Study On Functional Frame Network In Subjects Of Schizophrenia

Introduction:The widespread connectivity abnormalities in schizophrenia have led it to be considered as a disorder with dysconnectivity.Abnormal coupling between functional networks and Rich-club structure destruction in the brain of patients suggests that the damage of important connections may be involved in the pathological mechanism of this disease.Although the human brain showed relatively stable structural features,there are few studies on the stability and importance of functional connections.This study proposes a novel approach to identify the stable functional connections in brain network,which is called frame network.Meanwhile,we assume that the frame network is abnormal in schizophrenia.Method:By setting of 1% connections with smallest coefficient of variation,we found a widespread brain functional network(frame network)in healthy people(n=380,one dataset from Consortium for Reliability and Reproducibility project and the other from 1000 Functional Connectomes Project public database).We then explored the alterations in medicated group(from the Center for Biomedical Research Excellence project,including 60 schizophrenia subjects,C-SCH;and 71 mached controls,C-HC)and drug-naive first-episode group enrolled at the second Xiangya Hospital(FES,n=83),subjects at Ultra-High Risk(UHR,n=65).67 controls were compared with FES and UHR subjects.A linear support vector classifier(SVC)was constructed to distinguish patients and controls using the frame networks.NBS-predict was used to identify the frame connections which are strongly associated with disease classification.Meanwhile,FES subjects were clustered into two clinical subtypes based on pathological symptoms.Connectome-based predictive modeling(CPM)was used to identify the frame connections associated with negative symptoms in patients with different subtypes.Granger causality analysis was used to construct effective connections of dorsolateral superior frontal gyrus to explore the changes in the patient group,and multiple linear regression models were constructed based on the significantly changed effective connections to measure its effect on pathological symptoms.Results:We found most frame connections of healthy people had high strength,which were symmetrical and connected the left and right hemispheres.Conversely,patients’ frame network were more left lateralized,concentrating on the left frontal lobe.Significant differences in frame connections were observed in both patient groups.However,there was a difference pattern between the FES group and C-SCH group.In the FES group,the connections in the frontal parietal region increased,and the connections in the thalamus and basal ganglion decreased,while all connections in the C-SCH group were decreased.The frame networks of healthy people,the C-SCH group and FES group could effectively distinguish patients from the control group.The C-SCH group could distinguish both two kinds of patients from the controls.The frame connections associated with disease classification were concentrated in the prefrontal cortex(dorsolateral superior frontal gyrus,medial superior frontal gyrus and supplement motor areas)and the central cortex(anterior central gyrus,posterior central gyrus).All of three framework networks above could predict negative symptoms in the FES group.In the subtype with mild negative symptoms,the predicted connections were concentrated in the occipital lobe(medial occipital gyrus,calcarine fissure cortex)and the prefrontal lobe(dorsolateral superior frontal gyrus,medial superior frontal gyrus);in the subtype with severer negative symptoms,the predicted connection was between bilateral medial superior frontal gyrus and middle frontal gyrus.In the C-SCH and FES groups,the effective connection of some brain regions to the dorsolateral superior frontal gyrus were significantly reduced.In FES group,all the different connections were observed in the left brain,and all the significantly increased effective connections were sent from the left dorsolateral superior frontal gyrus,spread all over the brain.These changed effective connections contributed to variety symptoms in the FES group,especially cognitive symptoms and excitatory symptoms.Conclusion:These findings revealed the existence of stable high-strength connections in the brain.The abnormalities of these connections in schizophrenia subjects might serve as potential markers for disease classification and symptom prediction.Frame network could provide new insights into the dysconnectivity in schizophrenia and its clinical heterogeneity.