Integrated EEG-fNIRS for Characterizing Cortical Responses and Neurovascular Coupling in Automated and Discrete Gait Tasks
Fengxian Wu, Yaming Liu, Hucheng Jiang, Luyao Li, Chenglong Feng, Jiayi Sun, Wenxin Niu (2025) Integrated EEG-fNIRS for Characterizing Cortical Responses and Neurovascular Coupling in Automated and Discrete Gait Tasks IEEE Trans Neural Syst Rehabil Eng (IF: 5) PPAbstract
Walking is a fundamental human motor pattern supported by multi-level neural control. Previous research has extensively explored cortical responses during routine walking and complex gait scenarios. However, these studies often conflate basic gait control with cognitive demands, making it unclear how distinct cortical responses are elicited by automated versus discrete gait tasks. To address this, integrated EEG-fNIRS enables high spatiotemporal resolution characterization of cortical responses during these tasks in real-world conditions. This study proposes a framework that incorporates three gait tasks and simultaneously collects EEG and fNIRS data to characterize cortical responses and neurovascular coupling between automated and discrete gait tasks. Eighteen healthy participants performed continuous walking (CW), isolated gait phase (IGPT), and single-limb stance (SS) tasks during simultaneous EEG-fNIRS recording. Task-Related Component Analysis (TRCA) extracted task-related features, validated against channel-averaging methods. The coupling coefficient between EEG and fNIRS signals was computed using time-lagged maximum cross-correlation analysis. XGBoost classified tasks using different data inputs (channel averaging vs. TRCA, unimodal vs. bimodal). Repeated-measures ANOVA assessed inter-task differences. Results showed that beta-band suppression was stronger in IGPT vs. CW (p = 0.040), while SS showed higher fNIRS activation than CW (p = 0.026). TRCA significantly enhanced within-class similarity and between-class discriminability of EEG-fNIRS features across all tasks (p < 0.05), and revealed task-specific alpha-band neurovascular coupling, with stronger negative coupling in IGPT vs. CW (p = 0.046). Multimodal TRCA-based fusion achieved 74.51% classification accuracy, significantly outperformed EEG-Avg (49.02%, p = 0.042) and fNIRS-Avg (47.06%, p = 0.038). This study establishes an EEG-fNIRS framework that reveals task-specific cortical responses and neurovascular coupling differences between automated and discrete gait tasks, providing a foundation for further exploration of gait control and rehabilitation.
Links
http://www.ncbi.nlm.nih.gov/pubmed/40956734http://dx.doi.org/10.1109/TNSRE.2025.3610690
Similar articles
Tools
