Fig 1.
Exemplary EEG data recorded during different walking speeds before and after removal of movement artifacts.
Panels A,B,C and D depict high pass filtered EEG data recorded during TM walking at speeds of 0.4, 0.8, 1.6 and 2.2m/s, respectively. To the right, panels E, F, G and H depict the same time segments in the data after movement artefact removal. The x axis represents time in seconds, the y axis shows the voltage vs. time for 10 electrodes (voltage curves are shifted vertically for clarity).
Fig 2.
Cz electrode and accelerometer power spectra comparison.
The top row depicts the accelerometer’s power spectra in the various walking trials, while the middle and bottom rows depict the power spectra of the Cz electrode after preprocessing and MA-removal, respectively. X and Y axes of the plots stand for frequency and spectral power, respectively.
Fig 3.
Decaying harmonics at 0.5 multiples of the stepping frequency.
Frequency spectrum of an exemplary AMICA component, decomposed from a 1.6 m/s TM trial. Each spectral peak’s frequency and amplitude are depicted in the data-tips. The stepping frequency, as derived from the accelerometer, is shown as a red dashed line.
Fig 4.
W/S ratios are presented for the various walking trials for the preprocessed-only, AMICA-cleaned and ASR-cleaned EEG data in blue, orange and gray, respectively. X and y axes depict the various trials and the corresponding W/S ratios, respectively. The mean cross-subject regular and elevated OG walking speeds were 1.37 and 1.81 m/s, respectively. Shown are the cross-subject group average and standard deviation. For the preprocessed data (blue), all multiple comparisons are significantly different for TM walking (p<0.01, as indicated by the asterisks), as opposed to AMICA and ASR cleaned data.
Fig 5.
W/S ratios spatial distribution across the scalp.
Quantitative comparison between 6 EEG signals before and after MA removal. The spectral energy in the 5–80 Hz band was compared to the measured baseline’s spectral energy of the same frequency band. The values, spatially distributed across the scalp, depict the spectral energy ratios between walking and sitting (i.e., baseline), around the stepping frequencies associated with the different speeds (warmer colors reflect higher ratios). The mean cross-subject regular and elevated OG walking speeds were 1.37 and 1.81 m/s, respectively. Black dots depict scalp electrodes.
Fig 6.
Alpha and beta bands analysis during motor activity.
W/S ratios specifically calculated for the alpha [7.5–12 Hz] and beta [13–30 Hz] bands are presented in panels A and B, respectively. Ratios of the preprocessed-only, AMICA and ASR cleaned data are displayed in blue, orange and gray, respectively. The mean cross-subject regular and elevated OG walking speeds were 1.37 and 1.81 m/s, respectively. X and y axes depict the various trials and band’s W/S ratio, respectively. For preprocessed data (blue), significance in the pairwise comparisons are indicated by asterisks (single asterisks denotes p<0.01 and double asterisks p<0.05). AMICA and ASR cleaned data showed no statistically significant power increase.
Fig 7.
Increased electro-physiological signals during higher speeds.
Presented are 10-second temporal segments of occipital neck muscle components, as separated by the AMICA algorithm (exemplary component map presented on the right), recorded during the various TM speeds of a single participant. Speeds increase from top to bottom, x and y axes depict time and amplitude [V], respectively.