Early Detection of Parkinson's Disease Using Center of Pressure Data and Machine Learning

Area of Science:

• Neurology
• Biomedical Engineering
• Data Science

Background:

• Parkinson's disease (PD) is a neurodegenerative disorder characterized by motor symptoms like postural instability, increasing fall risk.
• Objective measurement of postural sway using center of pressure (COP) data shows potential for PD diagnosis.

Purpose of the Study:

• To investigate the efficacy of machine learning in differentiating PD patients from healthy controls (HC) based on postural sway features.
• To identify which types of features (time, frequency, time-frequency, structural) are most effective for PD classification.

Main Methods:

• COP data were collected from 19 PD patients and 13 HC.
• Various features were extracted from COP data, including time domain, frequency domain, time-frequency, and structural parameters.
• Machine learning models were employed to classify PD and HC using the extracted features.

Main Results:

• The Random Forest classifier achieved the highest accuracy, precision, and F1-score.
• Time domain features demonstrated superior performance in distinguishing PD patients from HC compared to other feature types.
• The study successfully differentiated PD patients from controls using machine learning analysis of postural sway.

Conclusions:

• Machine learning, particularly using time domain features of COP data, can effectively differentiate individuals with Parkinson's disease from healthy controls.
• Postural sway analysis offers a promising objective method for assessing PD-related motor deficits.