For a visual discrimination task, mice have been trained freely moving in a touchscreen chamber. During this training, mice not only perform behaviours relevant or necessarily related to the task, but a multitude of behaviours. The aim of this semester project is to use the video data recorded during the training of the mice to classify the behaviour and identify specific behavioural epochs using DeepLabCut (https://github.com/DeepLabCut/DeepLabCut), a deep-learning based video tracking software.
These behavioural epochs can be then related to the neuronal calcium imaging data acquired using a minituarized microscope.

Tasks:
• Set up DLC and optimize for tracking of the mice in the provided video data
• Use additional behavioural classification toolboxes (e.g. B-SOiD) to classify behavioural epochs
• Correlate with neuronal data