Spm8 manual vbm

Our findings show that mainly segmentation but also normalization procedures have a profound influence on VBM results. Even with a rather small sample as tested here, differences between methods reached significance which is important to consider when comparing studies that used different software packages for VBM analyses.

Irrespective of which segmentation method was used, we found lower GM and higher CSF volume in the elderly as compared to the young after correction for ICV. This result was highly expected based on previous literature Good et al. However, regarding GM volume, we found large differences across segmentation methods when quantifying the effect size of the age-related decline: SPM tended to over-estimate age specific differences, while DARTEL tended to under-estimate these structural changes when compared to the other methods.

One has to note however, that we tested two groups at the extreme end of the age range and this selection contributed to the large difference between SPM and DARTEL segmentation since the analysis of Peelle et al.

Our analysis suggests that methods using a-priory probability maps during the segmentation algorithm i. However, the largest part of the volume loss was driven by a profound reduction of high GM probability voxels.

All approaches revealed an increased CSF volume in the elderly vs. Although CSF volume is generally not at the core interest of VBM studies, this overestimation might indirectly influence the results when incorporated for the calculation of a covariate such as ICV Buckner et al. Typically this covariate is required when assessing tissue volumes relative to global brain volume and the advised method is to sum up thresholded GM, WM and CSF probability maps.

One common workaround is to approximate total intracranial volume by summing GM and WM volume whilst omitting CSF or by using other covariates depending on which aspect of age related GM changes is investigated Peelle et al.

We showed that method specific differences resulted from the segmentation and were particularly pronounced for the cortical surface close to large sulci and fissures. Method specific deviations were not only observed for voxel wise comparisons but also when anatomical regions of interest were used. Warping between the individual anatomy and a template requires non-linear, local deformations particularly around ventricles, sulci and fissures that are typically enlarged in elderly Blatter et al.

This presents a challenge for normalization procedures. However, it might be advisable to use a customized template when extreme age-groups are compared, as in our study. In agreement with our results, Klein et al. Alternatively, advanced segmentation methods can be used to avoid an age—related bias Ziegler et al. In summary, differences in segmentation algorithms that often rely on an initial spatial normalization step seem to be a major source for between-method differences when VBM is used to compare the brain structure of young vs.

We suggest that this effect is particularly pronounced when subjects have deviant anatomy as it is typically the case for elderly individuals and normalization employs only limited degrees of freedom for local deformations. Therefore, differences in the normalization algorithm might significantly influence the segmentation step when prior tissue probability maps are used, but also the localization of significant GM differences when individuals are compared within a common template space.

In this study we investigated whether different VBM procedures have a significant influence on GM estimates of young vs. Even though we found substantial differences between methods we cannot infer which VBM method is the most correct one since it is difficult to determine the ground truth for voxel-based methods. Typically this important step for method validation is done by the developers and published when a new method is introduced see for example Ashburner and Friston, However, it is rarely the case that previous datasets are reanalyzed what makes comparison across different studies difficult.

Here we report that age-specific GM differences are substantial across methods and should be considered when different findings are compared. Another caveat is the fact that our sample was relatively small and drawn from the extremes of the age spectrum.

Method specific differences are probably less pronounced when a smaller age range is considered Peelle et al. Finally, we applied parametric statistics whilst non-parametric approaches might be generally more appropriate for VBM studies Douaud et al. One has to keep in mind, though, that we were not interested in reporting age-specific GM differences per se , but rather whether these results depended on the algorithms provided by software packages frequently used by the neuroimaging community.

Morphometric measurements are increasingly applied for the detection of GM changes in healthy ageing as well as in neurodegenerative disease. Due to its automated and near hypothesis-free character, VBM has gained popularity as a substitute for manual demarcations of GM within volumes of interest. Here, we used the aging brain as a well-known model for structural atrophy and asked whether comparing GM between young and elderly by VBM depends on methodological differences between commonly used software packages, i.

