Exploring The Brain Structure And Function Of First-episode Obsessive-compulsive Disorder With Multimodal Magnetic Resonance Imaging

Obsessive-compulsive disorder(OCD))is a chronic disabling mental disorder characterized by persistent and intrusive thoughts,impulses,images,and repetitive behaviors.OCD is usually chronic,prolonged,and worsens gradually as the disease progresses.Recursive thinking and stereotypical behaviors that OCD cannot control often cause great pain and burden to patients and their families,and are a major cause of disability in worldwide.Clarifying the pathophysiological mechanism of OCD will be helpful.However,until now,the neurobiological mechanisms of OCD have been unknown.Neuroimaging studies have suggested that abnormalities in the corticostriato-thalamo-cortical(CSTC)loop may be involved in the morbidity of OCD.However,recent studies have suggested that abnormalities in brain regions outside the CSTC loop may also be involved in the pathogenesis of OCD and patients with OCD may have extensive cortical and subcortical abnormalities,suggesting the importance of considering the neural mechanism of OCD at the whole brain level.Resting-state functional magnetic resonance imaging(rs-fMRI)can reflect the functional activity characteristics of the brain in the resting state.High resolution structural magnetic resonance imaging(MRI)can reflect the information of the basic anatomical structure.and diffusion tensor imaging(DTI)can provide the structure information of the white matter of the brain.The combination of these three modes of MRI can not only explore the possible changes of brain functional activities of OCD,but also explore the anatomical structural basis of abnormal functional activities of OCD through the analysis of cortical and subcortical structures,so as to comprehensively reflect the possible neural mechanism of patients with OCD.In addition,most of the previous neuroimaging studies on patients with OCD were conducted on a single sequence or single index based on small samples,and most of the studies did not strictly control the factors of comorbidities and clinical intervention such as drugs,psychological and brain regulation,and results of related researches were inconsistent to a certain extent.Therefore,multimodal MRI techniques based on rs-fMRI,high-resolution structural MRI,and DTI were used in this study to conduct whole-brain level analysis on a large sample of first-episode and treatment-naive OCD patients,in order to fully explore the possible neural mechanisms of OCD patients.We carried out our research from the following three parts:Part 1 Multidimensional changes of resting-state brain function in first-episode OCD patients based on the whole brain levelBackground and ObjectiveFunctional magnetic resonance imaging(fMRI)is based on the blood oxygen level dependent response of different brain regions to evaluate the activity of brain regions and the changes of brain connectivity,which can reflect the changes of brain function in vivo noninvasively.Rs-fMRI can avoid the limitations of task-state fMRI and is with high repeatability,less restrictions on subjects and wide application.Regional homogeneity(ReHo)is a commonly used Rs-fMRI analysis method,which reflects the coordination of regional spontaneous brain activity.Compared with other methods such as the amplitude of low frequency(ALFF),fractional amplitude of low frequency(fALFF),ReHo is more stable in test-retest analysis and less affected by global signal interference.Degree centrality(DC)describes the centrality of the network through the evaluation of nodes.It is the most direct analysis method to describe the impact and function of nodes,and it is an important indicator to reflect the change of functional network.The high degree of synchronization of spontaneous neural activity between appositive regions is considered to be an important feature of normal brain function.Voxel-based mirror homotopy connectivity(VMHC)can reflect the changes of homostatic functional connectivity between left and right cerebral hemispheres,and measure the coordination between hemispheres.However,there are relatively few studies in OCD.Therefore,the purpose of this part is to use the whole brain level analysis based on rs-fMRI,using ReHo,DC and VMHC to explore the possible changes of brain functional activity in patients with first-episode and treatment-naive OCD from the three dimensions of homogeneity of regional brain spontaneous activity,central hub of brain topological network and functional connectivity of mirror apposition regions between left and right hemispheres,and to explore the relationship between abnormal functional activity changes and clinical symptoms.Materials and MethodsIn this part,a total of 106 patients with first-episode and treatment-naive OCD were enrolled with 106 healthy volunteers(normal control group)whose age,gender,and years of education matched the OCD group.Nine OCD patients and 6 healthy volunteers were excluded because their head motion exceeded the preset threshold.The magnetic resonance data was collected by a magnetic resonance physician who was skilled in MRI scanning on a General Electrical(GE)Discovery 750 3.0T magnetic resonance scanner with an 8-channel phased array head coil.1.In order to examine whether there were organic brain lesions,the