Artificial intelligence in Parkinson’s disease and other movement disorders

Abstract

Artificial intelligence has been making its way progressively, and movement disorders have not been an exception to this trend. The relative ease of data collection means that there are many artificial intelligence initiatives to try to diagnose or evaluate patients with movement disorders. Successful studies have been carried out collecting data from wearables or other devices, radiological images, videos of patients, or neurophysiological signals. The results are generally variable, but most studies present an accuracy above 80%. Currently, monitoring through wearables is the most advanced one, and this is reflected in the emergence of commercial devices that are already available. Despite all the hype around this new technology, there are still challenges regarding validity, safety, and accountability that must be resolved for widespread use in daily clinical practice.