Among movement disorders, essential tremor is by far the most common, as much as eight times more prevalent than Parkinson’s disease. Although these two conditions differ in their presentation and course, clinicians do not always recognize them, leading to common misdiagnoses. Proper and early diagnosis is important for receiving the right treatment and support. In this paper, the development of a portable and reliable tremor classification system based on a wearable device, enabling clinicians to differentiate between essential tremor and Parkinson’s disease-associated one, is reported. Inertial data were collected from subjects with a well-established diagnosis of tremor, and analyzed to extract different sets of relevant spectral features. Supervised learning methods were then applied to build several classification models, among which the best ones achieved an average accuracy above 90%. Results encourage the use of wearable technology as effective and affordable tools to support clinicians.
Automatic Classification of Tremor Severity in Parkinson’s Disease Using a Wearable Device
