Parkinson's Disease Drosophila Model

Advanced neuroscience research using machine learning to study motor control

Overview

During my research internship at Japan's premier comprehensive research institute, I worked under Dr. Fujiwara to explore the intricacies of Parkinson's disease using cutting-edge techniques combining behavioral neuroscience with machine learning.

Problem

Understanding the neural mechanisms underlying motor control deficits in Parkinson's disease requires sophisticated experimental setups and analysis tools. Traditional approaches to studying gait patterns provide limited insights into the complex neural dynamics.

Solution

Built a behavioral setup for freely walking drosophila using high-speed camera technology. Developed a Parkinson's disease model using the Split-gal4 technique to specifically inhibit dopaminergic neurons in the drosophila central complex. Analyzed complex gait patterns using Machine Learning tools including DeepLabCut for pose estimation and UMAP for dimensionality reduction.

Impact & Results

• •Advanced understanding of motor control in neurological conditions
• •Developed novel experimental paradigm combining behavioral setup with ML analysis
• •Mastered machine learning techniques (DeepLabCut, Bonsai, UMAP) despite no prior ML background
• •Contributed to international collaborative research at world-renowned institute

Key Skills Applied