Anhedonic Traits Do Not Impair Performance in a 3-Arm Bandit Task.

Area of Science:

• Cognitive Neuroscience
• Computational Psychiatry

Background:

• Anhedonia, a core symptom of diminished reward sensitivity, is often associated with impaired reinforcement learning (RL).
• Standard RL tasks can be demanding and may conflate valuation with other cognitive processes.

Purpose of the Study:

• To adapt and validate a three-arm bandit (3AB) task for assessing RL in anhedonia.
• To investigate group differences in reward and punishment learning, and sensitivity between anhedonic and non-anhedonic individuals.

Main Methods:

• An online sample (N=206) pre-screened for anhedonia completed a 3AB task.
• Hierarchical Bayesian modeling was employed to estimate RL parameters.
• Model-agnostic strategies (win-stay/lose-shift) were also analyzed.

Main Results:

• No credible group differences were found in reward learning rate, punishment learning rate, reward sensitivity, or punishment sensitivity (Bayes factors favored the null).
• Win-stay/lose-shift strategies also showed no significant group differences.
• The computational model demonstrated good predictive accuracy, but parameter estimates did not differentiate groups.

Conclusions:

• Anhedonia, in this 3AB task, is not reliably linked to core RL parameter differences or simple choice strategies.
• The adapted 3AB task provides a robust measure of RL, with good model performance and parameter recovery.
• Further research may explore other factors contributing to anhedonia's heterogeneity.