Dynamical modeling of individual sensory reactivity and habituation learning

Empirical research on sensory processing often focuses on group averages to uncover how sensory input translates into behavior. While this population-level approach has revealed important general principles of genetic and environmental control of behavior, it masks the rich and potentially meaningful variation observed at the level of individuals. Key features of sensory processing are reactivity, the degree to which a stimulus elicits a response, and habituation, a fundamental learning process that serves as a cognitive filter by reducing the response to repeated, irrelevant stimuli. We developed a model-based approach to quantify individual sensory processing dynamics, using a visual escape paradigm in <i>Drosophila</i>. We collected light-off jump responses in more than two hundred flies and quantified them individually via Bayesian inference of a dynamical model's parameters. Considering both reactivity and habituation, we found that quantitative properties of individual responses vary greatly even in an isogenic population raised under identical environmental conditions. By moving from population averages to individual-based analyses, we uncover and quantify pronounced stochasticity in individual responses. Furthermore, we found that individual reactivity and habituation parameters display substantial temporal stability over a two-week interval, indicating that these behaviors are stable, intrinsic properties of an individual. The presented framework enables robust stratification of individual behavioral phenotypes and provides a quantitative platform for modeling interindividual differences in sensory processing, including frequently observed atypical responses in neurogenetic diseases.