hide
Free keywords:
-
Abstract:
Alterations in reward processing are associated with various mental disorders. However, it is an open question to what extent intra-personal fluctuations of reward processing are influenced by mood and homeostatic body states across days. Here, we investigate the link between those states and parameters of reward processing such as reward sensitivity and learning from reward outcomes. To this end, we developed a gamified version of a reinforcement learning paradigm (“Influenca”) to collect up to 31 runs of value-based decisions in an unstable environment over multiple days. In-game choices were complemented by ecological momentary assessment (EMA) using state items. As part of an ongoing study, we collected 82 runs (9 participants). Based on pooled data, we predicted reinforcement learning parameters derived from the game using a cross-validated elastic net including state items as predictors in an exploratory analysis. Across participants, we found a higher learning rate if state ratings of alertness and sadness were high (r =.35). In contrast, reward sensitivity was lower if state ratings of thirst and distracting thoughts were high and if snacking occurred after the preceding meal (r =.46). Notably, snacking occurred more often during negative mood states (p =.007). In summary, reward-learning parameters are predicted by momentary mood and homeostatic states, but more data is needed to individualize prediction models. To conclude, the game collects data from participants in various states and mental conditions. Hence, it may provide an innovative tool for longitudinal assessment and early identification of risk factors such as reward-related dysfunctions.
