The Heterogeneity Of Autism Spectrum Disorder Based On Magnetic Resonance Imaging

Autism Spectrum Disorder（ASD）is a complex pervasive neurodevelopmental disorder whose core symptoms include impaired social interaction,communication,restricted and repetitive patterns of behavior and interests.ASD exhibits multi-level heterogeneity across genetics,brain systems,and behavioral manifestations.Heterogeneity has become the biggest obstacle to accurate diagnosis,individualized treatment,and prediction of clinical symptoms for ASD.Quantitative and individualized metrics for delineating the heterogeneity of brain structure in ASD are still lacking.Likewise,the extent to which brain structural metrics of ASD deviate from typical development（TD）and whether deviations can be used for parsing brain structural phenotypes of ASD is unclear.Similarly,how to use the brain functional connectivity in individual subspace to predict the clinical symptoms of ASD is a huge challenge.Therefore,with the development of magnetic resonance imaging technology,this research uses non-negative matrix factorization,Gaussian process regression,common and orthogonal basis extraction and other methods to quantitatively describe the brain structure and functional abnormalities of ASD individuals,and then try to explore the biomarker which could reveal the potential neural menchainsm underlying the ASD.The specific details are as follows:1.In view of the lack of low-dimensional and parameterized metrics for measuring the heterogeneity of ASD and the the degree to which ASD deviate from the normative range is unclear,second chapter uses the non-negative matrix factorization,and gray matter map is decomposed into 6 latent factors and corresponding weights,and Gaussian process regression is used to develop the normative model of the weight for each latent factor,and calculate the deviation of ASD individuals from the normative range.Deviations are used as the feature to identify the subtype of ASD.Compared with TD,ASD showed increased weights and deviations in five factors.Three subtypes with distinct neuroanatomical deviation patterns were identified.ASD subtype 1 and subtype 3 showed positive deviations,whereas ASD subtype 2 showed negative deviations.Distinct clinical manifestations in social communication deficits were identified among the three subtypes.Our findings suggest that individuals with ASD have heterogeneous deviation patterns in brain structure.The results highlight the need to test for subtypes in neuroimaging studies of ASD.This study also presents a framework for understanding neuroanatomical heterogeneity in this increasingly prevalent neurodevelopmental disorder.2.Aiming at the problem of how to overcome the heterogeneity of ASD and extract effective brain network features to predict the clinical symptoms of autism,the third chapter adopts the common orthonormal basis extraction algorithm,and the difference of functional connectivity between the ASD and the TD were decomposed into the common subspace and individual subspace.The results show that the functional connectivity difference matrix can be decomposed into three components in the common space.Component1 is mainly manifested as hyperconnectivity within and between the occipital network and the cerebellum,lower connectivity within and between the default mode network,frontoparietal network,cingulo-opercular network,sensorimotor network,and lower connectivity between the occipital network and the cerebellum and other networks.Component 2 is mainly characterized as the the hyperconnectivity within and between cinguloopercular network,the sensorimotor network,the occipital network and the cerebellum,the lower connectivity within and between default model network and frontoparietal network and other networks.Component 3 is mainly manifested as hyperconnectivity within and between all networks.At the same time,we found that the differences of functional connectivity in individual subspace could improved the predicton of clinical symptoms.Our study shows that the functional connectivity in the individual subspace of ASD is of great significance for predicting clinical symptoms,and provides a new perspective for the realization of individualized and precise diagnosis and treatment of ASD.Based on the structure of ASD and functional magnetic resonance brain imaging,our research proposes a quantitative,individualized,low-dimensional analytical framework to characterize the heterogeneity of ASD,and uses the differences in functional connectivity of individual subspace to predict the clinical symptoms of ASD.This study will help us to further understand the neuropathological mechanism of ASD,and provide a new perspective for revealing the heterogeneity of ASD and realizing precise clinical diagnosis and treatment services.