Reinforcement learning and working memory in mood disorders: A computational analysis in a developmental...

Area of Science:

• Neuroscience
• Computational Psychiatry
• Developmental Psychology

Background:

• Mood disorders frequently emerge in adolescence and young adulthood.
• These disorders are linked to abnormalities in reinforcement learning.
• The distinction between acute symptom severity and lifetime diagnosis in these abnormalities is unclear.

Purpose of the Study:

• To investigate reinforcement learning and working memory abnormalities in adolescents and young adults with mood disorders.
• To differentiate between state-like (current symptoms) and trait-like (lifetime diagnosis) abnormalities.
• To examine the impact of working memory load on reward processing.

Main Methods:

• A computational model of reinforcement learning and working memory was applied to an instrumental learning task.
• The study included 220 participants: adolescents and young adults with unipolar disorders, bipolar disorders, or no psychopathology.
• Model parameters were analyzed in relation to diagnoses and current symptom severity.

Main Results:

• Higher current manic or anhedonic symptoms correlated with poorer task performance.
• Lower reward learning rates were observed in participants with higher anhedonia or lifetime unipolar/bipolar disorders.
• Increased manic symptoms were associated with faster working memory decay and reduced working memory utilization.

Conclusions:

• Reinforcement learning processes are abnormal in mood disorders, correlating with either current symptom severity or lifetime diagnoses.
• These findings suggest distinct computational profiles for state- and trait-like aspects of mood disorders.
• Results highlight the role of reward processing anomalies in the pathophysiology of mood disorders.