Neuronal Reward and Decision Signals: From Theories to Data.
Level 5 - mechanism / opinion, no new human data
Narrative review of neurophysiological mechanisms and decision theory without systematic search or original human trial data.
PubMed 26109341 · doi:10.1152/physrev.00023.2014
What was done
This review synthesizes theoretical frameworks from reinforcement learning and economic choice theory with empirical neurophysiological data. It examines how constructs such as reward prediction errors, subjective value, and utility are physically represented across dopamine neurons, the striatum, the amygdala, and the frontal cortex.
What was found
The abstract reports no quantitative metrics or pooled statistical analyses. Qualitatively, it outlines that dopamine neurons broadcast global reward prediction error and utility prediction error signals. Specific neuronal populations in the striatum, amygdala, and frontal cortex encode distinct decision variables, including object value, action value, difference value, and chosen value, incorporating parameters such as risk, delay, effort, and social factors.
Why it matters
The paper demonstrates that theoretical constructs from economics and computation correspond directly to measurable neuronal implementations in the brain. This grounds abstract models of subjective value and decision-making in concrete neurobiology.
Limits
The abstract provides a purely narrative synthesis without systematic review methodology, meta-analytic data, sample sizes, or quantitative effect estimates. It relies heavily on animal models and mechanistic reasoning rather than direct clinical trial evidence.
Cited by
- supports Dopamine is essential in both human and animal brains for effort-reward decision-making and for reinforcement learning from prior actions.