subjects’brains were scanned using T1 fluid attenuated inversion recovery(FLAIR)sequence,T2-weighted imaging(T2WI)sequence and T2-FLAIR sequence which were commonly used in clinical.After confirmed the absence of organic brain lesions,the subjects were scanned by Rs-fMRI.Rs-fMRI data were obtained by gradient echo(GRE)single ’shot echo planar imaging(EPI)sequence.2.Two psychiatrists trained in scale consistency conducted clinical interviews and the Yale-Brown obsessive-compulsive scale(Y-BOCS)was used to evaluate obsessive and compulsive symptoms.3.The data processing assistant for resting state fMRI software(DPARSF)was used for image preprocessing,which was based on statistical parameter mapping(SPM)12 and Matlab platform.The specific processing steps included:the conversion of data format;the removal of the first 5 time points;slice timing correction;head motion correction;spatial standardization;the removal of linear trends;nuisance covariate regression;band-pass filtering(0.01-0.08 Hz);scrubbing.In order to eliminate the influence of head motion,we excluded the subjects whose translation in any direction was greater than 2 mm or whose rotation in any direction was more than 2° and calculated the average frame displacement(FD).The subjects whose mean FD was greater than 0.5 were also excluded to minimize the influence of head motion on the results.4.DPARSF was used for ReHo analysis of the preprocessed data.The ReHo value was obtained by calculating the Kendall’s coordination coefficient(KCC)of 27 nearest neighboring voxels.Then,these maps were normalized to standard z scores.Subsequently,we used the 6 mm Gaussian kernel(FWHM)to smooth the ReHo maps.5.DPARSF was used for DC analysis of the preprocessed data.DC was obtained by calculating the sum of functional connections between each voxel and all other voxels.Then we constructed the brain connectivity matrix and calculated the number of voxels whose correlation threshold exceeded 0.25.Then,these maps were normalized to standard z scores and spatially smoothed with a Gaussian kernel(FWHM)of 6 mm.6.DPARSF was used for VMHC analysis of the preprocessed data.The Pearson correlation coefficient of the time series of each voxel and its corresponding position in the other half of the cerebral hemisphere was calculated as the isopotal resting state functional connection of the individual.Then the obtained Pearson correlation coefficient was transformed by Fisher z transform to generate the normalized VMHC maps.Finally,the standardised VMHC maps were spatially smoothed with a Gaussian kernel(FWHM)of 6 mm.7.The statistical package for social sciences(SPSS 21.0,IBM(?),USA)was a statistical analysis tool for clinical data of subjects.Student’s t-tests were used to analyse the differences between groups of normally distributed data.Nonparametric tests were used to analyse the differences between groups of data that did not conform to a normal distribution.P<0.05 was considered as statistically significant.The differences in ReHo,DC,VMHC between OCD patients and normal control group were conducted by student’s t-tests,with age,gender,years of education and mean FD of subjects as covariates.Gaussian random field(GRF)was used for multiple comparison correction.The threshold values for evaluating the differences between groups were P<0.05 for the cluster level and z>2.807 for the voxel level After that,the brain regions with statistical significance were saved as region of interest(ROI),and the functional index value of corresponding brain regions were extracted.The relationship between the functional index value and the Y-BOCS scores was explored by correlation analysis.P<0.05 was considered as statistically significant.Patients with OCD were divided into three groups according to the severity of symptoms,and the correlation between abnormal functional index values and Y-BOCS scores in different subgroups was further analyzed.Results1.Compared with healthy controls,the first-episode and treatment-naive OCD patients showed increased ReHo value in the left orbitofrontal cortex and left inferior frontal gyrus(mainly in triangular region),and decreased ReHo value in bilateral lingual gyrus and calcarine(P<0.05 for the cluster level and z>2.807 for the voxel level,GRF correction).2.Compared with healthy controls,the first-episode and treatment-naive OCD patients showed increased DC value in the right insula,left orbitofrontal cortex,and left inferior frontal gyrus(mainly in opercular and triangular regions),and decreased DC value in bilateral calcarine,lingual gyrus,and left middle frontal gyrus(P<0.05 for the cluster level and z>2.807 for the voxel level,GRF correction).3.Compared with healthy controls,patients with first-episode and treatment-naive OCD showed reduced VMHC value in bilateral insula,bilateral calcarine,and lingual gyrus(P<0.05 for the cluster level and z>2.807 for the voxel level,GRF correction).4.The ReHo value of the left orbitofrontal cortex and left inferior frontal gyrus triangle in patients with first-episode and treatment-na(?)ve OCD was negatively correlated with the total score of Y-BOCS(r=-0.258,P=0.011).The DC value of the left middle frontal gyrus in patients with severe OCD was positively correlated with the total score of Y-BOCS(r=0.379,P=0.019).Conclusions1.Changes of regional brain spontaneous