On-Body Sensor Position Identification with a Simple, Robust and Accurate Method, Validated in Patients with...

Area of Science:

• Biomedical Engineering
• Rehabilitation Technology
• Wearable Sensor Technology

Background:

• Body-worn inertial sensors are crucial for monitoring movement disorders.
• Accurate sensor placement is essential but often requires manual labeling, posing challenges for adherence and data integrity.
• Existing methods lack automation, creating a burden for patients and clinicians.

Purpose of the Study:

• To develop and validate an automated method for localizing body-worn inertial sensors on key body segments (torso, wrists, shanks).
• To reduce the burden associated with manual sensor placement and improve adherence in long-term patient monitoring.
• To enhance the reliability of data collected from inertial measurement units (IMUs) in clinical and home settings.

Main Methods:

• Implementation of an automated algorithm for identifying sensor locations on the torso, wrists, and shanks.
• Validation of the method in a multi-site clinical trial involving Parkinson's disease patients.
• Assessment of accuracy in identifying sensor placement and distinguishing between left/right limbs.

Main Results:

• The automated method achieved 100% accuracy in identifying sensor placement on the torso, wrists, and shanks.
• It demonstrated 100% accuracy in discriminating between left and right shanks.
• The system achieved 98% accuracy in differentiating between left and right wrists, even with parkinsonian motor symptoms present.

Conclusions:

• The automated sensor localization method is highly accurate and reliable for body-worn inertial sensors.
• This technology can significantly facilitate home-based monitoring for patients with movement disorders.
• The findings support the use of automated sensor placement in rehabilitation and ambient monitoring applications.