TY - GEN
T1 - Automatic Sleep Spindle Detection and Analysis in Patients with Sleep Disorders
AU - Chen, Chao
AU - Zhu, Xuequan
AU - Belkacem, Abdelkader Nasreddine
AU - Lu, Lin
AU - Hao, Long
AU - You, Jia
AU - Shin, Duk
AU - Tan, Wenjun
AU - Huang, Zhaoyang
AU - Ming, Dong
N1 - Publisher Copyright:
© 2021, Springer Nature Singapore Pte Ltd.
PY - 2021
Y1 - 2021
N2 - Nowadays, Sleep disorder is a common disease, and spindle spindles are important features of the second stage non-rapid eye movement (NREM) sleep. In this paper, we propose an improved automatic detection method of spindles based on wavelet transform. The spindles automatic detector is mainly composed of wavelet transform and clustering. We collected the electroencephalography (EEG) signals of six patients with sleep disorders all night for ten hours, and then preprocessed the data and other operations, and then used our improved method to detect the sleep EEG signals by spindles. By comparing with the previous automatic detection method not improved and another automatic detection method, the results show that the accuracy of sleep spindles detection can be effectively improved. The accuracy of the improved detector is 5.19% higher than before, and 9.7% higher than that of another method based on amplitude threshold. Finally, we made a simple comparison between people with sleep disorders and normal people. We found that there were significant differences in spindle density between people with sleep disorders and people without sleep disorders. The average spindle density in the normal population averaged 2.59 spindles per minute. People with sleep disorders had an average spindle density of 1.32 spindles per minute. In future research, our research direction is to improve the accuracy of spindles automatic detection by improving the spindles detector and study the difference of spindles between patients with sleep disorders and normal people in a large number of samples, so that the difference of spindles can be used as the basis for the diagnosis of sleep disorders.
AB - Nowadays, Sleep disorder is a common disease, and spindle spindles are important features of the second stage non-rapid eye movement (NREM) sleep. In this paper, we propose an improved automatic detection method of spindles based on wavelet transform. The spindles automatic detector is mainly composed of wavelet transform and clustering. We collected the electroencephalography (EEG) signals of six patients with sleep disorders all night for ten hours, and then preprocessed the data and other operations, and then used our improved method to detect the sleep EEG signals by spindles. By comparing with the previous automatic detection method not improved and another automatic detection method, the results show that the accuracy of sleep spindles detection can be effectively improved. The accuracy of the improved detector is 5.19% higher than before, and 9.7% higher than that of another method based on amplitude threshold. Finally, we made a simple comparison between people with sleep disorders and normal people. We found that there were significant differences in spindle density between people with sleep disorders and people without sleep disorders. The average spindle density in the normal population averaged 2.59 spindles per minute. People with sleep disorders had an average spindle density of 1.32 spindles per minute. In future research, our research direction is to improve the accuracy of spindles automatic detection by improving the spindles detector and study the difference of spindles between patients with sleep disorders and normal people in a large number of samples, so that the difference of spindles can be used as the basis for the diagnosis of sleep disorders.
KW - Automatic detection
KW - EEG
KW - Sleep disorders
KW - Sleep spindles
KW - Wavelet transform and clustering
UR - http://www.scopus.com/inward/record.url?scp=85104699377&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85104699377&partnerID=8YFLogxK
U2 - 10.1007/978-981-16-1288-6_8
DO - 10.1007/978-981-16-1288-6_8
M3 - Conference contribution
AN - SCOPUS:85104699377
SN - 9789811612879
T3 - Communications in Computer and Information Science
SP - 113
EP - 124
BT - Human Brain and Artificial Intelligence - Second International Workshop, HBAI 2020, Held in Conjunction with IJCAI-PRICAI 2020, Revised Selected Papers
A2 - Wang, Yueming
PB - Springer Science and Business Media Deutschland GmbH
T2 - 2nd International Workshop on Human Brain and Artificial Intelligence, HBAI 2020 held in Conjunction with IJCAI-PRICAI 2020
Y2 - 7 January 2021 through 7 January 2021
ER -