Investigating neural correlates of locomotion transition via temporal relation of EEG and EOG-recorded eye movements

Abstract

The present study examines a temporal relation of walking behavior during locomotion transition (walking to stair ascent) to electrooculography (EOG) signals recorded from eye movement. Further, electroencephalography (EEG) signals from the occipital region of the brain are processed to understand the relative occurrence in EOG and EEG signals during the transition. The dipole sources in the occipital region with reference to EOG detection were estimated from independent components and then clustered using the k means algorithm. The dynamics of the dipoles in the occipital cluster in different frequency bands revealed significant desynchronization in the β and low γ bands, followed by resynchronization. This transitional behavior coincided with transient features suggesting possible saccadic movement of the eyes in the EOG signal. With the data from six able-bodied participants, the desynchronization in EEG from the occipital region was detected by nearly 2.2 ± 0.5s before the transition. Using preprocessing techniques on the EOG signal followed by detecting saccades from the derivative of the EOG signal, the eye movements were detected by nearly 2.5 ± 0.5s before the transition. The EOG decoded intention of transition appeared as early as 3.0 ± 1.63s before desynchronization in the EEG. A paired t-test analysis showed that the EOG-based intent decoding of transition reflects significantly earlier than occipital EEG (p < 0.00001). This study could lead to a multi-modal neural-machine interface that may produce results superior to previous attempts involving only EEG and EMG signals.