1. School of Computer and Information Technology, Beijing Jiaotong University, Beijing 100044, China 2. Beijing Key Laboratory of Traffic Data Analysis and Mining, Beijing Jiaotong University, Beijing 100044, China 3. Key Laboratory of Intelligent Passenger Service of Civil Aviation, CAAC, Beijing 100105, China
A deep convolutional neural network model was proposed aiming at the problem that the current data and features determine the upper limit of the classification accuracy of the sleep staging model. The parallel convolutional neural network automatically learns the time-domain and frequency-domain features of the original signals in terms of model construction. The feature fusion neural network fuses multi-features through dilated convolution and residual connection. The classification neural network recognizes the sleep stages based on fused features. Synthetic minority oversampling technique (SMOTE) method was applied to enhance data in order to reduce the effect of classification imbalance on classification effect, and two-step training method was applied to optimize the model. The original single-channel electroencephalogram (Fpz-Cz channel) of the Sleep-EDF data set was used to evaluate the proposed model by the 20-fold cross-validate scheme. The overall accuracy and macro-averaging F1-score were 86.73% and 81.70% respectively. The proposed deep convolution neural network was an end-to-end deep learning model without any prior knowledge. The experimental results showed that the classification accuracy of the proposed model was better than traditional deep learning models.
Tab.2The best structure and corresponding parameters of pretrained model
分期
W
N1
N2
N3
REM
PR/%
RE/%
F1/%
W
7295
271
131
37
193
88.67
92.03
90.32
N1
369
1396
606
16
417
65.57
49.79
56.60
N2
378
283
15582
853
703
91.21
87.54
89.34
N3
33
3
270
5397
0
85.53
94.63
89.85
REM
152
176
495
7
6887
83.99
89.24
86.54
Tab.3Confusion matrix and various evaluation indicators based on deep convolutional neural network model
Fig.4Comparison of expert manual sleep staging and model automatic sleep staging
%
模型
ACC
MF1
F1
W
N1
N2
N3
REM
文献[9]模型
78.9
73.7
71.6
47.0
84.6
84.0
81.4
文献[11]模型
74.8
69.8
65.4
43.7
80.6
84.9
74.5
文献[12]模型
82.0
76.9
84.7
46.6
85.9
84.8
82.4
文献[13]模型
84.3
79.7
89.2
52.2
86.8
85.1
85.0
提出模型
87.1
82.5
90.3
56.6
89.3
89.9
86.5
Tab.4Comparison of advanced research results with results of proposed model
[1]
WULFF K, GATTI S, WETTSTEIN J G, et al Sleep and circadian rhythm disruption in psychiatric and neurodegenerative disease[J]. Nature Reviews Neuroscience, 2010, 11 (8): 589- 599
doi: 10.1038/nrn2868
[2]
HOBSON J A A manual of standardized terminology, techniques and scoring system for sleep stages of human subjects[J]. JAMA Psychiatry, 1969, 20 (2): 246
[3]
BERRY R B, BROOKS R, GAMALDO C E, et al. The AASM manual for the scoring of sleep and associated events [M]. Darien, Illinois: American Academy of Sleep Medicine, 2012.
