49.4 Data-Driven Methodology to Describe Behavioral Differences in SCN2A Haploinsufficient Rats

Animal behavior provides the foundation upon which we interpret neurophysiology and model disease. However, we remain quite limited in our understanding of the complexities of animal behavior. When studying learning and memory, it is common to interpret behavior based upon simple assumptions, often based off the algorithm that governs the rules of the task. With spatial alternation behavior—a behavior where, to get a reward, an animal must alternate between locations as a function of past actions—differences in performance on the task are often interpreted as uniquely exposing differences in memory of the animals. Using Reinforcement Learning agents, we have shown that this assumption does not account for the way rats learn a spatial alternation task. Rather, variability in learning is consistent with the rats utilizing a variety of spatial preferences in addition to memory. However, this enhanced understanding of the task leaves us in a challenging place if we want to use spatial alternation behavior to expose differences between groups of animals.