Temporal Dynamics of Reward Dysfunction in Psychiatric Disorders: Memory-Derived Value, Incentive Salience, Hedonic Impact, and Value Updating.

Reward processing unfolds across time: Prior experiences shape cue valuation and motivation, anticipated rewards bias action selection, reward encounters generate hedonic and evaluative signals, and outcomes update future value representations. Existing accounts have often examined these processes separately. Here, we propose a temporally organized and mechanistically constrained framework for reward dysfunction across psychiatric disorders. The revised framework distinguishes episodic reward memory, cue-outcome learning, stored value representations, model-based prospective simulation, model-free habit learning, incentive salience, anticipatory pleasure, hedonic impact, outcome evaluation, effort valuation, and value updating. We argue that psychiatric disorders are not characterized best by single broken links in a reward cycle but rather by partially overlapping profiles of altered coupling among these processes. Major depressive disorder may involve impaired positive value updating and reduced access to positive reward memories; schizophrenia may involve unstable value representations, impaired effort-cost integration, and reduced model-based use of reward information; bipolar disorder may involve state-dependent reward hypersensitivity and regulatory instability; and addiction may involve cue-triggered incentive salience, habitization, and negative reinforcement. We specify how these profiles can be tested using reinforcement learning, effort-based decision making, memory-guided choice, ecological momentary assessment, and longitudinal paradigms. This framework generates falsifiable predictions and mechanistically informed intervention hypotheses while avoiding one-to-one mappings between diagnoses, symptoms, and neural circuits.