Predicting Adult Zebrafish Anxiety Behaviours Using the Behavioural Deep Context Analysis Network (BDCAN)

The BDCAN model predicts adult zebrafish anxiety-related behaviors using machine learning. Thus, knowing zebrafish behavior patterns is important in neurosciences, pharmacology, and anxiety disorders and therapies. The proposed BDCAN uses multi-modal spatio-temporal feature extraction and adaptive attention techniques to identify and characterize zebrafish anxiety subtleties based on BDCAN3D. Anxiety-related behaviors including irregular swimming, freezing, and thigmotaxis may be assessed using movement trajectory metrics like speed and spatial preferences identified by the network. The model design uses Convolutional Neural Networks to capture spatial characteristics, Long Short-Term Memory Networks to analyze temporal variables, and an attention layer to identify critical behavioral transitions. BDCAN has 98% accuracy in identifying anxiety-like behaviors with proper class labeling and filtering out sounds during testing.State-of-the-art deep machine learning can investigate zebrafish behavior with improved resolution and temporal stochastic dynamics. Aquatic behavioural analysis relies on the innovative, precise, and fast BDCAN framework to study anxiety and therapeutic chemicals in preclinical animals.