Study Of Method On Dynamic Brain Based On Clustering

Brain research was initially based on static assumptions.After years of development,brain research has formed a system based on the dynamic change hypothesis and using a combination of time and frequency domains.With various methods emerging,dynamic method systems need to be integrated and innovated on the original basis.How to better reflect or mine the temporal and spatial changes of brain dynamics from new methods has become an important issue that needs to be studied.The research in this paper is mainly combined with the clustering method under the dynamic method system.On the one hand,a new index is proposed based on the clustering calculation,results-cluster deviation,which is applied to the study of abnormal brain activity in patients with epilepsy and explore the meaning of the index.On the other hand,a software system,BRAINPLAYER,is proposed to improve the dynamic calculation system from an engineering perspective,which is specially used to view the clustering results generated by dynamic calculation.The main content of this article includes the following two parts:1.This article proposes a new indicator,cluster deviation,and uses it to investigate abnormal physiological activity in patients with epilepsy.Our study recruited three groups including temporal lobe epilepsy,generalized tonic-clonic epilepsy and age-,sex-matched normal controls.First,the standard deviation was compared and calculated,and no difference was found.Next,we clustered all the samples,and the clustering results are two resident states.Among them,state 1 accounted for the main state,showing that the amplitude of dynamic low-frequency oscillation in the left precuneus,the right angular gyrus and the left medial superior frontal gyrus was increased.State 2 is a non-primary state,showing an increase in the right middle occipital gyrus,and a decrease in the medial superior frontal gyrus.It can be inferred from previous studies that state 2 is a state characterized by epileptic abnormalities.It can be guessed from previous studies that State 2 is a state characterized by abnormal epilepsy.The calculated cluster deviation performance results are that the normal subjectsepilepsy patients are significantly higher in the hippocampus,the para hippocompal gyrus,and the temporal pole position than the normal subjects,and the temporal lobe epilepsy patients are more significant.According to previous studies,this location is closely related to the hippocampal sclerosis and focal area of temporal lobe epilepsy.This can be combined with the clustering results to infer that the significance of the low degree of clustering in this experiment is the severity of the abnormal behavior of the individual in the entire data sample.2.There is no software specifically for dynamic data visualization in the widely used software.From the perspective of data analysis,the process of data visualization is also very important for the discovery,determination,and verification of clustering results.This study identified the development of three main visualization modules,namely dynamic functional connection matrix,dynamic networks,and dynamic voxel level data display module.The main parts of the three modules all have dynamic data playing and color customization functions,and can perform simple data processing according to the corresponding display data characteristics,including dynamic function connection threshold processing,dynamic network node and edge size adjustment,dynamic voxel data template apply.The invention of this software provides a new auxiliary tool for the clustering process of dynamic data.In summary,this study includes two aspects,theory and practice.Theoretical study proposes a new cluster deviation index and applies the index to the analyse the normal subjects epilepsy data,so as to explore the theoretical significance of the index;on the other hand,from practical perspective,a new dynamic data visualization software is provided.Both works in the article provide new method and tool for the entire dynamic computing and clustering-extraction framework.