Correlation Between Dynamic Functional Connectivity And Clinical Features And Cognitive Function In First-episode Depression

Background and purposeDepression is a common mood disorder,with low mood,slow thinking and cognitive behavioral disorders as main clinical features.The incidence of depression is high,and it is easy to relapse,which causes harm to patient's physical and mental health and social behavior.While the pathogenesis of depression has not been fully understood yet,studies in neuroimaging have found evidences through various imaging techniques that brain structure and function of patients with depression have changed.Dynamic functional connectivity(dFC)is an emerging technology that can detect the dynamic changes of brain function connections(FC)on a shorter time scale.Based on group independent components analysis(Group ICA),this study uses a sliding time window technology and the cluster analysis to study the dFC of untreated first-episode depressed patients in various status and its correlation with clinical features and cognitive function of depression.Materials and methodsThis study includes 184 untreated depressed patients and 129 controls with comparable gender,age and years of education.All subjects are scanned with a GE Discovery 750 3.0 T magnetic resonance scanner for MR regular sequence and fMARI squence.The resting state fMRI data are preprocessed with the DPARSFA module in the DPABI software,and the resting state fMRI Group ICAand dFC are conducted with DPARSFA the GIFT software.Statistical analyses are performed using SPSS20.0 and GIFT software Stats module.The measurement data that follow a normal distribution are expressed by x±s,and the difference between two groups is compared with two independent samples t test;for those do not follow a normal distribution,their median(Upper and lower quartiles)are adopt,and the difference between the two groups is compared with Mann-whitney U test.A Spearman rank correlation analysis on the total time scores,average transit time,conversion times of the dFC indicators,the scores of the Hamilton rating scale for depression-17(HAMD-17),the course of disease,the wisconsin card sorting(WSCT)and trail marking test(TMT)is also performed.The test level is a=0.05.Results1.The results of clinical data analysis show that there is no significant difference in gender,age and years of education between the depression group and the control group(P>0.05),and the difference of HAMD-17 scores is statistically significant(P<0.05).2.The results of cognitive function analysis show:(1)The total time of WSCT of depression group is significantly(P<0.05)higher than that of the control group;(2)The completion time of TMT-A of the depression group is significantly(P<0.05)higher than that of the control group;(3)There are no significant differences between the two groups in the rest indicators of WSCT and TMT(P>0.05).3.The results of the resting state fMRI analysis show that:(1)The Group ICA analysis selects 37 independent components(ICs)that correlates with the resting-state networks(RSNs)from 100 ICs.These 37 ICs are categorized into 8 functional networks,including default mode network(DMN),salience network(SAN),auditory network(ADN),executive control network(ECN),and vision network(VN),language network(LAN),sensorimoter network(SMN)and precuneus network(PRN);(2)sliding window width is 30TR,the step size is 1TR,and the number of clusters is 5.After clustering,Five kinds of connection modes(state)are obtained from a clustering with a sliding window length of 30TR,step size of 1TR and five clusters.The proportions of these states are 24%,17%,20%,22%,and 17%,respectively;(3)There is a significant difference in the total time score and mean transit time between the depression group and the control group(P<0.05)in state 4.For other indicators and states,there is no significant difference between the two groups(P>0.05).(4)The difference of FC between state 1 and state 5 for the depression group is significantly different from the control group(P<0.05).For other states,there is no significant difference in FC between the two groups(P>0.05).4.Spearman rank correlation analysis on the three indexes of dFC and HAMD-17 score,disease course,WSCT and TMT show that:(1)total time score of state4 is positively correlated with HAMD-17 score(r=0.174),P=0.036)and the course of disease(r=0.169,P=0.039);(2)the total time score of statelis positively correlated with the number of completion of the first classification response(r =0.232,P = 0.017),correct thinking time(r = 0.232,P =0.017),and non-sustained error number(r = 0.232,P = 0.017),while it is negatively correlated with error thinking time(r=-0.257,P=0.008),continuous response number(r=-0.216,P=0.027),the number of persistent errors(r=-0.257,P= 0.008),and the number of classifications that cannot be completed(r =-0.257,P = 0.008);(3)the average transit time of state 1 is positively correlated with the number of error responses(r=0.231,P=0.018),complete the first classification response number(r=0.302,P=0.002),total response number(r=0.226,P=0.021),correct thinking time(r=0.302,P= 0.002),non-sustained error number(r =0.302,P=0.002),yet it is negatively correlated with continuous response number(r=-0.283,P= 0.003),error thinking time(r=-0.203,P=0.03 8),the number of persistent errors(r=-0.203,P=0.038),number of classifications that cannot be completed(r=-0.203,P= 0.038);(4)the number of conversions is positively correlated with the completion time of TMT-A(r=0.214,P=0.045)),while it is negatively correlated with the wrong thinking time(r=-0.236,P = 0.016),the number of persistent errors(r =-0.236,P = 0.016),the number of classifications that cannot be completed(r=-0.236,P= 0.016);(5)The dFC indicators of state 3 and 4 do not show any significant correlations with the scores of WSCT and TMT.5.The verification analysis results show that the result is highly reproducible with changing sliding time window width and the number of clusters.Conclusions1.The longer the course of the first-episode depressed patients,the more severe their symptoms,and the more likely that they stay in the weak-function connection mode.2.The dFC method can effectively identify FC changes in a shorter time window,and its characteristic indicators are helpful to reflect cognitive behaviors.