Related Experiment Video
Updated: Jul 30, 2025
Home-Based Monitor for Gait and Activity Analysis
Published on: August 8, 2019
Development of a Gait Feature-Based Model for Classifying Cognitive Disorders Using a Single Wearable Inertial Sensor
Jeongbin Park1, Hyang Jun Lee1, Ji Sun Park1
1From the PlanB4U Research Institute (J.P., C.H.K., W.J.J., K.W.K.), Seongnam; Department of Neuropsychiatry (H.J.L., J.B.B., J.W.H., K.W.K.), Seoul National University Bundang Hospital, Seongnam; Department of Brain and Cognitive Science (J.S.P., K.W.K.), Seoul National University College of Natural Sciences; Medical Research Collaborating Center (S.W.), Seoul National University Bundang Hospital, Seongnam; and Department of Psychiatry (K.W.K.), Seoul National University, College of Medicine, Korea.
Gait analysis using wearable sensors can effectively identify cognitive disorders in older adults. This method shows diagnostic performance comparable to traditional cognitive tests.
Area of Science:
- Gerontology
- Neurology
- Biomedical Engineering
Background:
- Gait changes are recognized as potential early indicators of cognitive disorders (CDs).
- Distinguishing between older adults with CD and normal cognition is crucial for timely intervention.
Purpose of the Study:
- To develop and validate a model for classifying cognitive disorders in older adults using gait parameters.
- To compare the diagnostic performance of the gait-based model with the Mini-Mental State Examination (MMSE).
Main Methods:
- Utilized gait speed and variability data from wearable inertial sensors in community-dwelling older adults.
- Developed a logistic regression model using an 80% development dataset and validated it on a 20% validation set.
- Compared the gait model's performance against an MMSE-based model using receiver operator characteristic analysis.
Main Results:
- The gait-based model demonstrated good diagnostic performance in both development (AUC=0.788) and validation (AUC=0.811) datasets.
- The model's diagnostic accuracy for CD was comparable to that of the MMSE in both datasets.
- The optimal cutoff score for the gait-based model was determined to be >-1.56.
Conclusions:
- A gait-based model utilizing wearable inertial sensors shows promise as a diagnostic tool for cognitive disorders in older adults.
- Gait analysis offers a viable, non-invasive method for identifying cognitive impairment in the elderly population.