activity,topological network attributes,and interhemispheric coordination are involved in the pathophysiological process of OCD,suggesting that abnormal brain function and activity in patients with first-episode and treatment-naive OCD occurred in multiple dimensions.2.Abnormal brain functional activities in patients with first-episode and treatment-naive OCD are not limited in the classical CSTC circuit,but also exist outside the CSTC circuit.3.It is speculated that the abnormal brain functional activities in the left orbitofrontal cortex and left inferior frontal gyrus triangle and left middle frontal gyrus are the functional basis of obsessive-compulsive symptoms,and it provides a possible intervention direction for OCD.Part 2 An MRI study of structural changes in gray matter in first-episode OCD patients using voxel-based morphometryBackground and ObjectiveIn the first part of the study,we found that OCD patients had abnormal brain functional activity in multiple dimensions,and the study of high-resolution structural MRI in OCD patients could provide relevant information on the anatomical basis of functional activity changes.The voxel-based morphology(VBM)analysis using the computational anatomy toolbox(CAT 12)can better detect changes in gray matter volume than other software.Therefore,the purpose of this part is to use the CAT12-based VBM to explore the structural changes of the gray matter in the whole brain of patients with first-episode and treatment-naive OCD,and to explore the relationship between the structural changes of the gray matter in patients with OCD and their clinical symptoms,so as to provide the possible neural mechanism of the patients with OCD from the direction of the gray matter structure.Materials and MethodsIn this part,a total of 100 patients with first-episode and treatment-naive OCD were enrolled with 106 healthy volunteers(normal control group)whose age,gender,and years of education matched the OCD group.The magnetic resonance data was collected by a magnetic resonance physician who was skilled in MRI scanning on a GE Discovery 750 3.0T magnetic resonance scanner with an 8-channel phased array head coil.1.The T1-weighted imaging(TiWI)with the three-dimensional brain volume(3D BRAVO)sequence was used to obtain the high-resolution structural images.2.Two psychiatrists trained in scale consistency conducted clinical interviews and the Y-BOCS was used to evaluate obsessive and compulsive symptoms.3.Firstly,the dcm2niigui software based on MRIcron was used to convert DICOM format file of original structure images to NIFTI format.Then,the structural images were analyzed by the CAT 12 in the SPM 12 based on MATLAB platform to obtain voxel comparison of gray matter volume.The specific process included the removal of bias field and noise;the dissection of the skull;the segmentation;finally,on the 1.5 mm isotropic adult template provided by CAT12,using the DARTEL,the normalization to Montreal Neurological Institute(MNI)coordinate space was performed.The generated images were checked for homogeneity.Because all images were with a high correlation value(>0.85),no images were discarded.Subsequently,we used the 8 mm Gaussian kernel(FWHM)to smooth the gray matter images.4.The SPSS 21.0 was used for statistical analysis of clinical data of subjects.Student’s t-tests were used to analyse the differences between groups of normally distributed data.Nonparametric tests were used to analyse the differences between groups of data that did not conform to a normal distribution.P<0.05 was considered as statistically significant.The differences in gray matter volume between OCD patients and normal control group were conducted by student’s t-tests,with age,gender,years of education and total intracranial volume of subjects as covariates.GRF was used for multiple comparison correction.The threshold values for evaluating the differences between groups were P<0.05 for the cluster level and z>2.807 for the voxel level.After that,the brain regions with statistical significance were extracted as ROI,and the gray matter volume value of corresponding brain regions were extracted.The relationship between gray matter volume and the Y-BOCS scores was explored by correlation analysis.P<0.05 was considered as statistically significant.Patients with OCD were divided into three groups according to the severity of symptoms,and the correlation between abnormal gray matter volume and Y-BOCS scores in different subgroups was further analyzed.Results1.Compared with healthy controls,the gray matter volume of bilateral cingulate gyrus(mainly in medial cingulate and posterior cingulate),bilateral precuneus lobe and left paracentral lobule was significantly reduced in the first-episode and treatment-naive OCD patients(P<0.05 for the cluster level and z>2.807 for the voxel level,GRF correction).2.No significant region with increase in gray matter volume was found in patients with first-episode and treatment-naive OCD compared to healthy controls.3.In patients with moderate OCD,the gray matter volume value of the left paracentral lobule was negatively correlated with the total score of Y-BOCS(r=0.361,P=0.013).ConclusionOCD patients have damage in the gray matter structure of brain regions related to high emotion and cognition,which is presumed to be related to abnormal top-down