The Transferable Methodologies of Detection Sleep Disorders Thanks to the Actigraphy Device for Parkinson's Disease Detection

Abstract

Due to population aging, society is struggling with an increasing number of patients with neurodegenerative diseases. One of them is Parkinson's disease. Early detection of Parkinson's disease is very important since there is no cure and the treatment is more effective when administered early. Wearable devices can be of great help - they are cheap and reachable, they can last for many days without charging, can provide long time monitoring, and are minimally invasive to human life. In the paper, we briefly desribe the sensors and actigraphs suitable for the analysis of sleep disturbance in Parkinson's patients and noctural symptoms of Parkinson's disease. Moreover, we pointed out how to collect the data and what could have an influence on the final performance of the automatic models. Additionally, as the main aim of this paper, we have analysed and desribed the machine learning algorithms used in the area of analysis accelerometer singla for sleep / awake stages recognition or diseases which manifested in changes in sleep patterns. We though that these algorithms, because of the nature of Parkinon's patients' sleep patterns, will be simultaneously appropriate for the detection of Parkinon's disease.Due to population aging, society is struggling with an increasing number of patients with neurodegenerative diseases. One of them is Parkinson's disease. Early detection of Parkinson's disease is very important since there is no cure and the treatment is more effective when administered early. Wearable devices can be of great help - they are cheap and reachable, they can last for many days without charging, can provide long time monitoring, and are minimally invasive to human life. In the paper, we briefly desribe the sensors and actigraphs suitable for the analysis of sleep disturbance in Parkinson's patients and noctural symptoms of Parkinson's disease. Moreover, we pointed out how to collect the data and what could have an influence on the final performance of the automatic models. Additionally, as the main aim of this paper, we have analysed and desribed the machine learning algorithms used in the area of analysis accelerometer singla for sleep / awake stages recognition or diseases which manifested in changes in sleep patterns. We though that these algorithms, because of the nature of Parkinon's patients' sleep patterns, will be simultaneously appropriate for the detection of Parkinon's disease.