Multimode MRI Research On Brain Of Temporal Lobe Epilepsy Individuals In The Resting State

Part one:White matter microstructural damage in temporal lobe epilepsy individuals DTI-TBSS and structure network studyObjectiveThe purpose of present study is to explore white matter microstructure abnormalities of temporal lobe epilepsy (TLE) by using diffusion tensor imaging (DTI) and tract based spatial statistic (TBSS) methods, and the differences between two index of DTI (FA value and MD value) on evaluating white matter microstructure of TLE, in order to setup image basis for clinical diagnosis, analysis, assessment of treatment on TLE. In addition, we explored the change of "small world" properties and their clinical significance through the topology network of brain that created through DTI data.Materials and Methods1. SubjectsFifty subjects have participated in present study, including26TLE patients and24normal adults. All patients were recruited from Guangdong999hospital, and underwent MR scan, Video-EEG monitoring, scalp EEG and/or sphenoidal electrodes. The inclusion criteria included the following:①seizure types and epileptic syndromes as diagnosed according to the classification of the International League Against Epilepsy (Anon.,1981,1989);②TLE diagnosis when continuous interictal-ictal scalp video electroencephalography showed interical epileptiform discharges (IEDs) of temporal origin. Unilateral or bilateral hippocampal sclerosis, focal cortical dysplasia (FCD) of temporal lobe were showed on MR imaging. Main clinical symptoms as follows:hoven, dyspnea and dither were indicators of a possible medial TLE, acousma, psychotic symptoms, somatosensory abnormality were indicators of a possible lateral TLE. The final diagnosis of TLE from patients’clinical performance and electro-cephalographic results. The exclusive criteria:with ace-occupying lesion (such as tumor, parasite and vascular malformation), white matter lesion, encephalomalacia on MRI, EEG showed suspicious abnormality, the results of EEG and clinical performance were not consistent. Eight subjects were excluded from this study (including6TLE patients and2health controls), remaining42subjects were included in this study, which included20TLE patients (mean age=25.55±8.67years,16males,4females),22health controls (mean age=25.23±6.67years,13males,9females).2. DTI-MRI data acquisitionDTI-MRI data were collected using a1.5-Tesla scanner (Philips Gyroscan Intera) with a6channel neurovascular (NV) coil to receive the signal. Firstly, axial T1-weighted, T2-weighted images and FLAIR images were acquired for detecting incranial lesions. A Single-shot turbo spin echo sequence was used to obtain the DTI data, TR (repetition time)=11000ms; TE (echo time)=71.614ms; flip angle=90°; matrix size=144×144; FOV (field of view)=230X230; NEX=1; thickness=2mm; slice=67; slice gap=0. Thirty-two diffusion gradient directions with b-value=800s/mm2. In addition, images without diffusion weighting were acquired corresponding to b-value=0.3. DTI data preprocessing and TBSS analysisThe DICOM data format of B0image and32diffusion weighted images of each subject were converted to NIFTI data format with four dimensionDTI data preprocessing was partly carried out using FSL (FMRIB Software Library, www.fmrib.ox.ac.uk/fsl, version4.19). The procedure of DTI data process included:①Diffusion tensor images were corrected for head movement by using FDT tool of the FSL software.②Mask for each brain was created by using each subject’Bo image with BET tool of the FSL software.③Fractional anisotropy (FA) and mean diffusivity (MD) were obtained with diffusion tensor calculation by using dtifit function, the FA and MD output images were used as input for tract based spatial statistic (TBSS).4. Brain structure network graph visualizationDTI data preprocessing was partly carried out using SPM5software (http://www.fil. ion.ucl.ac.uk/spm) based on Matlab R2009b. The reprocessing as follows:①Dsta with DICOM format were converted to NIFTI format;②Each individual structural image (Ti-weighted image) was coregistered to the Bo image in the DTI space using a linear transformation, the transformed structural image was then mapped to the T1template in the Montreal Neurological Institute (MNI) space using a nonlinear transformation;③The resulting inverse transformation was then used to warp the AAL-90mask from the MNI space to the DTI native space;④The diffusion tensor matrix was then calculated voxel-by-voxel and diagonalization was performed to yield eigenvalues;⑤Fiber bundles of white matter were reconstructed by using FACT algorithm with DTIstudio software;⑥The topological properties of brain structural networks were defined on the basis of a90X90binary graph, which was consisted of nodes and edges, the edges between nodes could be constructed by applying a correlation matrice, the regional centroid of each node was positioned by using Brainnet viewer Version1.1software.5. Statistic analysisTwo-sample t-test was performed with age, years of education between TLE patients and controls, chi-square test was performed with sex between TLE patients