Acquisition and Automated Segmentation of Inertia Sensor Data for Mobile Camptocormia Assessment

Area of Science:

• Biomedical Engineering
• Neurology
• Wearable Technology

Background:

• The camptocormia angle is a key metric for assessing Parkinson's disease progression and treatment effectiveness.
• Current measurement methods are often static and may not capture disease dynamics accurately.

Purpose of the Study:

• To develop and validate a wearable system for objective, long-term measurement of the camptocormia angle.
• To implement a machine learning approach for analyzing sensor data and segmenting patient activities.

Main Methods:

• Utilized five inertial sensors in a wearable setup to measure the camptocormia angle using the perpendicular method.
• Evaluated various inertial measurement unit sensors for suitability in mobile, long-term monitoring.
• Applied machine learning, specifically an artificial neural network (ANN) and a support vector machine (SVM), for data segmentation and activity recognition.

Main Results:

• An artificial neural network outperformed a support vector machine in recognizing patient activities.
• The ANN achieved high performance metrics: 92.4% accuracy, 82.9% sensitivity, and 82.1% F1-score.
• Identified suitable inertial sensors for reliable long-term, mobile measurements.

Conclusions:

• The proposed wearable system offers a promising solution for continuous, objective camptocormia angle monitoring in Parkinson's disease.
• This technology is expected to provide more refined parameters for evaluating disease progression and therapeutic efficacy compared to traditional methods.