AUDITORY THALAMUS NEURONAL DYNAMICS IN FREE EXPLORATION AND DECISION-MAKING

Auditory thalamus (medial geniculate body, MGB) dynamically encodes sensory and task-related information. A functional subset of MGB neurons shows ramping dynamics during reward-preceding delay periods in head-fixed Go/Nogo detection paradigms, suggesting a potential role of auditory thalamus in outcome prediction and short-term memory maintenance to guide the next action. We propose that MGB encodes several stages of self-paced, reward-driven decision-making such as the assessment of choices after presentation of sensory stimuli and the execution of decisions. To test this, we used a freely moving, two-choice, automated T-maze paradigm in which mice self-pace their behavior. Single cell neuronal activity in MGB was simultaneously tracked using a miniature microscope imaging approach. Upon task learning, we found that mice showed optimized goal trajectories and increased movement speed. To quantify neuronal activity during decision-making, we defined a decision-point approximation (i.e., allocentric angle changes in choice zone) and tested correlations of neuronal activity with behavioral variables during task execution. In naïve mice, MGB neurons exhibit sparse, adapting responses to initially neutral auditory cues in the maze, which transition to stable, outcome-specific stimulus-driven responses after cue-reward associations are learned in expert animals. Furthermore, MGB neurons display different response patterns across task parameters, including tone presentation, decision, and reward consumption. Besides fast-responding cue-sensitive neurons, we find cells with slow ramping activity during decision and reward periods. This data shows that MGB neurons dynamically encode learned-cues, decisions and reward-related information in a self-paced, complex, freely moving paradigm, further underlining the role of sensory thalamus in memory-related cognitive functions.