Reconstructing gait cycle patterns from non-invasive recorded low gamma modulations

This work presents a simple to set-up system for reconstructing gait cycle patterns from non-invasive recorded electroencephalographic (EEG) signals. It is based on the prior finding that low gamma amplitudes are modulated locked to the gait cycle in central sensorimotor areas. Therefore, we focused on a Laplacian Cz derivation and low gamma amplitude modulations to reconstruct the gait patterns. Our results show that this method was successful in reconstructing gait cycle patterns in 8/10 subjects during active walking and in every subject during passive walking. The median reconstruction error was 0.24±0.13 s for active and 0.26±0.10 s for passive walking. The presented methods and findings are a further step towards analysing and monitoring ongoing cortical activity during human upright walking.