State-space modeling uncovers brain-behavior dynamics of inhibitory control.

Abstract

Inhibitory control, the ability to suppress prepotent but inappropriate action, thought and emotion, plays a crucial role in goal-directed behavior. Neuroimaging studies suggest that this process engages dynamic interactions across canonical brain networks. However, the precise dynamic brain mechanisms underlying inhibitory control and their relationship to individual differences remain poorly understood. To address this gap, we applied a novel state-space model to a large-scale fMRI dataset to investigate latent brain states and their dynamics during a gender-Stroop task. Our study revealed four distinct latent brain states, each characterized by unique functional connectivity patterns. Temporal properties of latent brain states, such as occupancy rates (OR) and mean lifetimes (ML), were significant predictors of the Stroop effect. Specifically, the ML of state S1 showed negative association with the Stroop effect, suggesting that prolonged engagement in this state facilitated inhibitory control. State-specific connectivity patterns also predicted Stroop effects. Compared to S2, S1 showed stronger within- and between-network connectivity. These findings suggest that S1 represents an optimal brain state for resolving conflicts and promoting inhibitory control. Together, our findings shed light on the neural dynamics involved in overcoming sensory-motor conflicts triggered by automatic responses and highlight their potential implications for cognitive interventions.