Recognizing Pathological Movement Patterns - With the Help of Artificial Intelligence

Göttingen research project "Deep Movement Diagnostics" receives around EUR 1.2 million for the development of three-dimensional reconstructions of movement patterns

The research team of the "Deep Movement Diagnostics" project
The research team of the "Deep Movement Diagnostics" project (from left): Prof. Melanie Wilke an Prof. Mathias Bähr, both University Medical Center Göttingen, Prof. Alexander Gail, German Primate Center, Prof. Florentin Wörgötter, University of Göttingen, and Prof. Hansjörg Scherberger, German Primate Center. Photo: Karin Tilch

Reliably evaluating walking and gripping movements of patients is essential for the diagnosis and therapy of movement disorders, for example after a stroke or in Parkinson's syndromes. However, the success of this challenging diagnostic procedure depends to a large extent on the experience and skills of the attending physician.

"Movement disorders such as trembling, paralysis or muscle tension disorders affect many patients suffering from strokes or neurodegenerative diseases such as Parkinson's or multiple sclerosis. The precise and reproducible recording of mobility restrictions represents a major challenge in diagnostics and therapy control, as this requires experienced doctors who are not always available," explains Mathias Bähr, Director of the Neurological Clinic at Göttingen University Hospital.

This is where the "Deep Movement Diagnostics" project, coordinated by Alexander Gail, scientist at the German Primate Center - Leibniz Institute for Primate Research, comes in. The team, which also includes Mathias Bähr and Melanie Wilke, both University Medical Center Göttingen, Florentin Wörgötter, University of Göttingen, and Hansjörg Scherberger, German Primate Center, wants to use findings from machine learning and robotics to improve the objective assessment of movement patterns. "We will bundle our expertise in body and eye movement research in humans and monkeys, neurophysiology and clinical neurology as well as prosthetics and robotics," says Alexander Gail.

Recording of finger movements
Video (12 sec.) opens upon click on the image: Detailed video-based recording of the movement of all finger joints of a hand when gripping an object using artificial intelligence. (Photo and Video: Swathi Sheshadri)

Using state-of-the-art digital methods, walking and gripping movements will be measured and modeled with previously unattainable precision to provide diagnostic tools for individualized therapeutic approaches, for example for Parkinson's or stroke patients. The studies on motor skills in monkeys play an important role in this, they are the basis for later applications to humans. "Our goal is to develop an inexpensive, easy-to-use system for a widespread use in diagnosis and monitoring of therapies of movement disorders," says project leader Alexander Gail.

"Ease of use and low effort for carrying out the examinations are important factors in increasing patient acceptance," says Melanie Wilke, Director of the Institute of Cognitive Neurology. "We expect the new video-based methods to allow a qualitative leap forward compared to current clinical investigation techniques.

In addition to diagnostics, the research team wants to investigate complex movement sequences in healthy volunteers and monkeys in order to better understand the neurophysiological basics of movement disorders.

The research team will receive about 1.2 million euros over a period of three years from the "Big Data in the Life Sciences of the Future" funding line, which was put out to tender by the "Niedersächsisches Vorab" initiative of the Volkswagen Foundation.