These method-specific differences reached significance when tested at various levels of description total brain volume, regions of interest, voxel based. We argue that the segmentation procedure can have a major influence on cross-sectional VBM results, particularly when anatomical deviations are more outspoken in one group than the other. The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

We would like to thank Prof. Patrick Dupont for his valuable comments. Allen, J. Normal neuroanatomical variation due to age: the major lobes and a parcellation of the temporal region. Aging 26, — Methods for studying the aging brain: volumetric analyses versus VBM Commentary.

CrossRef Full Text. Andersson, J. Available online at: www. Ashburner, J. Diffeomorphic registration using geodesic shooting and Gauss-Newton optimisation. Neuroimage 55, — A fast diffeomorphic image registration algorithm. Neuroimage 38, 95— Voxel-based morphometry-the methods.

Neuroimage 11, — Why voxel-based morphometry should be used. Neuroimage 14, — New York, NY: Elsevier. Unified segmentation. Neuroimage 26, — Computing average shaped tissue probability templates. Neuroimage 45, — Blatter, D. Quantitative volumetric analysis of brain MR: normative database spanning 5 decades of life. AJNR Am. Pubmed Abstract Pubmed Full Text. Bookstein, F. Voxel-based morphometry should not be used with imperfectly registered images. Buckner, R. A unified approach for morphometric and functional data analysis in young, old, and demented adults using automated atlas-based head size normalization: reliability and validation against manual measurement of total intracranial volume.

Neuroimage 23, — Chetelat, G. Using voxel-based morphometry to map the structural changes associated with rapid conversion in MCI: a longitudinal MRI study. Neuroimage 27, — Courchesne, E. Normal brain development and aging: quantitative analysis at in vivo MR imaging in healthy volunteers. Radiology , — D'Agostino, E. A viscous fluid model for multimodal non-rigid image registration using mutual information.

An information theoretic approach for non-rigid image registration using voxel class probabilities. A unified framework for atlas based brain image segmentation and registration. Davatzikos, C. Why voxel-based morphometric analysis should be used with great caution when characterizing group differences. Neuroimage 23, 17— Davies, R. Development of an MRI rating scale for multiple brain regions: comparison with volumetrics and with voxel-based morphometry.

Neuroradiology 51, — Douaud, G. Distribution of gray matter atrophy in Huntington's disease patients: a combined ROI-based and voxel-based morphometric study. Neuroimage 32, — Draganski, B. Neuroplasticity: changes in gray matter induced by training.

Nature , — Ferreira, L. Neurostructural predictors of Alzheimer's disease: a meta-analysis of VBM studies. Aging 32, — Fjell, A. High consistency of regional cortical thinning in aging across multiple samples. Cortex 19, — Friston, K. Statistical parametric maps in functional imaging: a general linear approach. Brain Mapp. Giuliani, N. Voxel-based morphometry versus region of interest: a comparison of two methods for analyzing gray matter differences in schizophrenia.

Gonoi, W. Age-related changes in regional brain volume evaluated by atlas-based method. Neuroradiology 52, — Good, C. A voxel-based morphometric study of ageing in normal adult human brains. Neuroimage 14, 21— Automatic differentiation of anatomical patterns in the human brain: validation with studies of degenerative dementias. Neuroimage 17, 29— Greenberg, D. Aging, gender, and the elderly adult brain: an examination of analytical strategies.

Aging 29, — Grieve, S. Preservation of limbic and paralimbic structures in aging. Hammers, A. Statistical neuroanatomy of the human inferior frontal gyrus and probabilistic atlas in a standard stereotaxic space. Implementation and application of a brain template for multiple volumes of interest.

Hasan, K. Improving the reliability of manual and automated methods for hippocampal and amygdala volume measurements. Neuroimage 48, — Hutton, C. Compared to the conventional algorithm, the DARTEL approach can provide more precise spatial normalization to the template 3 , 14 , In the present study, as part of the modulation step we performed a non-linear deformation on the normalized segmented images with both the VBM8 and CAT12 toolboxes.

