Linking Regulation Strategies to Emotion Dynamics: A Bayesian Ornstein-Uhlenbeck Approach

Google map

Many phenomena in the behavioral and health sciences — emotion dynamics, stress responses, physiological processes, cognitive performance — behave as continuous, self-regulating processes that fluctuate in response to perturbation but tend to return toward equilibrium. The Ornstein-Uhlenbeck (OU) process can characterize a mean-reverting time series via three parameters: homebase, regulation rate, and volatility, while handling irregular time intervals typical of intensive longitudinal data (e.g., ecological momentary assessment). We present a hierarchical Bayesian OU model (BOUM) with person-level covariates to explore links between emotion regulation strategies and emotional dynamics. We first examined the performance of the model via a simulation study. Then we applied BOUM to a substantive question: can habitual emotion regulation strategies explain individual differences in how emotions unfold in daily life? We analyzed 14 days of EMA-measured valence (N = 164) from the MindTrack study. Results show that frequent use of cognitive reappraisal for emotion regulation was associated with a more positive homebase. Moreover, problem-solving with lower volatility, and rumination with both a more negative homebase and greater volatility. These findings point toward potential strategy-specific intervention targets for psychopathologies characterized by mood-related symptoms. More broadly, the framework generalizes to any mean-reverting process and could serve as a tool for identifying targeted intervention points across diverse research domains.