Divergent spatiotemporal integration of whole-field visual motion in medaka and zebrafish.

Animals in different ecological niches have evolved different sensory capacities to detect behaviorally relevant sensory signals. How behavioral algorithms and neural networks adapt to environmental demands remains unknown. We compared spatiotemporal visual motion processing in larval zebrafish (Danio rerio) and medaka (Oryzias latipes) using whole-field motion stimuli. We find that medaka integrated motion over visual fields twice as large as zebrafish and showed greater sensitivity to peripheral visual information. Temporally, zebrafish respond robustly to stimuli with lifetimes as short as 100 ms, while medaka require lifetimes exceeding one second and retain motion information for several seconds after stimulus offset. These different time constants suggest that medaka prioritize object classification and persistence detection, such as conspecifics in structured social groups, while zebrafish specialize in rapid background motion detection for navigation in variable flow environments. Our findings demonstrate how neural computation adapts to species-specific ecological and social demands.