A Research On Multi-level Markov Relation Of Resting State Data In Brain

Functional magnetic resonance imaging(fMRI)is a non-invasive technique to measure the activation level of various regions of the brain,that is one of the important technologies to study brain activity.The analysis of fMRI data has obtained a lot of results and deepened people's understanding of brain activity.However,the brain is a very large and complex system,people's understanding of the brain is far from being thorough.Therefore,the research of the brain still has important theoretical and practical significance.This article obtained fMRI data of 758 healthy subjects,analyzed the resting fMRI data of the brain,and explored the relationship between the characteristics of resting fMRI data analysis and the cognitive ability of the subjects.First of all,this paper carried out dynamic functional connection analysis and coactivation mode analysis on fMRI data.In the analysis of dynamic functional connections,this paper uses a sliding window to capture dynamic functional connections,and uses the K-means clustering algorithm to determine five repetitive dynamic functional connection modes,and deeply analyzes the functional connection characteristics of each connection mode.And calculate the relevant statistics of each connection mode in the entire data set.In the analysis of co-activation patterns,this paper uses the K-means clustering algorithm to determine eight recurring co-activation patterns,and analyzes in detail the activation of each brain region in each co-activation pattern.Similarly,the relevant statistics for each co-activation mode are calculated.Dynamic functional connection mode and coactivation mode are the results of fMRI data analysis from two different angles.This article explores the synchronization between the two modes.This article uses the statistics of dynamic functional connection mode and coactivation mode as features to explore its relationship with cognitive ability.First,this paper proposes to combine the features of resting state data with the features of task state data to train the model and predict cognitive ability.In this paper,three cognitive ability scores of the same subjects as the resting state fMRI data and corresponding task state fMRI data are obtained.In this paper,dynamic functional connection analysis and coactivation mode analysis are performed for each task state fMRI data set.Using the neural network model,the statistics of dynamic functional connection mode and co-activation mode are used as features,and the resting state features are trained respectively.Model,task state feature model,and joint feature model of resting state and task state.The results show that the joint feature model is significantly improved compared to the model trained only with resting state features or task state features.Second,explore the statistical relationship between resting state characteristics and cognitive abilities.It mainly adopts the method of correlation significance test to test whether the statistical features of dynamic function connection mode and co-activation mode have a significant correlation with cognitive ability.The results show that certain cognitive abilities have significant correlations with certain statistics,which provide a reference for understanding the relationship between resting brain activity and cognitive abilities.