Quasi-Brain-Death EEG Diagnosis Based on Tensor Train Decomposition

The quasi-brain-death diagnosis based on electroencephalogram (EEG) signal analysis is of great significance for early detection of quasi-brain-death patients which can avoid brain death misjudgment. Tensor is the multi-way array and tensor decomposition is a natural way to analyze high-order data. In this paper, we apply tensor train (TT) decomposition to EEG-based quasi-brain-death diagnosis. By reshaping the EEG data from matrix to higher-order tensor, we use a new algorithm to extract more valuable features from the data. The support vector machine (SVM) classifier is then used to complete the classification task of the extracted features. The experimental result shows that our method is well performed in evaluating the difference between coma patients and brain-death patients.