Exploring The Characteristics Of Altered Spontaneous Brain Activity In Major Depressive Disorder With Different Subtypes Based On Resting-state FMRI

Objective To investigate the characteristics of altered spontaneous brain activity in major depressive disorder(MDD)with different subtypes based on resting-state fMRI.The aim is to analyse specific brain abnormalities between Anxious-somatizing depression(ASD)and Melancholic depression(MD),and whether these abnormalities can differentiate the subtypes,in an attempt to find potential brain imaging markers for subtyping.Methods According to the inclusion criteria,545 patients with major depressive disorder and 56 healthy controls were included.General demographic information was collected from all subjects.The 17-item Hamilton Depression Rating Scale(HAMD-17)and the 30-item Inventory of Depressive Symptomatology(IDS-30)were collected from the case group to classify patients with MDD into 244 cases of ASD and 301 cases of MD,and to compare the clinical characteristics and course of the disease among the different subtypes.From which 25 cases of ASD,30 cases of MD and 23 cases of HC who were willing to cooperate with MRI data acquisition were selected for resting-state functional MRI scans.A total of 19 cases of ASD,26 cases of MD and 20 cases of HC were excluded from the analysis after exclusion of unqualified brain images,and their regional homogeneity(ReHo)and fractional amplitude of low-frequency fluctuation(fALFF)were calculated.Clinical data were statistically analysed using SPSS 26.0 with independent samples t-test or ANOVA for comparison of quantitative data between groups and chi-square test or Fisher’s exact test for comparison of count data.The functional brain imaging data were statistically analysed using DPABI based on Matlab R2018 b software.Using age,gender and years of education as covariates,analysis of covariance was used to compare the differential brain areas of ReHo and fALFF between the three groups,and LSD correction was further applied to compare between the two groups.The Gaussian random field(GRF)was used to identify brain areas with statistically significant differences at P < 0.001 for individual voxel levels and P < 0.05 for clusters of brain areas.Subsequently,the ReHo and fALFF values of the differential brain areas in each group were tested for bias correlation with HAMD-17 scores using gender,age,marriage,occupation and years of education as covariates.Finally,ROC curve was used to analyse the diagnostic classification value of ReHo and fALFF differential brain areas between ASD and MD,between ASD and HC,and between MD and HC.Results1.Comparison of general demographic and clinical characteristics between the three groups There were no statistically significant differences between the three groups in terms of ethnicity and years of education(P>0.05).The differences in age,sex,occupation,marriage,HAMD-17 score and factor scores were statistically significant(P<0.05).The ASD group was found to be older than the MD group;Compared to the HC group,the ASD and MD groups were more female and less male,more unemployed and less employed,while the MD group had a higher proportion of students;the ASD group had more married people and the MD group had more unmarried people.The ASD and MD groups had higher HAMD-17 scores and factor scores than healthy controls,and the ASD group had higher anxiety/somatization factor scores than the MD group,and lower total HAMD-17 scores,weight factor scores,cognitive disturbance and retardation,and total IDS-30 scores than the MD group.the ASD group had a higher age of first onset than the MD group,and there was no statistically significant difference between the two groups in terms of treatment modality,form of onset,duration of current and total illness,number of episodes and family history(P>0.05).2.Comparison of the differences in ReHo and fALLF between the three groups Conclusion An analysis of covariance of ReHo between the three groups,using sex,age and years of education as covariates,showed that there were differences in ReHo values in the left fusiform,left Inferior/middle occipital gyrus and left putamen,and differences in fALFF values in the right Cerebelum＿8/9,left Cerebelum＿8,right fusiform/parahippocampal gyrus,right caudate nucleus,left putamen and left fusiform/inferior temporal gyrus.These differences were statistically significant(P<0.05,GRF corrected). Comparison between the two groups yielded: ReHo and fALLF values in the right inferior temporal gyrus were elevated in the ASD group compared to the MD group.Compared to the HC group,the ASD group showed higher ReHo values in the bilateral precuneus,lower ReHo values in the left putamen and bilateral inferior/middle occipital gyrus,and lower fALFF values in the right fusiform/parahippocampal gyrus,right caudate nucleus,right Cerebelum＿8 and left fusiform / putamen nucleus.Compared to the HC group,the MD group showed higher ReHo values in the right precuneus and left superior frontal gyrus dorsolateral,lower ReHo values in the left putamen,left fusiform and left inferior occipital gyrus,and lower fALFF values in the right superior frontal gyrus orbital part,right fusiform/inferior temporal gyrus,right Cerebelum＿8/9 and left fusiform /inferior temporal gyrus,all with statistically significant differences(P<0.05,the GRF corrected).3.Correlation analysis of ReHo and fALLF differences with clinical symptoms The mean time series of ReHo and fALLF differences in the brain regions of the three groups were extracted separately,and a partial correlation analysis was performed after controlling for gender,age,marriage,occupation and years of education.Compared to the MD group,there was a negative correlation between the increased ReHo values in the right inferior temporal gyrus and its sleep disturbance factor score in the ASD group(r=-0.576,P=0.031).Compared to the HC group,there was a positive correlation between the decreased ReHo values in the left inferior/middle occipital gyrus and the anxiety/somatization factor score in the ASD group(r=0.524,P=0.045).Compared to the HC group,the increased right precuneus ReHo values in the MD group was negatively correlated with its total HAMD-17score(r=-0.447,P=0.029)and sleep disturbance factor score(r=-0.444,P=0.044),the decreased left fusiform ReHo values was positively correlated with its retardation(r=0.449,P=0.041),and the decreased the right fusiform/ Inferior temporal gyrus fALLF values was negatively correlated with its sleep disturbance factor score(r=0.439,P=0.046).4.Diagnosis of disease classification by ReHo and fALLF differences The mean time series of ReHo and fALFF difference brain regions between the two groups were extracted separately,and based on binary logistic regression and ROC curve analysis,it was found that right inferior temporal gyrus ReHo values(AUC=0.87),right inferior temporal gyrus fALFF values(AUC=0.891),the combination of which predicted(AUC=0.935)better classification of ASD with MD.Right inferior/middle occipital gyrus ReHo values(AUC=0.921),left fusiform/putamen fALFF values(AUC=0.897),the combination of which predicted(AUC=0.992)better classification of ASD with HC.Right precuneus ReHo values(AUC=0.919)and the right superior frontal gyrus orbital part fALFF values(AUC=0.931),the combination of which predicted(AUC=1.0)better classification of MD with HC.The accuracy,sensitivity and specificity of the combined index predictions were significantly higher than those of the individual indexes,and might be used as a potential specific diagnostic classifier to distinguish the ASD from the MD group,the ASD from the HC group,and the MD from the HC group.Conclusion1.The two subtypes of ASD and MD have their own specific mechanisms of brain function abnormalities,mostly involving the default mode network,frontalstriatum-limbic system,reward loop and other related brain areas,and their abnormalities are correlated with clinical symptoms.2.Abnormal ReHo or fALLF in the right inferior temporal gyrus,Abnormal ReHo in the right inferior/middle occipital gyrus occipital gyrus and abnormal fALLF in the left fusiform / putamen,abnormal ReHo in the right precuneus and abnormal fALLF in the right superior frontal gyrus orbital part may promise as a potential imaging biomarker for predicting ASD and MD,diagnosing ASD and MD respectively.