Research i_need_contribute
Using Nonlinear Dynamics and Multivariate Statistics to Analyze EEG Signals of Insomniacs
source:NCBI 2020-12-08 [Research]
with the Intervention of Superficial Acupuncture

Shi-Yi Qi, 1 Dong Lin, 1 Li-Li Lin, 1 Xiao-Zhen Huang, 2 Shen Lin, 1 Yun-Ying Yu, 3 Chuan-Hai Cao, 4 and Zhi-Xin Wang5

Author information Article notes Copyright and License information Disclaimer

1College of Acupuncture, Fujian University of Traditional Chinese Medicine, Fuzhou, Fujian Province, China

2Department of Traditional Chinese Medicine Rehabilitation, Anxi County Hospital, Quanzhou, Fujian Province, China

3Department of Sleep Medicine, Rehabilitation Hospital Affiliated to Fujian University of Traditional Chinese Medicine, Fuzhou, Fujian Province, China

4Department of Neurosurgery and Brain Repair, Morsani College of Medicine, University of South Florida, Tampa, FL, USA

5Department of Internal Medicine, Florida Orthopedic Institute, University of South Florida, Tampa, FL, USA

Dong Lin: moc.anis@1ekileulb

Academic Editor: Chan-Yen Kuo

 

 

Abstract

Objective

As a noninvasive and nonpharmacological therapeutic approach, superficial acupuncture (SA) is a special method of acupuncture. In this study, using nonlinear dynamics and multivariate statistics, we studied the electroencephalography (EEG) of primary insomnia under SA intervention to investigate how brain regions change.

Method

This study included 30 adults with primary insomnia. They underwent superficial acupuncture at the Shangen acupoint. The EEG signals were collected for 10 minutes at each state, including the resting state, the intervention state, and the postintervention state. The data were conducted using nonlinear dynamics (including approximate entropy (ApEn) and correlation dimension (CD)) and multivariate statistics.

Result

The repeated-measures ANOVA results showed that both ApEn and CD values were not significantly different at the three states (p > 0.05). The paired t-test results showed that the ApEn values of electrodes O2 (the right occipital lobe) at the postintervention state have decreased, compared with the resting state (p < 0.05), and no difference was detected in CD (p > 0.05). The cluster analysis results of ApEn showed that patients' EEG has changed from the right prefrontal lobe (electrode Fp2) to the right posterior temporal lobe (electrode T6) and finally to the right occipital lobe (electrode O2), before, during, and after the SA intervention. In addition, the factor analysis results of CD revealed that patients' EEG of all brain regions except for the occipital lobes has changed to the frontal lobes and anterior temporal and frontal lobes from pre- to postintervention.

Conclusion

SA activated the corresponding brain regions and reduced the complexity of the brain involved. It is feasible to use nonlinear dynamics analysis and multivariate statistics to examine the effects of SA on the human brain.