This modulation provides a comparison of the absolute amounts of tissue corrected for individual differences in brain size All segmented, modulated, and normalized GM and WM images were smoothed using 8-mm full-width-half-maximum Gaussian smoothing and then fed into a flexible factorial analysis in SPM8 and SPM12, separately.

The extent threshold was set at voxels. All of the statistical analyses were done using SPSS software, ver. Based on those findings, we decided to use two versions of the widely applied SPM toolbox i. In the VBM analysis using the older toolbox i. In addition, the results we obtained using CAT12 are broadly consistent with the pathology-based knowledge describing neuronal loss in the hippocampus of TLE-HS patients This finding is in agreement with those of earlier studies that demonstrated TLE with amygdala enlargement 24 — One limitation of our study might be that the subject groups were gender imbalanced; the LTLE-HS group in particular was predominantly female, and the healthy controls were mostly male.

In addition, given that statistical significance can sometimes be affected by various factors, we should pay careful attentions to interpreting the significance of the results.

The authors in Ref. As part of a future study, we plan to evaluate the amygdala and hippocampus volumes as the main regions affected by epilepsy, using different approaches such as SPM i. Although in the present investigation we used robust statistics and obtained results that are concordant with past studies, further studies using different samples and methods could be informative.

These analyses provided disparate results. The reason for this may be various improvements of the normalization and segmentation methods provided by SPM12 compared to the older program SPM8. Our findings also demonstrate that brain morphological abnormalities in TLE patients identified using CAT12 are consistent with other studies that investigated the gray- and white-matter abnormalities in TLE using different methods such as optimized VBM 21 and diffusion tensor imaging We suggest that future VBM analyses use the CAT12 toolbox as an advanced neuroimaging procedure in regional volumetric studies.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest. We thank the reviewers for their constructive comments.

National Center for Biotechnology Information , U. Front Neurol. Published online Aug Author information Article notes Copyright and License information Disclaimer. Specialty section: This article was submitted to Epilepsy, a section of the journal Frontiers in Neurology. Received Apr 26; Accepted Aug 8. The use, distribution or reproduction in other forums is permitted, provided the original author s or licensor are credited and that the original publication in this journal is cited, in accordance with accepted academic practice.

No use, distribution or reproduction is permitted which does not comply with these terms. This article has been cited by other articles in PMC. Keywords: voxel-based morphometry, VBM8, CAT12, temporal lobe epilepsy, hippocampus sclerosis, statistical parameter mapping. Introduction Identifying brain morphological changes is a challenging task in neuroimaging studies.

Open in a separate window. Figure 1. Experimental Procedures Data Collection All data used in this study were obtained from the National Center of Neurology and Psychiatry Hospital Tokyo for patients examined during the period from November through January Table 1 Characteristics of the healthy controls and TLE patients.

Figure 2. Anatomical regions were derived from the Talairach Client program. Figure 3. Conflict of Interest Statement The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest. Acknowledgments We thank the reviewers for their constructive comments.

References 1. See also the release notes. The book Statistical Parametric Mapping: The Analysis of Functional Brain Images provides the background and methodology for the analysis of all types of brain imaging data, from functional magnetic resonance imaging to magnetoencephalography. Critically, "Statistical Parametric Mapping" provides a widely accepted conceptual framework which allows treatment of all these different modalities.

The book takes the reader from the basic concepts underlying the analysis of neuroimaging data to cutting edge approaches that would be difficult to find in any other source. The material is presented in an incremental way so that the reader can understand the precedents for each new development.

This book will be particularly useful to neuroscientists engaged in any form of brain mapping; who have to contend with the real-world problems of data analysis and understanding the techniques they are using. It is primarily a scientific treatment and a didactic introduction to the analysis of brain imaging data. It can be used as both a textbook for students and scientists starting to use the techniques, as well as a reference for practicing neuroscientists.