On the Robustness of EEG Tensor Completion Methods

FENG DUAN; HAO JIA; ZHIWEN ZHANG; FAN FENG; YING TAN; YANGYANG DAI; ANDRZEJ CICHOCKI; ZHENGLU YANG; CESAR F. CAIAFA; SUN ZHE; JORDI SOLE-CASALS

SCIENCE CHINA Technological Sciences

Springer

Lugar: Beijin; Año: 2021

During the acquisition of electroencephalographic (EEG) signals, datamay be missing or corrupted by noise and artifacts. To reconstruct the incomplete EEG signals, we rst convert EEG signals into a three-order tensor (multi-dimensional data) of shape time channel trial. Then, the missing data can be eciently recovered by applying a tensor completion method (TCM). However, there is not a unique way to organize channels and trials in a tensor, and dierent numbers of channels are available depending on the EEG setting used, which may aect the quality of the tensor completion results. The main goal of this paper is to evaluate the robustness of EEG completion methods with several designed parameters such as the ordering ofchannels and trials, the number of channels and the amount of missing data. In this work, we compare the results of completing missing (or corrupted) data in EEG signals by using several available TCMs. To emulate dierent scenarios of missing data, three dierent patterns of missing data were designed. Firstly, the amount of missing data on completion eects was analysed, including the time lengths of missing data and the number of channels or trials aected by missing data. Then, the numerical stability of the completion methods was analysed by shuing the indices along channels or trials in the EEG data tensor. Finally, the way the number of electrodes of EEG tensors in uences completion eects was assessed by changing the number of channels. Among all the applied TCMs, the Simultaneous Tensor Decomposition and Completion(STDC) method achieves the best performance in providing stable results when the amount of missing data or the electrode number of EEG tensors is changed. In other words, STDC proves to be an excellent choice of TCM, since permutations of trials or channels have almost no in uence on the completion results. The STDC method can eciently complete the missing EEG signals. The designed simulations can be regarded as a procedure to validate whether or not a completion method is useful enough to complete EEG signals.