Early Detection of Mild Cognitive Impairment Through Balance Assessment Using Multi-Location Wearable Inertial...

Area of Science:

• Biomedical Engineering
• Neurology
• Gerontology

Background:

• Mild Cognitive Impairment (MCI) is an early stage of dementia, making early detection crucial for intervention.
• Postural instability is a key indicator of MCI, but traditional assessments can be subjective.
• Wearable inertial measurement unit (IMU) networks offer a potential solution for objective balance assessment.

Purpose of the Study:

• To develop and validate a multi-sensor balance assessment framework for identifying MCI-related postural instabilities.
• To evaluate the effectiveness of a wearable IMU network in capturing balance dynamics during static tasks.
• To determine the feasibility of non-invasive MCI screening using wearable sensor systems.

Main Methods:

• Utilized a synchronized network of five IMUs placed on the waist, thighs, and shanks.
• Recorded balance dynamics during four static tasks: Eyes-Open, Eyes-Closed, Right-Leg Lift, and Left-Leg Lift.
• Employed a three-stage feature selection strategy and Leave-One-Subject-Out cross-validation with SVM and ensemble models for classification.

Main Results:

• The multi-sensor system achieved classification accuracies between 71.7% and 79.2% for MCI detection.
• The Eyes-Open condition yielded the highest classification performance.
• A compact 10-feature subset demonstrated robust discriminative power across all tasks, outperforming a single-sensor baseline.

Conclusions:

• A multi-sensor wearable IMU network provides an effective and non-invasive method for assessing balance and identifying MCI.
• The proposed framework demonstrates the feasibility of using wearable technology for early MCI screening and intervention.
• Balance-driven assessment using compact sensor configurations shows promise for widespread clinical application.