Removal of EEG artifacts for BCI applications using fully Bayesian tensor completion

Abstract

High accuracy of electroencephalogram (EEG) classification can hardly be achieved if the signals are contaminated by severe artefacts. One helpless way to avoid such artefacts is usually to directly discard the severely disturbed EEG segments. This study considers a more elegant way that tries to recover the disturbed segments from other undisturbed segments. The possible artefacts in EEG are treated as missing values. A Bayesian tensor factorization (BTF) based method is proposed to implement EEG completion for artefact removal. By specifying a sparsity-inducing hierarchical prior, the underlying low-rank tensor is discovered from incomplete EEG tensor with automatically inferred model parameters. The EEG missing values are effectively predicted with robustness to overfitting. Effectiveness of the BTF algorithm is demonstrated on EEG data recorded from seven subjects in a brain-computer interface paradigm based on event-related potentials.

Deriving Electrophysiological Brain Network Connectivity via Tensor Component Analysis During Freely Listening to Music

“…3). One of them is dominated by low-frequency oscillations (Delta and Theta rhythms, 3-8 Hz) and another is centered at Beta rhythm (20)(21)(22)(23)(24)(25)(26)(27)(28)(29)(30). The regions involved by the two anterior higher-order cognitive networks are part of the default mode network (DMN), which here contains temporal poles, the ventromedial prefrontal cortex and posterior cingulate cortex.…”

“…Tensor component analysis (TCA), as a well-established tool for signal processing and machine learning [18]- [20], This work is licensed under a Creative Commons Attribution 4.0 License. For more information, see http://creativecommons.org/licenses/by/4.0/ has shown to be powerful for neuroimaging data processing and analysis in cognitive neuroscience [21]- [26]. A tensor is a multi-dimensional representation of data or a multi-way array.…”

Classification Analysis of Tensor-Based Recovered Missing EEG Data

“…Moreover, In [39] classification accuracy is improved by selecting the channels on the basis of their inter-correlation. To achieve high classification accuracy severe artefacts have been removed from EEG in [40] by employing fully Bayesian tensor completion. To improve computational efficiency of tensor factorization Fast nonnegative tensor factorization method has been proposed in [41] which achieves higher computational efficiency.…”

Single-channel EEG completion using Cascade Transformer

“…Tensor completion methods (TCMs) have been proposed in several literature to perform the EEG completion by treating the recorded EEG as multi-channel tensor [7][8][9][10]. [9] demonstrated that the simultaneous tensor decomposition and completion (STDC) could achieve a better and more robust performance among several TCMs.…”