#293: Dissecting branchiomotor neuron circuits in zebrafish - Toward high-throughput automated analysis of jaw movements


We use the zebrafish model to understand brain circuits containing branchiomotor neurons that control jaw movements in vertebrates. In one approach, we compare circuit output, evaluated by dynamic analysis of jaw movement from video recordings, between wildtype zebrafish and mutants with specific circuit defects. While such analyses have been informative, they require extensive manual annotation. We developed an automated image analysis system to extract deformable motion features using video microscopy, for rapid and accurate detection of jaw movement termed gape or mouth opening. The approach uses a robust fusion process combining multiple local deformable motion models into a likelihood response trace curve followed by peak processing to reliably detect jaw movement. Performance on 24 labeled videos of zebrafish larvae across three age groups achieved 97.2% recall, 93.5% precision and 98.7% accuracy. The automated system demonstrates high fidelity gape analysis scalable for high-throughput studies in neuroscience and pharmacology.