[4]
SORS A, BONNET S, MIREK S, et al A convolutional neural network for sleep stage scoring from raw single-channel EEG[J]. Biomedical Signal Processing and Control, 2018, 42: 107- 114
doi: 10.1016/j.bspc.2017.12.001
[5]
LACHNER-PIZA D, EPITASHVILI N, SCHULZE-BONHAGE A, et al A single channel sleep-spindle detector based on multivariate classification of EEG epochs: MUSSDET[J]. Journal of Neuroscience Methods, 2018, 297: 31- 43
doi: 10.1016/j.jneumeth.2017.12.023
[6]
王卫星, 孙守迁, 李超, 等 基于卷积神经网络的脑电信号上肢运动意图识别[J]. 浙江大学学报: 工学版, 2017, 51 (7): 1381- 1389 WANG Wei-xing, SUN Shou-qian, LI Chao, et al Recognition of upper limb motion intention of EEG signal based on convolutional neural network[J]. Journal of Zhejiang University: Engineering Science, 2017, 51 (7): 1381- 1389
[7]
杨帮华, 何美燕, 刘丽, 等 脑机接口中基于BISVM的EEG分类[J]. 浙江大学学报: 工学版, 2013, 47 (8): 1431- 1436 YANG Bang-hua, HE Mei-yan, LIU Li, et al EEG classification based on batch incremental SVM in brain computer interfaces[J]. Journal of Zhejiang University: Engineering Science, 2013, 47 (8): 1431- 1436
[8]
DA S, THIAGO L T, KOZAKEVICIUS A J, et al Single-channel EEG sleep stage classification based on a streamlined set of statistical features in wavelet domain[J]. Medical and Biological Engineering and Computing, 2017, 55 (2): 343
doi: 10.1007/s11517-016-1519-4
[9]
ORESTIS T, MATTHEWS P M, GUO Y Automatic sleep stage scoring using time-frequency analysis and stacked sparse autoencoders[J]. Annals of Biomedical Engineering, 2016, 44 (5): 1587
doi: 10.1007/s10439-015-1444-y
[10]
PRUCNAL M, POLAK A G Effect of feature extraction on automatic sleep stage classification by artificial neural network[J]. Metrology and Measurement Systems, 2017, 24 (2): 229
doi: 10.1515/mms-2017-0036
[11]
ORESTIS T, MATTHEWS P M, GUO Y, et al. Automatic sleep stage scoring with single-channel EEG using convolutional neural networks [EB/OL]. (2016-10-05). https://arxiv.org/abs/1610.01683.
[12]
SUPRATAK A, DONG H, WU C, et al DeepSleepNet: a model for automatic sleep stage scoring based on raw single-channel EEG[J]. IEEE Transactions on Neural Systems and Rehabilitation Engineering, 2017, 25 (11): 1998- 2008
doi: 10.1109/TNSRE.2017.2721116
[13]
MOUSAVI S, AFGHAH F, ACHARYA U R SleepEEGNet: automated sleep stage scoring with sequence to sequence deep learning approach[J]. PloS One, 2019, 14 (5): e0216456
doi: 10.1371/journal.pone.0216456
[14]
HE K, ZHANG X, REN S, et al. Deep residual learning for image recognition [C]// Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition. Las Vegas: IEEE, 2016: 770-778.
[15]
CHAWLA N V, BOWYER K W, HALL L O, et al SMOTE: synthetic minority over-sampling technique[J]. Journal of Artificial Intelligence Research, 2002, 16: 321- 357
doi: 10.1613/jair.953
[16]
COHEN M X. Analyzing neural time series data: theory and practice [M]. Cambridge: MIT Press, 2014.
[17]
YANG J, YANG J Y, ZHANG D, et al Feature fusion: parallel strategy vs. serial strategy[J]. Pattern Recognition, 2003, 36 (6): 1369- 1381
doi: 10.1016/S0031-3203(02)00262-5
[18]
JANG E, GU S, POOLE B. Categorical reparameterization with gumbel-softmax [EB/OL]. (2017-08-05). https://arxiv.org/abs/1611.01144.
[19]
SRIVASTAVA N, HINTON G, KRIZHEVSKY A, et al Dropout: a simple way to prevent neural networks from overfitting[J]. Journal of Machine Learning Research, 2014, 15 (1): 1929- 1958
[20]
GOLDBERGER A L, AMARAL L A N, GlASS L, et al PhysioBank, PhysioToolkit, and PhysioNet: components of a new research resource for complex physiologic signals[J]. Circulation: Journal of the American Heart Association, 2000, 101 (23): e215- e220
[21]
KEMP B, ZWINDERMAN A H, TUK B, et al Analysis of a sleep-dependent neuronal feedback loop: the slow-wave microcontinuity of the EEG[J]. IEEE Transactions on Biomedical Engineering, 2000, 47 (9): 1185- 1194
doi: 10.1109/10.867928
[22]
金欢欢, 尹海波, 何玲娜 端到端单通道睡眠EEG自动分期模型[J]. 计算机科学, 2019, 46 (3): 242- 247 JIN Huan-huan, YIN Hai-bo, HE Ling-na End-to-end single-channel automatic staging model for sleep EEG signal[J]. Computer Science, 2019, 46 (3): 242- 247
