Smart Health
Neurodegenerative Mouse Phenotyping
This project develops machine learning workflows to detect subtle behavioral indicators of neurodegenerative disease in mouse models using video analysis.
Back to Research
Project Summary
Early indicator detection from behavioral video data with VAME and DeepLabCut pipelines.
Alzheimer’s and related neurological diseases often show early signs that are difficult to detect visually, even though early detection is crucial for effective intervention and improved outcomes. Use behavioral video data to detect subtle early indicators of disease and quantify dose-response effects more accurately. The workflow uses DeepLabCut for pose estimation and VAME (Variational Animal Motion Embeddings) for behavior representation and analysis. The project focuses on machine learning over animal motion patterns rather than relying on manual visual inspection alone. A core objective is earlier and more sensitive diagnosis from complex longitudinal video data.