DISTINCT BEHAVIOURAL STATES DURING TACTILE DISCRIMINATION IN MICE REVEALED BY A MODULAR AUTOMATED OPEN-SOURCE MAZE FOR SENSORY-GUIDED FORAGING

Rodents have sophisticated capacities for context-dependent decision-making and for abstraction of sequential rules. We developed an experimental maze to allow mice to express these capacities. The maze includes automated stimulus presentation and reward delivery triggered by mouse tracking. This approach enables study of foraging behaviour without restricting mouse movement or access to nutrition, and provides robust experimental control while ensuring flexible design of tasks and open protocol sharing through the use of accessible components. To allow flexible reconfiguration of the maze and set up arbitrary associations between tactile stimuli, locations and rewards, we replaced standard wall panels with devices for reward dispensing or stimulus delivery. Device motion was triggered in real time by animal entry into regions of interest (ROIs). An animal could encounter multiple stimuli as it moved from the maze’s origin to any endpoint, and could shuttle freely between cage and maze. We found that mice habituated to the maze within minutes and were intrinsically motivated to explore it with no need for fluid or food restriction. Mice then learned to navigate to arbitrary locations for rewards. We provide examples of shifts between engaged (exploratory or goal-directed) and settled/nest-building behavioural states. Trials ended when the mouse chose to leave the maze, and mice shifted between behavioural states during the course of a single trial, often triggered by collecting a food reward. Our results demonstrate the complexity of state modulation during naturalistic exploration and suggest the need for fuller characterisation of the state-dependent modulation of tactile processing.