain dopamine neurons and interconnected structures (striatum, amygdala, and frontal cortex) encode reward prediction errors.

What was found

The abstract reports no quantitative values or effect sizes. It describes neurobiological signaling characteristics: midbrain dopamine neurons increase firing during positive prediction errors (reward greater than predicted), maintain baseline activity for fully predicted rewards, and decrease firing during negative prediction errors (reward less than predicted), nonlinearly coding formal economic utility.

Why it matters

It outlines the neuronal implementation of reward prediction errors, providing a mechanistic link between reinforcement learning models, economic utility, and addiction pathology.

Limits

The abstract lacks systematic search criteria, quality appraisal of included studies, and quantitative data. Findings rely largely on mechanistic animal neurophysiology alongside human neuroimaging.

Cited by