and controls. Two-sample t-test was performed with whole brain mean FA value and MD value between TLE patients and controls, we used a statistical significance level of P<0.05. Then, we used Pearson correlation analysis to investigate the underlying relationship between mean FA value of TLE patients and epilepsy duration.Two-sample t-test was performed with network topological measures(λ、σ、Cp、 Lp、Eglob、Eloc、Cost) between TLE patients and controls. Then, we used Pearson correlation analysis to investigate the underlying relationship between properties measures (Eglob, Cost) of the brain structure networks and epilepsy duration (P<0.05).Results1. The results of TBSS:Reduced FA values were found in many brain regions, which included bilateral temporal lobes, frontal lobes, parietal lobes and part of the occipital lobes, bilateral internal capsule, centrum semiovale (P<0.05). Increased MD values were found almost in left hemisphere, including left frontal lobe, temporal lobe, parietal lobe and occipital lobe (P<0.05). Moreover, it has a significant negative correlation (r=-0.5535, P=0.0114) for the mean FA value in FA skeleton of each patient with the epilepsy duration.2. The results of structural network topological measures:Compared to normal controls, the global efficiency (Eglob) showed significantly decreased in TLE patients, and the local efficiency (Eglob) showed significantly decreased also in TLE patients, which included left middle frontal gyrus, bilateral inferior occipital gyrus, left angular gyrus, left Heschl gyrus, right anterior cingulate and paracingulate gyri. Moreover, the shortest path length (t=-4.18, P=1.55e-04) and cost (t=-4.04, P=2.35e-04) showed significantly increased in TLE. It also showed a significant negative correlation (r=-0.63, P=0.0053) for Eglob of left Heschl gyrus in patients with the epilepsy duration. The TLE patients showed statistically significant decrease inEglob, Eloc, and increase in Lp, Cost, compared to health controls. However, it has no statistically significant difference for Cp between TLE patients and controls. ConclusionIt offered the evidence that there have extensive damages of brain white matter on TLE patients, and the severity of this damage increased with the epilepsy duration. Moreover, it seen to be more sensitivity to damage for left hemisphere than right hemisphere. Though graph-theory analysis, we found that the structural network of TLE patients also have "small-word" attributes. However, the topology of structural network parameters have altered remarkably in TLE patients compared to health controls, which included increase of Cp and Cost, decrease of Eglob and Eloc, the location of brain Hubs have also altered. This demonstrated that there have extensive damages of brain white matter on TLE patients, which mean decrease in efficiency of information transfer, that would be related to hypothymergasia, memory losses, learning ability and cognitive decline in TLE patients. Part two:Research on brain functional network of temporal lobe epilepsy via BOLD-fMRI and graph theoryObjectiveTo investigate alterations of functional connectivity and small-world topological properties related to temporal lobe epilepsy through functional connection technology and graph-theory based on blood oxygen level-dependent functional MRI and its clinical significance.Materials and Methods1. SubjectsSixty-three subjects participated in the study, including35TLE patients and28 health adults, all TLE patients were recruited from Guangdong999hospital. The inclusion criteria included the following:①seizure types and epileptic syndromes as diagnosed according to the classification of the International League Against Epilepsy (Anon.,1981,1989);②TLE diagnosis when continuous interictal-ictal scalp video electroencephalography showed interical epileptiform discharges (IEDs) of unilateral temporal origin. The exclusive criteria:with ace-occupying lesion (such as tumor, parasite and vascular malformation), white matter lesion, encephalomalacia on MR image, suspicious abnormality of the results for EEG, the results of EEG suggested bilateral TLE, or not consistent to clinical performance of TLE patients. The inclusion criteria for health controls was as same as part one. Nine participants were excluded from this study, remaining54subjects were included in this study, which included25health controls (mean age=24.24±5.31years,17males,8females),16left TLE patients (mean age=23.13±7.14years,11males,5females),13right TLE patients (mean age=26.31±10.10,8males,5females).2. Data acquisitionMRI data were collected using a1.5-Tesla scanner, participants were instructed to rest with their eyes closed and to be still, without considering the specific problems, and not to fall asleep. BOLD functional images covering the whole brain were acquired axially using an gradient echo-echo planar imaging sequence (GRE-EPI), TR=3000ms, TE=50ms, flip angle=90, matrix=128×128, FOV=230mm×230mm, slice=4.5mm, slice gap=0. For each subjects, the resting state fMRI scanning lasted seven minutes, thus collecting160volumes.3. Data processing3.1Data preprocessingData