Analysis Of The Topological Properties Of The Brain Network And Its Correlation With Cognitive Function In Patients With Temporal Lobe Epilepsy

Background and purposePatients with epilepsy account for about 1 percent of the global population,and about one-third of epilepsy patients are drug-tolerant and the seizures are difficult to control,called refractory epilepsy.TLE(temporal lobe epilepsy)is the most common refractory epilepsy,cognitive decline is a common complication of TLE,patients tend to experience a serious decline in quality of life,so epilepsy is not only a purely neurological disease,but is gradually becoming a difficult social problem.A process study on abnormal discharge in epilepsy found that single neuron is not sufficient to cause seizures,abnormalities in a group of cells are necessary to produce clinical symptoms or abnormal discharge of EEG Therefore,the abnormal synergistic activity of neurons in the brain network constitutes the basic pathophysiological mechanism of epilepsy.As a mathematical description and analysis method for complex systems,graph theory has been applied to analyze artificial and biological networks.Recently,considerable attention has been paid to brain research because it provides a powerful form to quantitatively describe the topological properties of brain connections.The construction of complex networks can more flexibly and naturally demonstrate the real state of the network and the underlying dynamic evolution process.Based on the BOLD-fMRI sequence,this study aimed to investigate the changes of global and local topological attributes of TLE whole brain network.The correlation between brain network parameters and cognitive function was analyzed to explore the relationship between the brain network parameters and cognitive impairment,and to provide the basis for the pathophysiological mechanism of cognitive impairment in epileptic patients.Materials and methods1.This study used a German siemens Prisma 3.0T MR scanner for routine head plain scan and resting-state BOLD sequence scans of all subjects enrolled;the subjects were 33 patients with TLE and 30 HC(healthy controls)(sex,age,and years of education matched).All subjects were collected for clinical information:general information,MMSE(mini-mental state examination),AVLT(auditory verbal learning test),MES(memory and executive Screening),MOCA-B(montreal cognitive assessment).Patients with TLE were simultaneously assessed by the NHS3(national hospital seizure severity scale).2.Using DPARSFA software to preprocess resting state BOLD-fMRI data based on MATLAB platform;Using AAL(Anatomical Automatic Labeling,AAL)to segment the brain into 116 brain regions of interest,to the synchronism between time series of each brain region,calculate the symmetry correlation matrix of 116x116,then perform Fisher r-to-z conversion on each subject;Using GRETNA software for further graph theory analysis,the global network topological attribute indexes:Cp(Clustering coefficient),Lp(characteristic path length),small world attributes ?=?/?,assortativity,Eg(global efficiency),hierarchy,Eloc(local efficiency);the local network topological attributes indexes:nodal betweenness,nodal centrality,nodal efficiency.3.Using t test(normal distribution)or nonparametric test(Mann-whitney U test)of SPSS21.0 to compare the brain network indicators of the TLE group and the HC group,and compare the statistically significant network indicators with duration,cognitive function and NHS3 score were analyzed by Spearman correlation.Results1.Results of clinical data analysis:no significant difference was found in age,sex ratio and years of education between patients with TLE and HC group(P>0.05).Scores of instant memory(N1),short delay memory(N4),MES,MMSE,MOCA-B and recognition were statistically significant(P<0.05),and the scores of each item in the TLE group were smaller than those in the HC group.2.Global network topological attribute indexes:in the range of sparsity(0.05-0.5),the TLE group and the HC group were in accordance with the small world attribute:the ? value>1,the value was 1,and the small world attribute index a=?/?>1.The TLE group ? significantly decreased(P=0.023),decreased ?(P=0.037),decreased Eloc(P=0.005)and decreased hierachy(P=0.039)compared with HC.Moreover,the Cp,Eg,Lp,? and assortativity between the two groups were not statistically significant(P>0.05).3.Local brain network topological attribute indexes:the nodal betweenness of the right superior frontal gyrus,right hippocampus,left middle occipital gyrus,left superior parietal gyrus,left thalamus and area 7 of vermis of the TLE group were significantly lower than those in the HC group,but the nodal betweenness of the right triangular inferior frontal gyrus was significantly higher than that in the HC group.The nodal centrality of the left thalamus,area 6 of left cerebellum,area 6 of vermis and area 7 of vermis in the TLE group were significantly lower than that in the HC group,but the nodal centrality of the right Rolandic operculum and right superior occipital gyrus were significantly higher than that of the HC group.The nodal efficiency of the left middle frontal gyrus,orbital middle frontal gyrus,right parahippocampal gyrus,right amygdala,left inferior parietal horn gyrus,right paracentral lobule,left pallidum,left thalamus,area 6 of left cerebellum,area 7 of vermis were significantly lower than that of HC group(P<0.05),however,the nodal efficiency of the right paracentral lobule was significantly higher than that of the HC group.4.Correlation analysis:there was no significant correlation between global network attribute index and cognitive index,duration of disease,NHS3 in TLE group(P>0.05).However,some local attribute indexes in TLE group were significantly correlated with cognitive function.Nodal betweenness:the nodal betweenness of the right superior frontal gyrus was positively correlated with MES(P=0.027,r=0.461),the left superior parietal gyrus and left thalamus were positively correlated with MOCA-B score(P=0.009,0.023,r=0.542,0.481),and the left superior parietal gyrus and N4 were positively correlated(P=0.036,r=0.449).Nodal centrality:the nodal centrality in area 7 of vermis was positively correlated with MOCA-B score(P=0.049,r=0.425).Nodal efficiency:the nodal efficiency of left orbital middle frontal gyrus was positively correlated with the MOCA-B score(P=0.035,r=0.452).Conclusions1.The whole brain network of TLE patients still has the properties of small world,but some of the global topological parameters are lower than those of the control group,and generally tend to be random network characteristics.2.The topological properties of local brain network in patients with TLE(nodal betweenness,nodal centrality,nodal efficiency)have changed.Most of the areas where the nodal properties change locate in the default mode network,and there is a phenomenon of nodal properties increasing and decreasing.3.Some nodal attributes of TLE are positively correlated with cognitive scores,suggesting that nodal attributes can be used as an imaging marker for cognitive evaluation.