Related Experiment Video
Updated: Jul 2, 2026
07:24
Home-Based Monitor for Gait and Activity Analysis
Published on: August 8, 2019
Wearable Sensor-Based Mild Cognitive Impairment Identification: A Multi-Domain Gait Analysis Approach With
Summary
This study introduces a novel method for identifying Mild Cognitive Impairment (MCI) using wearable sensors. The approach combines multi-domain gait analysis with explainable AI, achieving high accuracy in detecting MCI through movement patterns.
Area of Science:
- Gerontology and Cognitive Neuroscience
- Biomedical Engineering and Wearable Technology
- Machine Learning and Data Science
Background:
- Mild Cognitive Impairment (MCI) signifies cognitive decline beyond normal aging, often preceding dementia, necessitating early detection for intervention.
- Conventional clinical methods for MCI identification face challenges in accuracy and accessibility.
- Inertial Measurement Units (IMUs) offer potential for portable, motion-based MCI screening, but existing feature extraction methods are limited.
Purpose of the Study:
- To develop and validate an integrated framework for MCI screening using multi-domain gait analysis and explainable machine learning.
- To systematically analyze comprehensive features from IMU data across time, frequency, nonlinear, and time-frequency domains.
- To enhance the interpretability and clinical applicability of IMU-based MCI detection.
Main Methods:
- Collected data using waist-mounted IMUs during standardized Short Physical Performance Battery (SPPB) assessments in older adults.
- Extracted comprehensive gait features across four domains: time, frequency, nonlinear, and time-frequency.
- Employed a hybrid feature selection strategy and association rule mining for dimensionality reduction and identification of key diagnostic feature combinations.
Main Results:
- The proposed framework achieved a mean test Area Under the ROC Curve (AUC) of 0.897 and a mean test accuracy of 0.828 using an SVM classifier.
- The study identified key diagnostic feature combinations strongly linked to MCI through association rule mining.
- The approach demonstrated significant potential for cost-effective and generalizable IMU-based MCI screening.
Conclusions:
- The integrated framework combining multi-domain gait analysis and explainable machine learning shows high efficacy in identifying MCI.
- This method offers a promising, interpretable, and clinically applicable tool for early MCI detection using wearable IMUs.
- The findings support the development of accessible, real-world screening tools for cognitive decline.
