Single trial recognition of anticipatory slow cortical potentials: The role of spatio-spectral filtering

Abstract

Abstract-Single trial recognition of slow cortical potentials (SCPs) from full-band EEG (FbEEG) faces different challenges to classical EEG such as noisy, high magnitude (∼ ±100 µV ) infra slow oscillations (ISO) with f ≤ 0.1 Hz and high frequency spatial noise from a variety of artifacts. We analyze offline the anticipation related SCPs recorded from 11 subjects over two days in a variation of the Contingent Negative Variation (CNV) paradigm with Go and No-go conditions in an assistive technology framework. The results suggest that widely used spatial filters such as Common Average Referencing (CAR) and Laplacian are sub-optimal for the single trial analysis of SCPs. We show that a spatial smoothing filter (SSF), which in combination with CAR enhances the spatially distributed SCP while attenuating high frequency spatial noise. We report, first, that a narrow band filter in the range [0.1 1] Hz captures anticipation related SCP better and effectively reduces ISOs. Second, the SSF in combination with CAR outperforms CARalone and Laplacian spatial filters. Third, we compare linear and quadratic classifiers calculated using optimally filtered Cz electrode potentials and report that the best methods resulted in single trial classification accuracies of 83±4%, where classifiers were trained on day 1 and tested using data from day 2, to ensure generalization capabilities across days (1-7 days).

Detection of anticipatory brain potentials during car driving

“…around the onset of the '1' on the screen, and it peaks at 0s which is the time of the appearance of 'Go/Stop' on the screen and back to zero after around 250ms later. This is consistent with the CNV potential reported in previous studies [6], [9], [12]. In addition, a clear difference can be observed between Go and Nogo epochs (increasing negativity for Go and almost flat or slightly positive response for all the other No-go epochs).…”

“…Experiments with 6 subjects, show event-related potentials consistent with CNV signals reported in the literature [7], [8], [12]. Detection The results presented in the current work are promising since previous studies of anticipation-related potentials use simpler experimental setups and stimuli [6], [7].…”

“…This number of features have been reported to sufficiently represent the evolution of the CNV potentials in a 1s window [12].…”

“…EEG was preprocessed using common average reference (CAR) [11] and then filtered by using bandpass Butterworth filter between 0.1 to 1 Hz, due to the frequency range of CNV potentials [12]. EMG signals were filtered with a bandpass Butterworth filter in the range of 20 to 50 Hz and smoothed with a moving average filter (time window = 25 samples).…”

“…This signal, which typically lasts from about 300 ms to several seconds with magnitudes up to 50µV has been termed contingent negative variation (CNV) potential [6].This potential has been linked to the preparatory processing required for appropriate actions at the arrival of future events [7], [8]. Nevertheless, apart from few examples [9], [10], [12], these potentials have been studied using simple procols and stimlui.…”