The Temporal Dynamics of Self-Control

Self-control—the ability to pursue long-term goals over short-term temptations—is a critical faculty of human cognition, but the cognitive processes enabling self-control are not well understood. Traditional models have focused on impulse inhibition: effortfully inhibiting prepotent motor responses towards a temptation, yielding a stage-based evolution of choice. Other models emphasize dynamic competition between goal and temptation, yielding a more integrative evolution of choice. Although these models represent fundamentally different conceptions of self-control, current methods are inadequate for investigating real-time dynamics, leaving the question of which model best describes self-control unresolved. We investigate these models using mouse-tracking: a dynamic, real-time measure of decision-making in which we measure participants’ computer mouse movements as they navigate tradeoffs between immediate and delayed gratification (e.g., $5 today vs. $20 in 3 months). We develop a novel quantitative approach that integrates the rich spatial and temporal information contained in mouse trajectories, and find evidence for both impulse inhibition and dynamic competition. Notably, impulse inhibition is less frequent, occurring in only one-quarter of choices favoring larger later rewards over smaller sooner ones. We further find substantial individual variability on who relies on impulse inhibition, with more present-biased individuals more likely to use impulse inhibition to choose larger-later options. Finally, our approach reveals the diverse variability within impulse inhibition and dynamic competition, and accounting for this variability greatly strengthened models predicting out-of-sample choices. Our findings clarify the mechanisms underlying self-control and introduce a robust tool for quantifying real-time decision-making dynamics.