Cocaine-exposed rats utilize learned value functions to accumulate evidence quicker than controls

Abstract

Included in this folder are all data files (.csv) file associated with this manuscript that was submitted to Psychopharmacology.

Rationale: Maladaptive decision-making is a hallmark feature of substance use disorders, with impairments in various decision-making components and related neural circuitry persisting even after active drug use has ceased. Though this general fact has long been appreciated with decades of support from clinical and preclinical work alike, behavioral studies have sometimes reported contradicting effects; while prior drug use is typically associated with disrupted decision-making, there is some evidence of paradoxical enhancements. Objectives: We combine behavioral data from four of our previous rat studies assessing the effect of prior cocaine use on performance in a reward guided decision-making task in order to zoom in on trial-by-trial differences in decision-making as rats respond to changed contingencies and examine mechanisms of choice via drift diffusion modeling. Methods: Rats underwent a 12-day protocol for self-administration of either sucrose pellets (control) or cocaine. Following a one month withdrawal period, rats were assessed on a reward- guided decision-making task where reward was manipulated by varying either the delay to receive or the size of a liquid sucrose reward. Results: Cocaine-exposed rats adapted to changed contingencies more rapidly than controls—exhibiting stronger biases for the higher value reward by the end of trial blocks—which was mirrored by an increased rate of evidence accumulation in our computational model. Conclusions: We suggest this apparent behavioral improvement may reflect an overestimation of task volatility. While such an enhancement was advantageous in our deterministic task, this could explain why previous work using tasks that incorporate uncertainty often report deficits following cocaine use.