Phase-Spatial Beamforming Renders a Visual Brain Computer Interface Capable of Exploiting EEG Electrode Phase Shifts in Motion-Onset Target Responses

Abstract

Objective: in this work, we aim to develop a more efficient visual motion-onset based Brain-computer interface (BCI). Brain-computer interfaces provide communication facilities that do not rely on the brain's usual pathways. Visual BCIs are based on changes in EEG activity in response to attended flashing or flickering targets. A less taxing way to encode such targets is with briefly moving stimuli, the onset of which elicits a lateralized EEG potential over the parieto-occipital scalp area called the motion-onset visual evoked potential (mVEP). Methods: We recruited 21 healthy subjects for an experiment in which motion-onset stimulations translating leftwards (LT) or rightwards (RT) were encoding 9 displayed targets.We propose a novel algorithm that exploits the phase-shift between EEG electrodes to improve target decoding performance. We hereto extend the spatiotemporal beamformer (stBF) with a phase extracting procedure, leading to the phase-spatial beamformer (psBF). Results: we show that psBF performs significantly better than the stBF (p < 0.001 for 1 and 2 stimulus repetitions and p < 0.01 for 3 to 5 stimulus repetitions), as well as the previously validated linear support-vector machines (p < 0.001 for 5 stimulus

A novel brain-computer interface based on audio-assisted visual evoked EEG and spatial-temporal attention CNN

“…But there are 5 characters that need to be traversed one by one in the second stage, which would lead to a decrease in presentation efficiency. To improve the efficiency of the BCI paradigm, we analyze further possible strategies, including novel paradigms to enhance the EEG features, such as the leftwards or rightwards motion-onset stimuli translating (Libert et al, 2022b) and the two-dimensional auditory stimuli with both pitch (high/medium/low) and direction (left/middle/right) (Hohne et al, 2011), and the stable classification algorithm of ERP for cross subjects or scenarios, such as the analytic beamformer transformation (Libert et al, 2022a), ternary classification method (Zhang et al, 2021) and some transfer learning methods.…”

“…In sum, there are still some current challenges to the application of the EEG-based BCI, including the friendly cognitive load and EEG characteristics-guided BCI classification algorithms (Xu et al, 2021). Compared with the flashing or flickering visual BCIs, the mVEP is a convenient way to encode targets with briefly moving stimuli (Libert et al, 2022b). On single trial classification, CNN can achieve comparable performance to both the LDA and support vector machine, but slightly less stable and interpretable (Vareka, 2020).…”

Analytic beamformer transformation for transfer learning in motion-onset visual evoked potential decoding

“…The motion-onset data stemmed from a previous experiment [34] and is therefore only briefly described. Subjects were seated approximately 70 cm from the screen.…”

“…To complete our investigation, we checked whether one of the chosen decoders benefitted more from one of the data representations. Here we have excluded the comparison with the spatiotemporal epochs as we have already reported on this in previous work [34,48]. Boxplots of the results can be found in figure 4 for the decoding of individual datasets and figure 5 for population-trained decoders.…”

“…In contrast to the aforementioned paradigms, which rely on sudden or repeatedly flickering stimuli, mVEPs are evoked in response to the onset of a moving non-flickering stimulus, which is reported to more comfortable to the user [29][30][31]. Motion direction has been shown to impact on the recorded mVEP responses [27,32,33] and the observed phase difference between EEG channels could be exploited to achieve a higher communication rate [34].…”