Reconstruction of missing channel in electroencephalogram using spatiotemporal correlation-based averaging

Abstract

Objective. Electroencephalogram (EEG) recordings often contain large segments with missing signals due to poor electrode contact or other artifact contamination. Recovering missing values, contaminated segments and lost channels could be highly beneficial, especially for automatic classification algorithms, such as machine/deep learning models, whose performance relies heavily on high-quality data. The current study proposes a new method for recovering missing segments in EEG. Approach. In the proposed method, the reconstructed segment is estimated by substitution of the missing part of the signal with the normalized weighted sum of other channels. The weighting process is based on inter-channel correlation of the non-missing preceding and proceeding temporal windows. The algorithm was designed to be computationally efficient. Experimental data from patients (N = 20) undergoing general anesthesia due to elective surgery were used for the validation of the algorithm. The data were recorded using a portable EEG device with ten channels and a self-adhesive frontal electrode during induction of anesthesia with propofol from waking state until burst suppression level, containing lots of variation in both amplitude and frequency properties. The proposed imputation technique was compared with another simple-structure technique. Distance correlation (DC) was used as a measure of comparison evaluation. Main results.:The proposed method with average distance correlation of 82.48±10.01 (µ ± σ)% outperformed its competitor with average distance correlation of 67.89±14.12 (µ ± σ)% . This algorithm also showed better performance for an increasing number of missing channels. Significance. the proposed technique provides an easy-to-implement and computationally efficient approach for the reliable reconstruction of missing or contaminated EEG segments.

Assigning channel weights using an attention mechanism: an EEG interpolation algorithm

“…Visualizing the reconstructed data and the real data, we observed that at most points the reconstructed data were close to the real values. However, when several bad channels are close to each other, the interpolation algorithm fits poorly at points with higher amplitudes, and a similar phenomenon was observed in the Bahador et al (2021) and Li et al (2022) experiments. This part of the data points causes a large reconstruction error.…”

“… Bahador et al (2021) proposed an interpolation method that consists of two steps. (1) The data is sliced into smaller segments in the temporal dimension and each segment is interpolated using good channel data.…”

Channel-independent recreation of artefactual signals in chronically recorded local field potentials using machine learning

“…Inter-channel interpolation When a channel in an array is impacted locally by an artefact, that segment can be replaced using the average or other methods that take into consideration the surrounding channels, which isn’t possible in a single channel approach [ 4 ].…”

iTa-DFiE: An Innovative Tensor Algebra-Based Detection Framework for Incomplete Noninvasive Electroencephalography

“…For most mathematical models in EEG-based BCI system applying for diagnosis, prediction and treatment, a complete dataset is imperative. Unfortunately, the real-world data from non-invasive EEG signals can be missing during the data acquisition process due to manual mistakes, hardware failure or software failure [16][17][18][19][20]. More worrisome, the phenomenon of missing observations dramatically causes inaccurate results that would lead to an unreliable diagnosis, or even a reduction in the performance of methodologies.…”

“…Performing missing data imputation on non-invasive EEG time series data has become a fundamental step before prosing further usage. Unlike most existing studies solves missing observations which have only considered original flat-view matrix or vector models [17], [20][21][22][23], the paper focus here is to deal with occlusion in tensorial structures without destroying inherent multiway nature of the EEG data. Ideally, the paper proposes a comprehensive model for tensor completion of EEG-based BCIs system by decoding the three main properties of brain signals, namely noise, temporal dynamics, and tensor representation simultaneously.…”