ZebraTrack: An Open-Source Object Detection Algorithm to Detect and Track Larval Zebrafish Motor Touch Responses

Abstract

MotivationZebrafish (Danio rerio) are a model organism used for the study of vertebrate development, disease and drug discovery. Two-day old larval zebrafish exhibit burst swimming behaviour that can be elicited by a light touch to the tail. Larval motor touch-responses are frequently video recorded and later analyzed. Methods to robustly analyze these videos in a reproducible and time-efficient manner are reliant on manual tracking, which is prone to experimenter bias and error. ResultsHere we present ZebraTrack, a machine learning-based program, which employs Ultralytics YOLOv8 nano (YOLOv8n) object detection algorithm to automatically analyze larval touch response videos. The program breaks down video files into their constituent frames and passes these through a custom-trained YOLOv8n algorithm to detect the presence of a single larval zebrafish. ZebraTrack then refines the tracking data output by the model and tabulates it into an excel spreadsheet. The program then computes and extracts four relevant swim metrics: swim duration (s), swim distance (mm), mean swim velocity (mm/s), and max swim velocity (mm/s). ZebraTrack rapidly accelerates the analysis process, while also eliminating the errors associated with manual tracking. Furthermore, it allows for high-throughput analysis of larval touch response videos and can detect subtle differences in motor metrics arising as a result of temperature differences, demonstrating that utility of this tracking algorithm. Availability and implementationZebraTrack is available for download at https://github.com/Armstrong-Lab-70/ZebraTrack.