Latent Characteristics and Neural Manifold of Brain Functional Network under Acupuncture
Kai Li,Jiang Wang,Shanshan Li,Bin Deng,Haitao Yu
DOI: https://doi.org/10.1109/tnsre.2022.3157380
IF: 4.9
2022-01-01
IEEE Transactions on Neural Systems and Rehabilitation Engineering
Abstract:Acupuncture can regulate the cognition of brain system, and different manipulations are the keys of realizing the curative effect of acupuncture on human body. Therefore, it is crucial to distinguish and monitor the different acupuncture manipulations automatically. In this brief, in order to enhance the robustness of electroencephalogram (EEG) detection against noise and interference, we propose an acupuncture manipulation detecting framework based on supervised ISOMAP and recurrent neural network (RNN). Primarily, the low-dimensional embedding neural manifold of brain dynamical functional network is extracted via the reconstructed geodetic distance. It is found that there exhibits stronger acupuncture-specific reconfiguration of brain network. Besides, we show that the distance travel along this manifold correlates strongly with changes of acupuncture manipulations. The low-dimensional brain topological structure of all subjects shows crescent-like feature when acupuncturing at Zusanli acupoints, and fixed-points are varying under diverse manipulation methods. Moreover, Takagi-Sugeno-Kang (TSK) classifier is adopted to identify acupuncture manipulations according to the nonlinear characteristics of neural manifolds. Compared with different classifier, TSK can further improve the accuracy of manipulation identification at 96.71%. The results demonstrate the effectiveness of our model in detecting the acupuncture manipulations, which may provide neural biomarkers for acupuncture physicians.