ZEBRA K+: A MODULAR AI-POWERED PLATFORM FOR HIGH-THROUGHPUT KINEMATIC ANALYSIS OF ACOUSTIC STARTLE RESPONSE PLASTICITY IN ZEBRAFISH EMBRYOS AND LARVAE

The acoustic startle response (ASR) and its modulation through short-term habituation and prepulse inhibition (PPI) are fundamental for investigating neural plasticity and sensorimotor gating, yet achieving reliable automated kinematic analysis across different life stages remains a significant technical challenge. This study presents Zebra K+, a modular high-throughput platform that integrates advanced machine learning tools for the precise kinematic analysis of ASR in zebrafish. While the module for adults has been previously established, we here validate the system’s application for embryos (5 days post-fertilization, dpf) and early larvae (6-7 dpf). The platform utilizes high-speed video recording (1000 fps) combined with automated image analysis based on DeepLabCut, leveraging artificial intelligence to achieve high-resolution quantification of movement kinematics that traditional software fails to capture accurately. Through pharmacological validation using ketamine, apomorphine, and haloperidol, we demonstrate that Zebra K+ effectively detects developmental changes in startle sensitivity and plasticity. Our results confirm the platform's ability to identify specific PPI deficits and track the maturation of dopaminergic signaling, notably capturing the reversal of apomorphine-induced PPI deficits by haloperidol. In conclusion, Zebra K+ provides a robust, scalable, and AI-driven tool for neurotoxicological screening and developmental neuropharmacology, facilitating the implementation of New Approach Methods (NAMs) by providing precise behavioral endpoints across the entire zebrafish life cycle.