Everyday Life Tremor Signal Processing in PD Patients using BSN

Abstract

Parkinson’s disease is a neurological disorder characterized by the deficiency of dopamine levels in the brain. More than 75 percent of these patients suffer from tremors. Parkinsonian tremor (PT) is more characterized to be a rest tremor, but some patients suffer from action tremor as well. Usually, patients suffering from this disease are diagnosed by their physicians who perform some battery MDS-UPDRS tasks to determine the disorder. Some sensors were used to diagnose the tremor objectively, but in this study, we are using a new Body Sensor Network (BSN) designed at our institute to be used in detecting the acceleration, gyroscope, and magnetometer of the tremor patients in the clinic. Signal processing of the recorded data is performed to determine and classify the number of times throughout the day the patient suffered from tremors. This is ensured through automatic signal segmentation, extraction of several signal features, and classification with the most accurate machine learning classifier. In this study, we have proved that our BSN sensor is capable of helping clinicians in classifying tremor occurrence in Parkinson diseased patients specifically, and tremor patients generally throughout monitoring their everyday life activities.