Contributed by matthew.kempton on Feb. 22, 2018
Description: Bipolar Patients > Controls, White matter volume processed with SPM DARTEL, full details in paper: Subjects The VBM analysis included 26 euthymic patients with BD (23 with bipolar I and 3 with bipolar II, 9 males and 17 females) and 23 healthy control subjects (7 males and 16 females). The patients were primarily recruited from a UK patient support group, healthy controls were recruited via advertisements in local media. The study was approved by the local ethics committee and written informed consent was obtained from all participants. All subjects were assessed using the Structural Clinical Interview for DSM-IV Axis I Disorders (SCID-CV). Patients were included if they fulfilled criteria for DSM-IV for BD and did not have any comorbidity for other DSM-IV Axis-I disorders. Healthy controls subjects were selected in order to match BD patients for age, sex, race/ethnicity, weight, height, handedness, premorbid IQ, years of education, lifetime drug and alcohol use. They were included if they had no DSM-IV Axis I disorders and no family history of psychiatric conditions. The mean age was 42.1 (± SD 14.8) for BD patients and 41.2 (± SD 14.0) for healthy controls. Demographic and clinical measures are given in table s1 and s2. MRI acquisition Participants were scanned using a 1.5 Tesla Siemens Magnetom Vision MRI scanner to obtain T1 weighted MPRAGE (Multi-Planar Rapidly Acquired Gradient Echo) scans. In order to confer good resolution and good contrast between grey and white matter in particular, the following parameters were selected: TR = 9.7 ms, TE = 4 ms, TI = 300 ms, Nex = 1, 256 x 192 matrix, flip angle = 8°, 128 slices, voxel size = 1.0 x 1.0 x 2.0mm. There was no significant difference in scan date between patients and controls (p=0.28). VBM DARTEL pre-processing We examined group-related differences in regional brain volume using voxel-based morphometry, as implemented in SPM8 software (http://www.fil.ion.ucl.ac.uk/spm/) running under MATLAB R2012b, version 8.0 (The MathWorks, Icn, Natick, Massachussetts). First the T1-weighted images were pre-processed using the DARTEL (Diffeomorphic Anatomical Registration using Exponentiated Lie algebra) algorithm (Ashburner, 2007) following the steps described by Ashburner (Ashburner, 2010). Firstly, each T1-weighted image was checked for scanner artefacts and gross anatomical abnormalities and then manually reoriented to the Anterior Commissure-Posterior Commissure line blind to diagnosis. The images were then segmented into grey matter, white matter and cerebrospinal fluid in native space. The DARTEL SPM8 toolbox was used to implement the high-dimensional DARTEL normalization through which the DARTEL template was created from the images of all the subjects of the study. During the template creation, flow fields were computed which contain information about the transformation from every native image to the DARTEL template (Peelle et al., 2012). This procedure increases the accuracy of the alignment between subjects by using millions of parameters to characterise the spatial transformations of each brain (Ashburner, 2010). In order to allow for inter-study comparisons, the segmented images were spatially normalized to MNI space including the flow fields in the process. The images were ‘modulated’ to conserve the information on absolute volume. Smoothing was applied to the images using a FWHM 8mm isotropic Gaussian kernel resulting in smoothed, segmented, normalized, and modulated images. VBM Statistical analysis A central aim of the study was to examine the volume of the white matter ROI created by the meta-analysis in an independent sample, however for completeness in the supplementary materials we present the whole VBM brain analysis of the independent dataset. Total intracranial volume was determined for each subject by summing grey matter, white matter and CSF segmentations. The regional differences in voxel-based parameters between BD and controls were assessed using a General Linear Model (GLM) with total intracranial volume and age as covariates of no interest. An absolute threshold masking of 0.05 was adopted in order to exclude voxels outside the brain. A height threshold of p < 0.05 FWE (family wise error) corrected was initially adopted to detect significant regional differences. In addition a more liberal height threshold of p < 0.001, uncorrected for multiple comparisons, was also applied with a cluster threshold of 10 voxels. Following this height threshold, a non-stationary cluster extend correction was implemented at the cluster threshold of p < 0.05 family-wise error (FWE) corrected for multiple comparisons in order to account for the non-isotropic (non-uniform) smoothness across the data (Hayasaka et al., 2004; Worsley et al., 1999). This correction was performed using the VBM8 toolbox (available online at http://dbm.neuro.uni-jena.de/vbm/download). Finally we implemented the same method excluding patients who were taking lithium as studies have demonstrated that lithium may increase total grey matter volume (Hallahan et al., 2011; Kempton et al., 2008; Monkul et al., 2007; Moore et al., 2009; Sassi et al., 2002). Montreal Neurological Institute (MNI) coordinates are reported in the results tables (supplementary table 3 and table 4), however these coordinates were converted to Talairach coordinates to determine the names of corresponding brain regions. MNI coordinates were converted to Talairach using GingerALE, version 2.1.1 (available online at http://www.brainmap.org/ale/) and brain region names were determined using Talairach Client, version 2.4.3 (available online at http://www.talairach.org/client.html). Supplementary Results Independent VBM whole brain study results No significant differences in white or gray matter volume were found at the height threshold of p < 0.05 FWE corrected. The analysis was then repeated with a height threshold of p < 0.001 uncorrected. Regions of significant white matter volume decreases at a height threshold of p<0.001 uncorrected are shown in supplementary table 3. No regions of significant increased white matter in bipolar patients compared to controls were found. Two clusters of voxels survived the additional non-stationary cluster extent threshold of p < 0.05 FWE corrected for multiple comparisons in the white matter results. These clusters encompassed white matter adjacent to the cingulate gyrus and in the corpus callosum (supplementary figure 3). Grey matter volume differences between the two groups are also shown in supplementary table 3. Finally, we found regions of decreased and increased grey matter in bipolar patients that were not taking lithium compared to healthy controls (supplementary table 4). The T-maps of each contrast are freely available to download from www.bipolardatabase.org. The white matter results have been used in the main paper to validate the region of interest found in our meta-analysis.
Tags: vbm bipolar disorder white matter
Task View 3D View| Field | Value |
|---|
If you use these data please include the following persistent identifier in the text of your manuscript:
https://identifiers.org/neurovault.image:61000
This will help to track the use of this data in the literature. In addition, consider also citing the paper related to this collection.