Tensor Based Singular Spectrum Analysis for Automatic Scoring of Sleep EEG

• Published in: IEEE transactions on neural systems and rehabilitation engineering Vol. 23; no. 1; pp. 1 - 9
• Main Authors: Kouchaki, Samaneh, Sanei, Saeid, Arbon, Emma L., Dijk, Derk-Jan
• Format: Journal Article
• Language: English
• Published: United States IEEE 01.01.2015; The Institute of Electrical and Electronics Engineers, Inc. (IEEE)
• Subjects: Adult; Algorithms; Automation; Channels; Computer Simulation; Decomposition; Electroencephalogram (EEG); Electroencephalography; Electroencephalography - methods; Electroencephalography - statistics & numerical data; empirical mode decomposition (EMD); Female; Humans; Male; Mathematical analysis; Narrowband; Noise; Rehabilitation; Scoring; Signal Processing, Computer-Assisted; single channel source separation; singular spectrum analysis (SSA); Sleep; Sleep - physiology; Source separation; Spectrum analysis; Subspaces; Tensile stress; tensor factorization; Tensors; Time series analysis
• ISSN: 1534-4320; 1558-0210; 1558-0210
• DOI: 10.1109/TNSRE.2014.2329557

A new supervised approach for decomposition of single channel signal mixtures is introduced in this paper. The performance of the traditional singular spectrum analysis algorithm is significantly improved by applying tensor decomposition instead of traditional singular value decomposition. As another contribution to this subspace analysis method, the inherent frequency diversity of the data has been effectively exploited to highlight the subspace of interest. As an important application, sleep electroencephalogram has been analyzed and the stages of sleep for the subjects in normal condition, with sleep restriction, and with sleep extension have been accurately estimated and compared with the results of sleep scoring by clinical experts.