Open-BEATBox: an open-source autonomous operant chamber for high-throughput data acquisition of complex behaviour

Abstract

Behavioural neuroscience increasingly relies on rodent models to investigate the neural mechanisms underlying cognition in health and disease. However, conventional operant testing is typically restricted to short daily sessions that require frequent experimenter intervention and often rely on food or water restriction. These constraints limit behavioural sampling, disrupt natural activity patterns, and introduce variability associated with handling and experimental schedules.Here, we present the Open Platform for Experimental Neuroscience Behavioural and Autonomous Box (Open-BEATBox), an open-source, autonomous home-cage platform that enables continuous, self-paced assessment of complex cognitive behaviour in mice. By allowing animals to live in the testing environment throughout the experiment, Open-BEATBox supports longitudinal behavioural assessment across different cognitive tasks while minimizing experimenter intervention and preserving natural circadian activity. The platform is modular, customizable, and designed to accommodate a broad range of operant paradigms and integration with chronic neural recording and neuromodulation technologies.Using a visual reversal-learning task, we show that continuous autonomous testing generates approximately ten times more completed trials during reversal learning than conventional session-based protocols while reducing experimenter workload without evidence of major adverse effects on the welfare measures assessed. The flexibility of the platform further enables implementation of additional cognitive paradigms, including uncertainty-based decision-making tasks.Together, Open-BEATBox provides an accessible open platform for longitudinal cognitive phenotyping, facilitating the collection of large longitudinal behavioural datasets and creating new opportunities to combine continuous behavioural assessment with modern systems neuroscience approaches.