Daniel Krauß at Spaulding Rehabilitation Hospital in Boston

During my research stay at Spaulding Rehabilitation Hospital in Boston, Massachusetts, I worked under the supervision of Professor Paolo Bonato from Harvard Medical School. My research focused on developing a novel machine-learning technique to automatically annotate video recordings for Parkinson’s disease (PD) studies. Specifically, I used a dataset that had been previously recorded in the Motion Analysis Lab in Boston as part of Study 3 of the larger Blue Sky project.

The aim of this project was to analyze posture changes to assess orthostatic dysfunctionality, which is a common issue faced by individuals with Parkinson’s disease. The project was a collaborative effort, involving Friedrich-Alexander-University Erlangen-Nürnberg and the Biomedical Data Science group from the Luxembourg Centre for Systems Biomedicine at the University of Luxembourg, who are performing heart rate variability (HRV) analysis in PD patients.

By utilizing this existing dataset, I applied machine learning techniques to automatically detect and annotate posture changes in the video recordings. These annotations focused on key movements such as sitting, standing, and transitions between these states. Understanding these posture changes provides crucial insights into orthostatic dysfunctionality in Parkinson’s patients, potentially improving clinical assessments and interventions.

I had the opportunity to present the preliminary results of this work at the BSN 2024 conference in Chicago, where I gave a poster presentation titled „Analyzing multi-camera video data for posture change detection to assess orthostatic dysfunctionality in Parkinson‘s disease.“ The presentation was well-received, and the feedback will help further develop the project.

This experience has been instrumental in advancing my knowledge of machine learning applications in clinical settings, and I look forward to further collaboration on this collaborative research work.