Feature Classification And Auxiliary Intervention Analysis Of Autistic Children Based On Brain Function Network

Autism spectrum disorder is a neurodevelopmental disorder associated with impaired executive function,language,mood,and social functioning.On the basis of constructing the brain function network,this paper analyzes the differences in the characteristics of the brain function network between autistic children and normal children,and realizes the effective classification of autistic children and normal children based on the support vector machine classifier.Further,based on the characteristics of the brain functional network,this paper explores the effect of transcranial direct current stimulation intervention on the improvement of the brain function of autistic children,hoping to provide help for the brain regulation of autistic children.Eeg and behavioral data were collected from a total of 40 children,including 20 autistic children(4-8 years old)who received tDCS intervention and 20 control children(4-8 years old)who received sham stimulation.The resting eeg was collected in a quiet environment and the behavior scale was tested.In this paper,the characteristic parameters of brain functional network were analyzed.Firstly,the brain network connection matrix of five frequency bands,namely the directional transfer function,was extracted from the eeg of 16 channels in autistic children.The selection of threshold will affect the network topology.In this paper,the method of selecting the best threshold is specially adopted,and the graph theory analysis method is applied to obtain the network characteristics of each frequency band.The characteristics of brain network include average degree,global efficiency,local efficiency,characteristic path length and clustering coefficient.The changes of brain functional networks in children with autism before and after tDCS stimulation were statistically analyzed.Finally,the network features under different thresholds were used as input feature vectors of the classifier,and three optimization algorithms were used to optimize the classification of support vector machines to obtain the most classification results,so as to explore the differentiation degree of the brain network features of autistic children before and after tDCS stimulation.The results showed that the expression ability of autistic children was improved after intervention by ABC scale analysis of behavioral data.There were significant differences between the two groups before and after stimulation in children with autism in the whole brain(p<0.001),but there was no statistical difference before and after stimulation in the control group with false stimulation(p>0.05).Subsequently,the analysis of the characteristics of the brain network graph theory before and after the stimulation under the optimal threshold showed that the statistical difference of the characteristic path length and clustering coefficient in Gamma frequency band was the most significant(p<0.01).Through optimization algorithm to optimize the function of support vector classification of network characteristics,the results show that the grid search optimization algorithm is the best effect,spectrum classification accuracy rate were 71.96%,82.01%,87.96%,87.96% and 84.79%,and the running time of 2.528 s,in a certain extent,transcranial direct current stimulation can serve as an effective means of intervention in autism,for children with autism early intervention treatment provides a train of thought.