Table of Contents Author Guidelines Submit a Manuscript
Computational Intelligence and Neuroscience
Volume 2017, Article ID 4574079, 8 pages
Research Article

A Comparison Study on Multidomain EEG Features for Sleep Stage Classification

1Key Laboratory of Advanced Control and Optimization for Chemical Processes, East China University of Science and Technology, Ministry of Education, Shanghai, China
2Department of Automation, School of Information Science and Engineering, East China University of Science and Technology, Shanghai, China
3Research Institute of Systems Control, Institute for Advanced Research and Education, Saga, Japan

Correspondence should be addressed to Bei Wang; nc.ude.tsuce@gnawieb

Received 10 July 2017; Revised 22 September 2017; Accepted 11 October 2017; Published 5 November 2017

Academic Editor: Pasi A. Karjalainen

Copyright © 2017 Yu Zhang et al. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.


Feature extraction from physiological signals of EEG (electroencephalogram) is an essential part for sleep staging. In this study, multidomain feature extraction was investigated based on time domain analysis, nonlinear analysis, and frequency domain analysis. Unlike the traditional feature calculation in time domain, a sequence merging method was developed as a preprocessing procedure. The objective is to eliminate the clutter waveform and highlight the characteristic waveform for further analysis. The numbers of the characteristic activities were extracted as the features from time domain. The contributions of features from different domains to the sleep stages were compared. The effectiveness was further analyzed by automatic sleep stage classification and compared with the visual inspection. The overnight clinical sleep EEG recordings of 3 patients after the treatment of Continuous Positive Airway Pressure (CPAP) were tested. The obtained results showed that the developed method can highlight the characteristic activity which is useful for both automatic sleep staging and visual inspection. Furthermore, it can be a training tool for better understanding the appearance of characteristic waveforms from raw sleep EEG which is mixed and complex in time domain.