NEUROMODULATORY CONTRIBUTIONS TO FLEXIBLE DECISION-MAKING REVEALED IN A STANDARDIZED BEHAVIORAL TASK

Decision-making is a dynamic process that evolves according to the animal’s learning and internal physiological states. Recent electrophysiological data from the International Brain Laboratory (IBL) revealed a widespread network of brain regions in proficient mice performing a visual two-alterative forced choice task, providing a high-resolution snapshot of the decision process. Here, we complement this static picture by correlating the activity of neuromodulatory (NM) nuclei, compact neural populations delivering specialized signaling molecules across the brain network, with flexible decision-making behavior. Some of these molecules like dopamine, serotonin, norepinephrine, and acetylcholine, play a prominent role in flexible decision-making by signaling reward prediction, uncertainty, attention, and behavioral arousal. However, their individual contributions remain unclear, as they are typically studied in isolation across diverse decision-making protocols. We built upon the standardized, reproducible behavioral framework of the IBL, using fiber photometry to record bulk calcium activity via GCaMP6f from genetically-targeted NM populations in five prominent nuclei across ~10⁶ trials in >50 mice. We find unique and overlapping signaling patterns during three distinct sub-trial events: pre-trial waiting, stimulus onset, and feedback delivery. We further correlate variability in this signaling to psychometric parameters and behavioral co-variates to provide a comprehensive picture of the neuromodulatory signals underlying shifts in decision-making. We present these results alongside an open-access dataset and an open-source Python toolbox for pre-processing and quality control. Our work aims to disentangle the specific contributions of these NM systems to decision-making, providing a better understanding of how they act in concert to modulate brain-wide decision processes.