Towards a neurocomputational and dynamical understanding of rumination

Abstract

Ruminative brooding, the maladaptive repetitive dwelling on abstract, self-referential thoughts, is a transdiagnostic feature of many psychiatric disorders, yet its neurocognitive mechanisms remain poorly understood. Brooding is both repetitive and perseverative, implicating deficits in cognitive control and maladaptive emotion regulation. We investigated these properties using experimental psychology, computational modeling, network neuroscience, and dynamical systems approaches. First, we reported brooding-associated inhibitory control deficits that were exacerbated by emotional cues, as reflected by performance on Stroop interference tasks. Numerically fitting parallel distributed processing models to participant Stroop data revealed a set of brooding-associated parameters that reflected network sensitivity to changing task demands and activity persistence of emotional processes, but not cognitive control processes. Next, we analyzed relationships between self-regulatory processes and brooding, and their corresponding EEG functional connectivity patterns. We found evidence of internal resistance to emotions and thoughts in brooding. Furthermore, functional connectivity patterns indicated that brooding and emotion dysregulation co-varied with aberrant top-down modulation from prefrontal regions to emotion and interoceptive systems, potentially reflecting compensatory regulation of internal emotional states. Finally, we proposed that brooding is an emergent property of an attractor state within the brain’s default mode network (DMN). Here, we hypothesised that brooding-related attractor dynamics would produce stable, recurrent neural dynamics that resist perturbation. Supporting this, DMN regions demonstrated persistent activity when switching from a cued rumination to a working memory task, consistent with resistance of ruminative thought to perturbation. In addition, nonlinear analyses of the recurrence properties of EEG demonstrated a positive association between brooding and neural recurrence, further supporting this attractor state hypothesis. Together, these studies advance a mechanistic understanding of brooding from neurocomputational, network, and dynamical perspectives.