Analysis of gait pattern related to high cerebral small vessel disease burden using quantitative gait data from wearable sensors
Kelin Xu,Yingzhe Wang,Yanfeng Jiang,Yawen Wang,Peixi Li,Heyang Lu,Chen Suo,Ziyu Yuan,Qi Yang,Qiang Dong,Li Jin,Mei Cui,Xingdong Chen
DOI: https://doi.org/10.1016/j.cmpb.2024.108162
IF: 6.1
2024-04-14
Computer Methods and Programs in Biomedicine
Abstract:Background and Objectives Sensor-based wearable devices help to obtain a wide range of quantitative gait parameters, which provides sufficient data to investigate disease-specific gait patterns. Although cerebral small vessel disease (CSVD) plays a significant role in gait impairment, the specific gait pattern associated with a high burden of CSVD remains to be explored. Methods We analyzed the gait pattern related to high CSVD burden from 720 participants (aged 55-65 years, 42.5% male) free of neurological disease in the Taizhou Imaging Study. All participants underwent detailed quantitative gait assessments (obtained from an insole-like wearable gait tracking device) and brain magnetic resonance imaging examinations. Thirty-three gait parameters were summarized into five gait domains. Sparse sliced inverse regression was developed to extract the gait pattern related to high CSVD burden. Results The specific gait pattern derived from several gait domains (i.e., angles, phases, variability, and spatio-temporal) was significantly associated with the CSVD burden (OR=1.250, 95% CI: 1.011-1.250). The gait pattern indicates that people with a high CSVD burden were prone to have smaller gait angles, more stance time, more double support time, larger gait variability, and slower gait velocity. Furthermore, people with this gait pattern had a 25% higher risk of a high CSVD burden. Conclusions We established a more stable and disease-specific quantitative gait pattern related to high CSVD burden, which is prone to facilitate the identification of individuals with high CSVD burden among the community residents or the general population.
engineering, biomedical,computer science, interdisciplinary applications,medical informatics, theory & methods