Detection methods of Parkinson's Disease based on physiological signals and machine learning methods

Abstract

Parkinson's disease (PD) is an extremely complex motor disorder due to the lack of dopaminergic neurons in the substantia nigra. and other dopaminergic and non-dopaminergic regions of the brain. The high rate of misdiagnosis in Parkinson's disease often causes patients to miss out on the best treatment opportunities. Since some of the symptoms of Parkinson's disease are mild in the initial stages and become severe over time, it is particularly important to correctly diagnose Parkinson's disease timely. The traditional tremor detection method of Parkinson's disease is more complex and the misdiagnosis rate is high. Methods based on physiological signals such as Local field potential (LFP), Electromyographic signal (EMG) and EEG signal et.al and research by using the machine learning strategies including the traditional machine learning and deep leaning methods are increasing. Get a precise diagnosis for Parkinson's disease, this paper analyzes physiological signals and machine learning methods that commonly used in PD detection, which may provide theoretical and practical references to future studies.