Development and Assessment of a Movement Disorder Simulator Based on Inertial Data

Area of Science:

• Biomedical Engineering
• Medical Informatics
• Computational Neuroscience

Background:

• Telemedicine for neurodegenerative diseases relies on low-cost sensors and classification algorithms.
• Acquiring large datasets for sensor/algorithm development requires extensive patient studies and ethics approval, causing significant delays.
• Software simulation offers a viable alternative for initial research stages, accelerating development.

Purpose of the Study:

• To develop, validate, and utilize a software simulator for generating movement disorder data.
• To create a tool that simulates data for both healthy and pathological conditions based on raw inertial measurement data.
• To demonstrate the simulator's application in Parkinson's disease tremor analysis and assess data quality for classification.

Main Methods:

• Development of a software simulator to process raw inertial measurement data into tri-axial acceleration and angular velocity.
• Focus on a case study of Parkinson's disease tremor for data generation and analysis.
• Implementation of a machine learning method to evaluate data suitability for classification.
• Analysis of data quality metrics for classification accuracy with low-performance sensors.

Main Results:

• Simulator validation showed high correlation coefficients: >0.94 for angular velocity and >0.93 for acceleration.
• Machine learning classification achieved over 98% accuracy for Parkinson's disease tremor under optimal conditions.
• The simulator provides flexible and user-friendly generation of pathological data.

Conclusions:

• The developed software simulator effectively generates reliable movement disorder data, aiding research in neurodegenerative diseases.
• The tool facilitates the development and testing of telemedicine applications using low-cost sensors.
• High classification accuracy demonstrates the simulator's potential for advancing Parkinson's disease diagnosis and monitoring.