Adult Zebrafish Engage in Path Integration-like Behavior When Exploring a Novel Environment
Sanghera, B.; Shraybman, P.; Kanwal, J. S.
Show abstract
Spatial navigation relies on the integration of self-motion (idiothetic) and external (allothetic) cues to construct an internal representation of space. While path integration has been extensively studied in insects and mammals, its presence and mechanistic basis in non-mammalian vertebrates remain poorly understood. Here, we show that adult zebrafish (Danio rerio) placed in a novel, visually impoverished environment exhibit systematic reorganization of exploratory behavior consistent with path integration-like navigation. Using markerless tracking, we identified a sequence of stereotyped bouts initiated by wall contact that consists of subsequent reorientations followed by a corrective sharp turn, returning the animal toward its initial contact location. Markov chain analyses reveal structured transitions linking wall-touch events, reorientation states, and high-curvature turns early in exploration, followed by a progressive decoupling of turning from boundary contact over time. Angular error-correction analyses demonstrate that cumulative reorientation predicts the magnitude and direction of subsequent sharp turns, consistent with vector-based error compensation. Multivariate analyses further show a graded shift in behavioral state space from early to late exploration, rather than an abrupt strategy switch. Together, these results provide evidence that zebrafish employ a go-and-touch strategy that integrates tactile boundary information with self-motion cues to estimate spatial layout. Our findings establish zebrafish as a vertebrate model for studying the computational and neurogenetic basis of path integration that develops by the fifth week post-fertilization.
Matching journals
The top 3 journals account for 50% of the predicted probability mass.
Similar papers in this journal
Similar papers in this journal
"Similar papers" are the closest papers from that journal in the model's embedding space. They show what the match is built on, but the ranking comes mostly from a classifier over the whole training set, not from these examples alone.