Quantification Platform for Touch Response of Zebrafish Larvae using Machine Learning

A touch-evoked response of zebrafish larvae provides information of the mechanism of the gene functional expressions. Recently, an automated system has been developed for precise and repeated touch-response experimentation with minor human intervention. The data collected by the system are analyzed with regard to an automated quantification pipeline for scientific conclusions, including five quantification criteria: latency time, C-Bend curvature maximum, C-Bend peak time, response time, and moving distance. To quantify these criteria, we propose a larva tracking based automatic quantification pipeline by using a U-Net for initialization of tracking, a particle filter as tracking strategy, and region growing for the segmentation of larvae. Experimental data with different treatments are analyzed by using the proposed quantification platform for demonstration, and the result proves that this platform can generate comparable touch-response behavior quantification readouts in an efficient
and automatic way. This platform provides an alternative to automatically screening the drugs for knowledge discovery according to the pattern of the touch-response behaviors of zebrafish larvae mutated by chemicals.