Gait Analysis using BCI and Multi-Sensor System
- Sundararajan Venkatadriagaram, Assistant Professor of Mechanical Engineering, University of California Riverside
- Ehsan Tarkesh Esfahani, Assistant Professor , University at Buffalo
Recent trends in rehabilitation emphasize the direct control of prosthetic and assistive devices by using brain activity leading to the development of Brain-Computer Interface (BCI) integrated devices. In our research, a multi-modal monitoring system is presented to estimate different phases of the human walking cycle from kinematic and ground reaction forces (GRF) to muscle and brain activity. First, a model of the biped walking trajectory was generated using polynomial spline and sinusoidal functions. Second, a compact portable wireless multi-sensor system was designed, which includes a three-dimensional accelerometer and two force sensing resistors at the ankle-foot for recording kinematic data and GRF at the toe and heel. By applying a filter and threshold method, the different phases of the walking cycle (toe-off, swing, and heel-strike) were detected. Third, to detect the human’s intention for undergoing the walking movement, electroencephalography (EEG) signals were recorded at 14 locations of the scalp with a non-invasive wireless neuro-headset (Emotiv). Our experimental study focuses on analyzing the steady-state and transient-state conditions during treadmill walking at different speed and standing. We computed the normalized power spectral density (NPSD) of the EEG signals in different bands. The outcome of our analysis indicated that during treadmill walking the NPSD in the theta band (~4-8hz) was primarily higher than during standing, and the NPSD in the alpha band (~8-12hz) was greater for standing than during walking. This observation is in-line with the occurrence of ‘event-related de-synchronization and synchronization’ (ERD/ERS) illustrating a variant synchronization of cortical rhythms thus allowing changes in the amplitude of frequency bands. The rhythms decrease or de-synchronize with movement and increase or synchronize after movement with relaxation. In our future studies, further analysis of the correlation between brain activity and walking will be done.