0065 A Novel Automated Task for Spatial Learning in Rodents

Abstract

Abstract Introduction Spatial learning in rodents is useful for addressing a variety of research questions. We set out to create a fully automated spatial learning task that required rodents to make a sequence of cued navigational decisions. We used this maze to investigate the impact of sleep on insight gain, a sudden rather than gradual acquisition of a skill or behavior. Methods We designed a circular four-quadrant elevated maze. Each quadrant had three choice points on a branched arm, consisting of a reward well and cue light. To discourage rats from backtracking, motorized doors were placed between quadrants. Rats were trained to follow the light for a water reward. Their position was tracked by custom MATLAB code processing live video from a ceiling-mounted camera; the code also controlled all components of the maze, calculated and presented novel routes through the maze, analyzed results in real time, and maintained records of each rat across sessions. After an average of five days rats reached criterion in pretraining (following lights in pseudorandom positions) and were taught a hidden rule: the direction cued on the second quadrant was the correct direction on the fourth (uncued) quadrant. Rats completed one session each in the morning and afternoon, separated by 3h of sleep opportunity or sleep disruption, achieved by gentle handling. This was repeated after a night of sleep opportunity or sleep disruption for a total of four sessions. Results Rats quickly learned and consistently chose the correct well, remaining motivated for 96 consecutive trials. Rats were said to have gained insight if they demonstrated significant learning of the hidden rule. Although no animals have yet met this criterion, there was a trend of sleep-disrupted rats performing worse, suggesting that sleep plays an important role in consolidating the spatial rules of the task. Conclusion The versatility of this maze can accommodate many tasks not limited to insight gain specifically. Delivering rewards consistently at multiple distinct locations is a valuable technique for investigating spatial learning, navigation, and decision making, making this maze a powerful tool for investigating interventions for disease models that implicate the hippocampus or striatum. Support (if any)