TEMPORAL DYNAMICS OF SELF-INITIATED VOLUNTARY ACTIONS IN MICE

To satisfy internal needs, individuals must interact with their environment to obtain specific outcomes. Goal-directed (GD) behaviors depend on action-outcome (A-O) associations and outcome value updating. While extensive work has focused on which action to select, when to act and how fast to act remain unresolved. This raises key questions: how do environmental parameters shape the internal pace of GD actions, and what are the neuronal circuits underlying such behavior? In a self-paced operant task, mice perform nose pokes to obtain food rewards in absence of external cues. We compared contingency degradation, outcome devaluation and partial extinction procedures and measured sensitivity to A–O contingencies and outcome value. To probe action initiation dynamics, we analyzed response rates and timing distributions across task time epochs (e.g., retrieval time, inter-response intervals, poke duration).We also fitted drift-diffusion models (DDM) to quantify underlying decision parameters including evidence accumulation rate and decision threshold. Mice successfully showed behavioral sensitivity to experimental manipulations, confirming goal-directed control. Timing analyses revealed structured temporal patterns in action initiation that varied with motivational state and task history. Preliminary DDM fitting suggests that associative statistics can differentially modulate evidence accumulation rates and decision thresholds during action initiation. This work presents an integrated experimental and computational framework for dissecting the temporal dynamics of self-initiated goal-directed actions in mice. We further develop a normative theory linking DDM parameters to optimal associative learning.