A Deep Shared Multi-Scale Inception Network Enables Accurate Neonatal Quiet Sleep Detection With Limited EEG Channels

被引:24
作者
Ansari, Amir H. [1 ]
Pillay, Kirubin [2 ]
Dereymaeker, Anneleen [3 ]
Jansen, Katrien [3 ,4 ]
Van Huffel, Sabine [1 ]
Naulaers, Gunnar [3 ]
De Vos, Maarten [4 ,5 ]
机构
[1] Katholieke Univ Leuven, Dept Elect Engn ESAT, B-3000 Leuven, Belgium
[2] Univ Oxford, John Radcliffe Hosp, Dept Paediat, Oxford OX1 2JD, England
[3] Katholieke Univ Leuven, Univ Hosp Leuven, Neonatal Intens Care Unit, Dept Dev & Regenerat, B-3000 Leuven, Belgium
[4] Katholieke Univ Leuven, Univ Hosp Leuven, Dept Dev & Regenerat, B-3000 Leuven, Belgium
[5] Katholieke Univ Leuven, STADIUS, Dept Elect Engn ESAT, B-3000 Leuven, Belgium
基金
英国惠康基金;
关键词
Electroencephalography; Pediatrics; Sleep; Feature extraction; Radio frequency; Brain modeling; Monitoring; Convolutional Neural Networks; CNN; Inception networks; multi-scale EEG analysis; sleep stage classification; neonatal EEG analysis; CLASSIFICATION;
D O I
10.1109/JBHI.2021.3101117
中图分类号
TP [自动化技术、计算机技术];
学科分类号
0812 ;
摘要
In this paper, we introduce a new variation of the Convolutional Neural Network Inception block, called Sinc, for sleep stage classification in premature newborn babies using electroencephalogram (EEG). In practice, there are many medical centres where only a limited number of EEG channels are recorded. Existing automated algorithms mainly use multi-channel EEGs which perform poorly when fewer numbers of channels are available. The proposed Sinc utilizes multi-scale analysis to place emphasis on the temporal EEG information to be less dependent on the number of EEG channels. In Sinc, we increase the receptive fields through Inception while by additionally sharing the filters that have similar receptive fields, overfitting is controlled and the number of trainable parameters dramatically reduced. To train and test this model, 96 longitudinal EEG recordings from 26 premature infants are used. The Sinc-based model significantly outperforms state-of-the-art neonatal quiet sleep detection algorithms, with mean Kappa 0.77 +/- 0.01 (with 8-channel EEG) and 0.75 +/- 0.01 (with a single bipolar channel EEG). This is the first study using Inception-based networks for EEG analysis that utilizes filter sharing to improve efficiency and trainability. The suggested network can successfully detect quiet sleep stages with even a single EEG channel making it more practical especially in the hospital setting where cerebral function monitoring is predominantly used.
引用
收藏
页码:1023 / 1033
页数:11
相关论文
共 46 条
[1]  
Alonso J.-M., 2009, Scholarpedia, V4, P5393, DOI DOI 10.4249/SCHOLARPEDIA.5393
[2]  
[Anonymous], 2016, P 4 INT C LEARNING R
[3]  
[Anonymous], 2007, P CIKM
[4]   Improved multi-stage neonatal seizure detection using a heuristic classifier and a data-driven post-processor [J].
Ansari, A. H. ;
Cherian, P. J. ;
Dereymaeker, A. ;
Matic, V. ;
Jansen, K. ;
De Wispelaere, L. ;
Dielman, C. ;
Vervisch, J. ;
Swarte, R. M. ;
Govaert, P. ;
Naulaers, G. ;
De Vos, M. ;
Van Huffel, S. .
CLINICAL NEUROPHYSIOLOGY, 2016, 127 (09) :3014-3024
[5]   A convolutional neural network outperforming state-of-the-art sleep staging algorithms for both preterm and term infants [J].
Ansari, Amir H. ;
De Wel, Ofelie ;
Pillay, Kirubin ;
Dereymaeker, Anneleen ;
Jansen, Katrien ;
Van Huffel, Sabine ;
Naulaers, Gunnar ;
De Vos, Maarten .
JOURNAL OF NEURAL ENGINEERING, 2020, 17 (01)
[6]   Neonatal Seizure Detection Using Deep Convolutional Neural Networks [J].
Ansari, Amir H. ;
Cherian, Perumpillichira J. ;
Caicedo, Alexander ;
Naulaers, Gunnar ;
De Vos, Maarten ;
Van Huffel, Sabine .
INTERNATIONAL JOURNAL OF NEURAL SYSTEMS, 2019, 29 (04)
[7]   Quiet sleep detection in preterm infants using deep convolutional neural networks [J].
Ansari, Amir Hossein ;
De Wel, Ofelie ;
Lavanga, Mario ;
Caicedo, Alexander ;
Dereymaeker, Anneleen ;
Jansen, Katrien ;
Vervisch, Jan ;
De Vos, Maarten ;
Naulaers, Gunnar ;
Van Huffel, Sabine .
JOURNAL OF NEURAL ENGINEERING, 2018, 15 (06)
[8]  
Baldassarre F., 2017, ARXIV171203400CS
[9]   Classification of myocardial infarction with multi-lead ECG signals and deep CNN [J].
Baloglu, Ulas Baran ;
Talo, Muhammed ;
Yildirim, Ozal ;
Tan, Ru San ;
Acharya, U. Rajendra .
PATTERN RECOGNITION LETTERS, 2019, 122 :23-30
[10]   COMPUTER CHARACTERIZATION OF TRACE ALTERNANT AND REM-SLEEP PATTERNS IN THE NEONATAL EEG BY ADAPTIVE SEGMENTATION - AN EXPLORATORY-STUDY [J].
BARLOW, JS .
ELECTROENCEPHALOGRAPHY AND CLINICAL NEUROPHYSIOLOGY, 1985, 60 (02) :163-173