RECORD: A high-throughput system for complex naturalistic decision-making in rodents

Abstract

Translational studies benefit from experimental designs where laboratory organisms use human-relevant behaviors. One such behavior is decision-making, however studying complex decision-making in rodents poses many problems. We overcame these limitations by designing a fully automated, inexpensive, high-throughput framework to study decision-making across multiple levels of rewards and costs in rodents: the REward-COst in Rodent Decision-making (RECORD) system. RECORD integrates three components: 1) 3D-printed configurable arenas, 2) custom electronic hardware, and 3) software for databasing, analysis, and computational modeling. RECORD sequentially offers subjects single cost/reward trade-offs that they can approach or avoid. We validated four different behavioral protocols, highlighting the versatility of our system. RECORD enables the execution of a decision-making task battery without requiring food or water deprivation, thus allowing us to characterize a profound and systematic impact of food deprivation on decision-making. We also demonstrate how RECORD enables analysis of heterogeneity in decision-making both within and across individuals and show, for the first time, that this behavioral variability is quantifiably constrained. Using oxycodone self-administration as a test case, we also reveal how analytic approaches that incorporate behavioral heterogeneity are more sensitive to detecting perturbations in decision-making than conventional approaches that average behavior across all individuals. Overall, RECORD is a more powerful and sensitive approach to studying decision-making in animals, with features that facilitate translational studies of DM in psychiatric disorders.

Model of a striatal circuit exploring biological mechanisms underlying decision-making during normal and disordered states

“…between offers which each have benefits and costs). Using this heuristic, we can infer the decision-spaces of behaving rodents 148 and humans 149 by regressing sSPN activity on experimental parameters (for example, temperature or music volume), as we demonstrate using synthetic data (Figures S1Q,R, Inferring decision-space from SPN activity and choice, STAR Methods).…”