A Study On Brain Plasticity Of Athletes Based On Auto-regressive Time-frequency Analysis

Since the development of neuroscience,brain plasticity has been the focus of brain science research.Athletes provide a good object for the study of brain plasticity because of the long training experience.Previous analysis of FMRI data has focused on the frequency range of 0.01 to 0.08 Hz,which is considered to be the frequency band for signals of spontaneous brain activity.But filtering in a uniform way may hide information carried by signals at other frequencies.Based on the time-frequency analysis of autoregressive time series and graph theory,this thesis describes the information transfer patterns of brain in time-frequency scale.The main work is as follows:Firstly,we established a resting state brain functional connectivity network model in the range of 0.01 to 0.08 Hz by using graph theory analysis method,and conducted large-scale brain network analysis on student athletes and general control groups by using graph theory network index of whole brain functional connectivity density.The results showed that compared with the normal control group,the functional connectivity density of the left inferior frontal gyrus and middle frontal gyrus decreased in the student athletes group.The faster the executive control response time,the lower the functional connectivity density of the whole brain in the inferior frontal gyrus of the triangle.This is the representation of brain plasticity caused by the training of fast ball motor skills,indicating that long-term sports training makes athletes' brains show higher attention motor regulation ability and neural efficiency of executive control.Secondly,in order to explore the frequency specificity of brain plasticity information,we constructed brain functional network maps at different frequency scales based on wavelet transform,and analyzed the constructed brain functional network using graph theory method.The results show that within our frequency range,the small-world attribute gradually increases with the decrease of frequency,indicating that brain information mainly exists in low-frequency signals.Different frequency bands of brain networks show different network characteristics,which shows that it is very necessary to study the frequency division of brain f MRI signals.At the same time,it was found that the effect of exercise training on brain plasticity was more obvious in low frequency signals.Finally,to solve the causal network modeling of non-stationary time series process,granger causality models based on empirical mode decomposition were established in different frequency bands and applied to the functional image data of alerting attention task to explore the plasticity of athletes' brain attention network in different training stages.The results showed that there were significant differences in the intensity of information transmission from pectin putamen to anterior cingulate and paracingulate gyrus among the professional group,the student group and the normal group at the third intrinsic mode function(approximately slow-5:0.01-0.027 Hz),and the intensity of node effect was significantly negatively correlated with the response time of the alert task.It indicates that the stronger the connection between these nodes,the faster the response speed,and the brain plasticity information is more obvious in slow-5 frequency band.Exercise training promotes the plasticity of the brain.The more exercise training,the more attention people pay to the outside world,the faster the reaction time,the shorter the reaction time.