Related Experiment Video
Updated: Mar 21, 2026
Identification and Classification of Position-specific GABAA Receptor Subunit Missense Variants for Their Role In Hippocampal Pyramidal Neurons
Published on: June 6, 2025
Adaptive Prediction Error Coding in the Human Midbrain and Striatum Facilitates Behavioral Adaptation and Learning
Kelly M J Diederen1, Tom Spencer2, Martin D Vestergaard3
1Department of Physiology, Development, and Neuroscience, University of Cambridge, Downing place, Cambridge CB2 3DY, United Kingdom; Department of Psychiatry, University of Cambridge, Robinson Way, Cambridge CB2 0SZ, United Kingdom.
This study shows that the brain adapts how it processes reward prediction errors based on reward variability. This neural adaptation, observed in the Substantia Nigra/Ventral Tegmental Area (SN/VTA) and ventral striatum, improves learning and task performance.
Area of Science:
- Neuroscience
- Cognitive Science
- Computational Psychiatry
Background:
- Effective learning relies on adjusting prediction errors to reward variability.
- Neurons may code prediction errors relative to the standard deviation (SD) of reward distributions for behavioral adaptation.
Purpose of the Study:
- To investigate if neural coding of prediction errors adapts to reward variability.
- To examine the relationship between neural adaptation, learning rates, and task performance.
Main Methods:
- Participants predicted reward magnitudes from distributions with varying SDs.
- Brain activity (BOLD response) in the Substantia Nigra/Ventral Tegmental Area (SN/VTA) and ventral striatum was measured.
- Behavioral learning rates were analyzed in relation to prediction errors and reward SDs.
Main Results:
- BOLD response slopes in SN/VTA and ventral striatum were steeper for prediction errors in smaller SD distributions, indicating adaptive coding.
- SN/VTA adaptation developed over trials, not instantaneously.
- SD-dependent changes in learning rate paralleled neural adaptation.
- Enhanced SN/VTA and ventral striatal adaptation correlated with improved task performance.
Conclusions:
- Adaptive coding of prediction errors in the SN/VTA and ventral striatum facilitates behavioral adaptation.
- This neural mechanism supports more efficient learning in variable reward environments.
