An online EEG BCI based on covert visuospatial attention in absence of exogenous stimulation

Abstract

Objective. In this work we present-for the first time-the online operation of an electroencephalogram (EEG) brain-computer interface (BCI) system based on covert visuospatial attention (CVSA), without relying on any evoked responses. Electrophysiological correlates of pure top-down CVSA have only recently been proposed as a control signal for BCI. Such systems are expected to share the ease of use of stimulus-driven BCIs (e.g. P300, steady state visually evoked potential) with the autonomy afforded by decoding voluntary modulations of ongoing activity (e.g. motor imagery). Approach. Eight healthy subjects participated in the study. EEG signals were acquired with an active 64-channel system. The classification method was based on a time-dependent approach tuned to capture the most discriminant spectral features of the temporal evolution of attentional processes. The system was used by all subjects over two days without retraining, to verify its robustness and reliability. Main results. We report a mean online accuracy across the group of 70.6 ± 1.5%, and 88.8 ± 5.8% for the best subject. Half of the participants produced stable features over the entire duration of the study. Additionally, we explain drops in performance in subjects showing stable features in terms of known electrophysiological correlates of fatigue, suggesting the prospect of online monitoring of mental states in BCI systems. Significance. This work represents the first demonstration of the feasibility of an online EEG BCI based on CVSA. The results achieved suggest the CVSA BCI as a promising alternative to standard BCI modalities.

Modulation of Posterior Alpha Activity by Spatial Attention Allows for Controlling A Continuous Brain–Computer Interface

“…Here, we show that a simple template matching approach yields performance comparable to other studies, both in terms of classification rate as well as information transfer rate per subtrial. On trial level, our paradigm yielded an average performance of around 66 % in the last session, which is in between the offline studies by van Gerven et al van Gerven et al (2009) and Kelly et al (2005) and the online study by Tonin et al (2013). On subtrial level, Subject 1 obtained a maximal information transfer rate of 8.9 bits/min which is comparable to the maximum ITR found by Kelly et al (7.5 In both paradigms, cueing one side was followed by a change in visual stimulation at that side.…”

“…While these studies do suggest that modulations in alpha power can be used for BCI control, all studies relied on offline analyses after data acquisition. Recently, Tonin et al (2013) showed for the first time that covert attention brain-computer interfacing is possible using electronencephalography in an online setup.…”

“…In contrast to the paradigm of Tonin et al (2013), we present subjects with continuous feedback while performing the BCI task. In addition, we hypothesize that the strength of lateralized alpha power predicts the individual subjects' classification rate.…”

A Stimulus-Independent Hybrid BCI Based on Motor Imagery and Somatosensory Attentional Orientation

“…Contrary to CVSA based BCI [27], [28], which enabled covert visual attention decoding without visual stimulation and especially without overt eye movement, in the somatosensory modality without physical tactile stimulation, the imagined sensation task has also induced oscillatory activation in the somatosensory area of the brain and enabled covert somatosensory attention decoding without any sensory input. Furthermore, during the SAO tasks, both in the STE-Group and SAO-Group, the contralateral ERD was stronger than the ipsilateral ERD, which is consistent with motor imagery studies (mentally simulated efferent output) [41] and tactile BCI based selective sensation (afferent input processing) [20].…”

“…the focus of attention on different regions of the visual field without overt eye movements, it has been observed that selective activation of the parieto-occipital α oscillation reflects endogenous shifts in the locus of visual attention [27]. Moreover the endogenous brain activation from the CVSA task has been shown to allow construction of an independent BCI without visual stimulation and especially without overt eye movement [28], [29]. Therefore, it is possible that covert visual attention is decoded even without the presence of the visual stimulus.…”