preprocessing was partly carried out using SPM5software (http://www.fil. ion.ucl.ac.uk/spm) based on Matlab R2009b. The reprocessing as follows:①Data with DICOM format were converted to NIFTI format;②The first10images were discarded to ensure the magnetization equilibrium;③Then the remaining images were realigned, the subjects would be excluded for head translation or rotation exceeded±1mm or±1°;④lice timing, correct differences in image acquisition time between slices;⑤Normalization, normalize images into a standard Talairach and Tournoux space by MNI template images which supplied with SPM5;⑥The images were proceed with linear detrend and low-frequency filtering, then the preprocessed time series were used for further functional connectivity and graph-theory analysis.3.2Image segmentThe images were segmented into90anatomical regions of interests (ROIs)(45ROIs for each hemisphere) using anatomically labeled-90(AAL-90) template. These anatomical ROIs were extracted by the MarsBaR toolbox (http://marsbar.sourceforge. net)3.3Computation of correlation matrixThe resting state BOLD time series were correlated region by region for each subject across the full length of the resting time series. Then a square90X90correlation matrix was obtained for each subject,4005(C920=90X89/2=4005) inter-regional correlations were subjected to statistic analysis.3.4functional network graph visualizationThe topological properties of the brain functional networks were defined on the basis of a90X90binary graph, which was consisted of nodes and edges, the edges between nodes could be constructed by applying a correlation matrice, the regional centroid of each node was positioned by using Brainnet viewer Version1.1software.4. Statistic analysis4.1Function connectivity comparisonTwo-sample t-test was performed with functional connectivity between left TLE patients and controls, and right TLE patients and controls base on matlab R2009b. To account for multiple comparisons, the false discovery rate method was applied, P﹤0.001was supposed to be a significantly functional connectivity. 4.2Network topological measures comparisonTwo-sample t-test was performed with network topological measures between left TLE patients and controls, and right TLE patients and controls base on matlab2009b. To account for multiple comparisons, the false discovery rate method was applied, P﹤0.001was supposed to be a significantly functional connectivity.Results1. Results of functional connectivity:Left TLE patients produced significantly stronger connectivity than healthy controls between specific ROIs, e.g. right precentral gyrus vs right hippocampus; right inferior frontal gyrus orbital part vs right postcentral gyrus, and produced significantly lower connectivity than healthy controls between left posterior cingulated gyrus vs left thalamus. Right TLE patients produced significantly stronger connectivity than healthy controls between specific ROIs, e.g. left precentral gyrus vs left parahippocampal gyrus; left dorsolateral superior frontal gyrus vs left olfactory cortex, and produced significantly lower connectivity than healthy controls between right medial superior frontal gyrus vs temporal pole of right middle temporal gyrus; left posterior cingulated gyrus vs temporal pole of right middle temporal gyrus; right hippocampus vs right caudate nucleus.2. Results of brain functional network:2.1Clustering coefficients (Cp) and local efficiency (Eloc) showed significantly larger value in left TLE patients when compared to normal controls. However, there have no alteration with graph theory measures between right TLE and normal controls.2.2In the left TLE patients, some nodal parameters showed significant increase, e.g. right middle temporal gyrus; right parahippocampal gyrus, and some nodal parameters showed significant decrease, e.g. right thalamus, left middle frontal gyrus orbital part, especially for left middle frontal gyrus orbital part (P=0.0057). In the right TLE patients, some nodal parameters showed significant increase, e.g. right inferior parietal gyrus, bilateral angular gyrus, right postcentral gyrus, especially for right inferior parietal gyrus (P=6.8817e-4). However, the left inferior temporal gyrus showed significant decrease with nodal parameters.ConclusionIt showed that there have alterations of functional connectivity between many brain regions both in left TLEs or right TLEs, though graph-theory analysis, we found that the clustering coefficients (Cp) and local efficiency (Eloc) showed significantly larger value in left TLE patients when compared to normal controls. However, there has no alteration with graph theory measures between right TLE and normal controls. Furthermore, the nodal parameters of right thalamus and left middle frontal gyrus orbital part showed significant decrease in left TLE patients. It demonstrated that the damages of brain with left TLE were more serious than right TLE. Furthermore, it suggested that the alternation of functional connectivity and nodal parameters would be related to memory losses, emotional changes and cognitive decline in TLE patients.