Cognitive Subtype Study Of Firstepisode Psychosis Based On Fuctional MRI

Neuropsychiatric diseases are people’s health and spiritual life destroyers,which bring a heavy burden on family members and society.Neuropsychological diseases are diseases caused by progressive damage to the nervous system during neurodevelopment,such as mental disorders,amnesia,depression,cognitive disorders,etc.,which cause great pain to patients and significantly reduce the quality of life of patients.Among them,FirstEpisode Pyschosis(FEP)which usually starts in adolescence is a common state of early onset of many chronic psychiatric diseases.Because the symptoms of FEP are not stable enough,the diseases cannot completely determine the specific disease classification and can only be diagnosed according to the differential clinical phenotype after the annual observation period.This period is a crucial period for the treatment of diseases.It would be of great help to patients if the period can be grasped for the treatment in time.The data which obtained from John Hopkins Medical School in the United States in this study are from first-episode pyschosis group and healthy control group,including Cognitve Test Score(CTS)data and functional Magnetic Resonance Imaging data(f MRI).In this paper,the data are processed and analyzed,and the disease characteristics of FEP are extracted,and then the FEP cognitive subtypes are divided and the brain network function connection activity analysis of cognitive subtypes is presented at the end of this paper.Specifically,this paper first introduces the CTS collection process and specific meanings,and presents the f MRI digitization process in detail,and then uses the general linear model to correct the f MRI data,so as to exclude the influence of different age,gender and race on the analysis results.Next,this paper extracts six cognitive ability factors from the cognitive test score data using factor analysis,and uses hierarchical clustering on f MRI data to divide the whole brain structure into eight functionally differentiated brain networks at the same time.Thus,this paper uses Canonical Correlation Analysis(CCA)combined with L1 regularization and permutation test to extract two essential basic features—memory related features and ideational-executive related features.In this paper,hierarchical clustering is used to classify FEP patient group based on the two essential basic features,and then four subtypes with stable and welldefined boundaries are obtained.Finally,this paper combines the brain network function connection mode to analyze the abnormality of brain network function connection and the brain network activity of cognitive ability factors in FEP subtype.