The Application Of MVPA And Change-point Analysis Approaches In The Researches Of Acupuncture Mechanism

Acupuncture, as a treasure of the splendid traditional culture of Chinese nations, possesses massive health resources with characters and advantages in Chinese medical and health care undertakings. Because of the globally increasing medical expenditure and the iatrogenic and drug-induced diseases caused by modern medication in conquering hazardous diseases, the medical community and patients are becoming very interested in the economic, low toxic and side-effect acupuncture subsectioned in Traditional Chinese Medicine good at regulation. The traditional Chinese Medicine of acupuncture theories, growing out of practices and deeply processed by Chinese classical philosophy, make it hard to reconcile with modern medicines in many aspects, which sets up shackles for the development of acupuncture theories. With the advanced development of the modern neuroimaging techniques, particularly the functional magnetic resonance imaging (fMRI), the exploration of the relations among acupoints, brain and viscera has become possible. The fMRI technique, with its non-invasive and higher resolution characteristics, has been widely applied in the detection of the brain functions. However, the fMRI acupuncture investigation is still in its early stage, and needs further improvements. The study in this paper focused on the application of data-driven fMRI data analysis approaches in the researches of acupuncture mechanism. The details of the researches in this dissertation are listed as follows:The first experiment"Exploring vision-related acupoint specificity with multi-voxel pattern analysis". In the investigation of acupuncture mechanism, the acupoint specificity is still a controversial research topic. Here we applied general linear model-based approach and a data-driven approach, multi-voxel pattern analysis (MVPA), to analyze the difference of spatial neutral response patterns induced by acupuncture stimulation at GB37 and one of its nearby nonacupoint. The results derived from the analysis based on general linear model showed no significant differences. However, the results of MVPA demonstrated that using the information abstracted from each voxels, those two spatial patterns of neutral responses could be classified with really high accuracies. Therefore, MVPA approach is more optimal to be applied in the research of acupoint specificity.The second experiment"Investigation of temporal characteristics of neutral responses induced by acupuncture stimulation based on change-point theory". According to the theory of Traditional Chinese Medicine and a variety of clinical reports, the effect of acupuncture can sustain beyond the needle manipulation period. Therefore, acupuncture stimulations could induce complex dynamic neutral responses which change with time, and we should apply proper analysis approaches to investigate the temporal information of the central nervous system. Here we introduced one state-related analysis approach of fMRI timeseries, which is based on change-point theory,to learn about the temporal characteristics of neural responses to acupuncture. Results showed that the neutral responses of the same brain cortex area, induced by acupuncture stimulations at ST 36 and the nearby non-acupoint, had different temporal characteristics except for their spatial distributions and intensities. Based on the analysis we have done, this state-related fMRI analysis approach is suitable to investigate the temporal characteristics of the neutral responses induced by acupuncture stimulation.