emotion,behavior,and cognitive regulation,suggesting the gray matter structural basis of the possible pathophysiological process of OCD.Part 3 Changes of white matter microstructure in first-episode OCD patients based on voxel level analysisBackground and ObjectiveBased on the above analysis of the differences in the structure and function of the gray matter,we found that there were abnormalities in the functional activities and the structure of the gray matter of distant brain regions in patients with OCD,but the changes in the subcortical anatomical connection in patients with OCD remained unclear.The study of white matter structural abnormalities in patients with OCD is very important for further understanding of the neural mechanisms associated with OCD.DTI provides the information on the integrity and microstructure of white matter based on the quantitative diffusion of water molecules in brain tissue.The fractional anisotropy(FA)is the main measurement index of DTI.It is suggested that the findings of voxel-based analysis(VBA)usually involve a wider range of brain regions than those of tract-based spatial statistics(TBSS).Therefore,the purpose of this study is to use whole-brain level VBA to explore the changes of white matter microstructure in patients with first-episode and treatment-naive OCD and the relationship between the changes of abnormal white matter microstructure and obsessive-compulsive symptoms.Materials and MethodsIn this part,a total of 91 patients with first-episode and treatment-naive OCD were enrolled with 106 healthy volunteers(normal control group)whose age,gender,and years of education matched the OCD group.The magnetic resonance data was collected by a magnetic resonance physician who was skilled in MRI scanning on a GE Discovery 750 3.0T magnetic resonance scanner with an 8-channel phased array head coil.1.DTI was obtained by single shot EPI sequence.2.Two psychiatrists trained in scale consistency conducted clinical interviews and the Y-BOCS was used to evaluate obsessive and compulsive symptoms.3.Firstly,the dcm2niigui software based on MRIcron was used to convert DICOM format file of original images to NIFTI format.Then the FMRIB Software Library(FSL)based on Linux system was used to the eddy current correction of the converted data.The image quality was checked immediately after correction to prevent the loss of signal after eddy current correction.Then the corrected images were stripped to remove the scalp and other non-brain tissue structures to improve the accuracy of the registration process.The quality of the images after skull stripping was also checked so as to find the problematic images in time.Then the dispersion index of DTI was calculated.The single exponential dispersion model was used to fit the single Gaussian dispersion tensor according to the principle of least square method.The dispersion eigenvalue and eigenvector were extracted,and FA was calculated according to the dispersion eigenvalue.Then,all individual FA images were registered to the FA template of MNI space using nonlinear registration,and resampled by voxels of 2 mm×2 mm×2 mm.Finally,a Gaussian kernal with the FWHM of 6 mm was used to smooth the normalized FA image.4.The SPSS 21.0 was used for statistical analysis of clinical data of subjects.Student’s t-tests were used to analyse the differences between groups of normally distributed data.Nonparametric tests were used to analyse the differences between groups of data that did not conform to a normal distribution.P<0.05 was considered as statistically significant.The differences in FA value between OCD patients and normal control group were conducted by student’s t-tests,with age,gender,and years of education of subjects as covariates.Gaussian random field(GRF)was used for multiple comparison correction.The threshold values for evaluating the differences between groups were P<0.05 for the cluster level and z>2.807 for the voxel level.After that,the brain regions with statistical significance were extracted as ROI,and the FA value of corresponding brain regions were extracted.The relationship between the FA value and the Y-BOCS scores was explored by correlation analysis.P<0.05 was considered as statistically significant.Patients with OCD were divided into three groups according to the severity of symptoms,and the correlation between abnormal FA value and Y-BOCS scores in different subgroups was further analyzed.Results1.Compared with healthy controls,first-episode and treatment-naive OCD showed significantly reduced FA values in the left insular lobe and right posterior cerebellar lobe(P<0.05 for the cluster level and z>2.807 for the voxel level,GRF corrected).2.No significant increase in FA values was found in patients with first-episode and treatment-naive OCD compared to healthy controls.3.In patients with severe OCD,the FA value of the right posterior cerebellar lobe was negatively correlated with the total score of Y-BOCS(r=-0.384,P=0.016).ConclusionsPatients with first-episode and treatment-naive OCD have abnormal white matter microstructures in several brain regions related to high emotion and cognition,which is speculated to be related to the dynamic interaction of emotion,cognition and motivation,suggesting the white matter structural basis of the possible pathophysiological process of OCD.