From Traditional to Transformative: Gait Analysis With Wearable Technology and Machine Learning in CSVD Diagnosis and Research.
Saved in:
| Title: | From Traditional to Transformative: Gait Analysis With Wearable Technology and Machine Learning in CSVD Diagnosis and Research. |
|---|---|
| Authors: | Yi, Ming (AUTHOR), Fan, Shaoyi (AUTHOR), Xiao, Chi (AUTHOR), Yang, Jing (AUTHOR), Guo, Jiayu (AUTHOR), Yu, Lei (AUTHOR), Hu, Bin (AUTHOR), Dang, Chao (AUTHOR), Xu, Fuping (AUTHOR), Fan, Yuhua (AUTHOR), Cumming, Paul (AUTHOR) |
| Source: | Acta Neurologica Scandinavica. 5/16/2025, Vol. 2025, p1-11. 11p. |
| Subjects: | Cerebral small vessel diseases, Machine learning, Early diagnosis, Diagnosis, Prediction models, Wearable technology, Physiological aspects of walking |
| Abstract: | Background: Cerebral small vessel disease (CSVD) often manifests with gait impairment, a critical yet overlooked aspect of early disease progression. Our study is aimed at leveraging wearable sensors and machine learning to analyse gait characteristics, providing a cost‐effective and scalable method for early CSVD diagnosis. Methods: We collected baseline and gait data from 115 individuals diagnosed with CSVD and 120 community healthy controls. All participants underwent a quantitative gait assessment utilizing the wearable device Ambulosono. The study applied an affordable digital 6‐min walk test (6MWT) for standardized assessment, employing machine learning to build a prediction model. Results: Traditional binary logistic regression statistical analysis revealed that the most distinguished gait thresholds during a 6‐min walk were walking speed (≤ 70.34 m/min; sensitivity 0.625, specificity 0.791, AUC 0.760) and cadence (≤ 117.45; sensitivity 0.658, specificity 0.748, AUC 0.738). Gait variability was not statistically significantly different. Compared with traditional statistics, the machine learning model greatly improved the ability of gait characteristics to predict CSVD. We used a random forest model to train the selected features, and the AUC of the CSVD prediction mode increased from 0.83 to 0.94 (p = 0.006 DeLong's test), with 82% accuracy, 78% specificity, 86% sensitivity, 79% precision, and an F1‐score of 0.82. Conclusions: Our findings underscore the innovative application of gait features and machine learning in CSVD diagnosis. The integration of the affordable digital 6MWT gait tool with machine learning represents a promising approach for early detection and large‐scale population screening. [ABSTRACT FROM AUTHOR] |
| Copyright of Acta Neurologica Scandinavica is the property of Wiley-Blackwell and its content may not be copied or emailed to multiple sites without the copyright holder's express written permission. Additionally, content may not be used with any artificial intelligence tools or machine learning technologies. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.) | |
| Database: | Psychology and Behavioral Sciences Collection |
|
Full text is not displayed to guests.
Login for full access.
|
|
| Abstract: | Background: Cerebral small vessel disease (CSVD) often manifests with gait impairment, a critical yet overlooked aspect of early disease progression. Our study is aimed at leveraging wearable sensors and machine learning to analyse gait characteristics, providing a cost‐effective and scalable method for early CSVD diagnosis. Methods: We collected baseline and gait data from 115 individuals diagnosed with CSVD and 120 community healthy controls. All participants underwent a quantitative gait assessment utilizing the wearable device Ambulosono. The study applied an affordable digital 6‐min walk test (6MWT) for standardized assessment, employing machine learning to build a prediction model. Results: Traditional binary logistic regression statistical analysis revealed that the most distinguished gait thresholds during a 6‐min walk were walking speed (≤ 70.34 m/min; sensitivity 0.625, specificity 0.791, AUC 0.760) and cadence (≤ 117.45; sensitivity 0.658, specificity 0.748, AUC 0.738). Gait variability was not statistically significantly different. Compared with traditional statistics, the machine learning model greatly improved the ability of gait characteristics to predict CSVD. We used a random forest model to train the selected features, and the AUC of the CSVD prediction mode increased from 0.83 to 0.94 (p = 0.006 DeLong's test), with 82% accuracy, 78% specificity, 86% sensitivity, 79% precision, and an F1‐score of 0.82. Conclusions: Our findings underscore the innovative application of gait features and machine learning in CSVD diagnosis. The integration of the affordable digital 6MWT gait tool with machine learning represents a promising approach for early detection and large‐scale population screening. [ABSTRACT FROM AUTHOR] |
|---|---|
| ISSN: | 00016314 |
| DOI: | 10.1155/ane/1369705 |
