AI-Integrated Wearable Sensor Accurately Measures Step Length for Evaluation of Neurological Diseases

Step length is recognized as a sensitive and non-invasive metric for assessing a variety of conditions and diseases such as aging, neurological deterioration, neurodegenerative diseases, cognitive decline, and disorders like Alzheimer’s, Parkinson’s, and multiple sclerosis. Traditional devices used to measure step length, including camera-based systems and force-sensitive gait mats, are stationary, bulky, and typically confined to specialized clinics and laboratories. While these methods are accurate, they capture only a momentary glimpse of a person's walking ability, which may not accurately reflect their daily functioning. Daily walking patterns can vary with changes in a person’s fatigue, mood, and medication use. To address these limitations, researchers have developed a machine learning-based interdisciplinary model that accurately estimates step length and can be integrated into a wearable device attached to the lower back with tape, facilitating continuous step monitoring in everyday settings.

Researchers at Tel Aviv University (TAU, Tel Aviv, Israel) and Ichilov’s Tel Aviv Sourasky Medical Center overcame the limitations of existing step length measurement devices by utilizing IMU (inertial measurement unit) systems. These lightweight and relatively inexpensive sensors are already embedded in every smartphone and smartwatch and can measure walking-related parameters. Previous investigations into IMU-based wearable devices for step length assessment were conducted solely on healthy individuals without walking impairments, relied on small sample sizes that prevented generalization, and involved devices that were uncomfortable and often required multiple sensors. The goal was to create an effective, user-friendly solution suitable for individuals with mobility issues, including the elderly and those with medical conditions, enabling continuous, day-long data collection on step length within the patient's usual environment.

In an article describing the research published on May 25, 2024, in the journal Digital Medicine, the team utilized IMU sensor-based gait data along with step length measurements previously gathered in a conventional setting from 472 individuals with various conditions, including Parkinson’s disease, mild cognitive impairment, healthy elderly individuals, younger healthy adults, and people with multiple sclerosis. A substantial and varied dataset of 83,569 steps was compiled from these participants. The researchers employed this dataset to train several machine learning models that convert IMU sensor data into accurate step length estimates. To assess the models’ effectiveness, they tested how well these models could generalize by evaluating their ability to accurately analyze new data not previously used in training, demonstrating the models' robustness in real-world settings.

“We found that the model called XGBoost is the most accurate and is 3.5 times more accurate than the most advanced biomechanical model currently used to estimate step length,” said Assaf Zadka, a graduate student in the Department of Biomedical Engineering at TAU, who led the research. “For a single step, the average error of our model was 6 cm, compared to 21 cm predicted by the conventional model. When we evaluated an average of 10 steps, we arrived at an error of less than 5 cm, a threshold known in the professional literature as ‘the minimum difference that has clinical importance,’ which allows identifying a significant improvement or decrease in the subject’s condition. In other words, our model is robust and reliable, and can be used to analyze sensor data from subjects, some with walking difficulties, who were not included in the original training set.”

Related Links:
Tel Aviv University
Tel Aviv Sourasky Medical Center

Visit expo >

Gold Member

SARS‑CoV‑2/Flu A/Flu B/RSV Sample-To-Answer Test

SARS‑CoV‑2/Flu A/Flu B/RSV Cartridge (CE-IVD)

Flocked Fiber Swabs
Puritan® patented HydraFlock®

New

Ultra-Flat DR Detector

meX+1717SCC

New

Mobile Autoclave

3870 HSG