Related Experiment Video
Updated: Apr 18, 2026
10:28
Dynamic Digital Biomarkers of Motor and Cognitive Function in Parkinson's Disease
Published on: July 24, 2019
16.6K
Human-Centered Multi-Sensor Framework for Identifying Driving Patterns Associated with Cognitive Decline Through
Sonia Moshfeghi1, Seyedeh Gol Ara Ghoreishi1, Muhammad Tanveer Jan1
1College of Engineering & Computer Science, Florida Atlantic University, Boca Raton, 33431, FL, USA.
Research Square
|April 17, 2026
Summary
Telematics data from cars can help detect Mild Cognitive Impairment (MCI) by analyzing driving behavior. While promising, gender imbalance in the study impacted results, necessitating larger, balanced trials for clinical use.
Area of Science:
- Neuroscience
- Gerontology
- Computer Science
Background:
- Mild Cognitive Impairment (MCI) affects cognition and everyday activities like driving.
- Driving behavior analysis using telematics offers a novel approach for MCI detection.
Purpose of the Study:
- To investigate the feasibility of using telematics data to detect MCI in older drivers.
- To identify key driving behavior indicators associated with MCI.
- To assess the required data collection duration for reliable MCI screening.
Main Methods:
- Deployed AutoPi telematics units in 51 older drivers (10 MCI, 41 controls) for 28 months.
- Collected 20,145 trips using GPS, IMU, and OBD-II sensors.
- Employed K-means clustering, Random Forest, Welch's t-tests, and logistic regression for analysis.
Main Results:
- Achieved an AUC of 0.698 with a sensitivity of 0.800, with throttle variability and mean throttle application as key predictors.
- Model performance significantly decreased when gender was excluded, highlighting demographic confounding.
- Approximately 50 trips (four months) may be needed for reliable screening.
Conclusions:
- Telematics-based driving behavior analysis shows potential for MCI monitoring.
- Demographic factors, particularly gender, significantly influenced the model's discriminative power.
- Further validation in larger, gender-balanced cohorts is crucial before clinical deployment.
