Distributional Assumptions in Voxel-Based Morphometry

doi:10.1006/nimg.2002.1153

NeuroImage

Volume 17, Issue 2, October 2002, Pages 1027-1030

https://doi.org/10.1006/nimg.2002.1153 Get rights and content

Abstract

In this paper we address the assumptions about the distribution of errors made by voxel-based morphometry. Voxel-based morphometry (VBM) uses the general linear model to construct parametric statistical tests. In order for these statistics to be valid, a small number of assumptions must hold. A key assumption is that the model's error terms are normally distributed. This is usually ensured through the Central Limit Theorem by smoothing the data. However, there is increasing interest in using minimal smoothing (in order to sensitize the analysis to regional differences at a small spatial scale). The validity of such analyses is investigated. In brief, our results indicate that nonnormality in the error terms can be an issue in VBM. However, in balanced designs, provided the data are smoothed with a 4-mm FWHM kernel, nonnormality is sufficiently attenuated to render the tests valid. Unbalanced designs appear to be less robust to violations of normality: a significant number of false positives arise at a smoothing of 4 and 8 mm when comparing a single subject to a group. This is despite the fact that conventional group comparisons appear to be robust, remaining valid even with no smoothing. The implications of the results for researchers using voxel-based morphometry are discussed.

References (10)

J. Ashburner et al.
Multimodal image coregistration and partitioning—A unified framework
Neuroimage
(1997)
F.G. Woermann et al.
Voxel-by-voxel comparison of automatically segmented cerebral gray matter—A rater-independent comparison of structural MRI in patients with epilepsy
Neuroimage
(1999)
J. Ashburner et al.
Voxel based morphometry—The methods
Neuroimage
(1999)
K.J. Friston et al.
Spatial Registration and Normalisation of Images
Hum. Brain Map.
(1995)

There are more references available in the full text version of this article.

Cited by (216)

Parkinson's disease may disrupt overlapping subthalamic nucleus and pallidal motor networks
2023, NeuroImage: Clinical
There is an ongoing debate about differential clinical outcome and associated adverse effects of deep brain stimulation (DBS) in Parkinson’s disease (PD) targeting the subthalamic nucleus (STN) or the globus pallidus pars interna (GPi). Given that functional connectivity profiles suggest beneficial DBS effects within a common network, the empirical evidence about the underlying anatomical circuitry is still scarce. Therefore, we investigate the STN and GPi-associated structural covariance brain patterns in PD patients and healthy controls.
We estimate GPi’s and STN’s whole-brain structural covariance from magnetic resonance imaging (MRI) in a normative mid- to old-age community-dwelling cohort (n = 1184) across maps of grey matter volume, magnetization transfer (MT) saturation, longitudinal relaxation rate (R1), effective transversal relaxation rate (R2*) and effective proton density (PD*). We compare these with the structural covariance estimates in patients with idiopathic PD (n = 32) followed by validation using a reduced size controls’ cohort (n = 32).
In the normative data set, we observed overlapping spatially distributed cortical and subcortical covariance patterns across maps confined to basal ganglia, thalamus, motor, and premotor cortical areas. Only the subcortical and midline motor cortical areas were confirmed in the reduced size cohort. These findings contrasted with the absence of structural covariance with cortical areas in the PD cohort.
We interpret with caution the differential covariance maps of overlapping STN and GPi networks in patients with PD and healthy controls as correlates of motor network disruption. Our study provides face validity to the proposed extension of the currently existing structural covariance methods based on morphometry features to multiparameter MRI sensitive to brain tissue microstructure.
A novel approach for extracting functional brain networks involved in mesial temporal lobe epilepsy based on self organizing maps
2022, Informatics in Medicine Unlocked
We propose a novel data-driven approach to extract and present large-scale functional brain networks from functional magnetic resonance imaging (fMRI) data using spatiotemporal self-organizing maps (STSOM) clustering, accounting for the properties of the brain functional networks being spatially structured and interhemispherically symmetric. Also, a novel group-wise analysis is proposed based on restricted Frechet mean to identify group-level networks. The alteration of resulted networks in left and right mesial temporal lobe epilepsy (mTLE) is studied.
Thirty-five unilateral mTLE patients (21 left-mTLE (LTLE) and 14 right-mTLE (RTLE)), were prospectively studied. Eleven healthy control (HC) subjects were also recruited. To determine the functional networks of the whole brain, we extracted individual and group-level networks using spatiotemporal self-organizing maps and the restricted Frechet mean method, respectively. We applied the resulted networks to specify within and between-network alteration in functional connectivity (FC) in the LTLE and RTLE patients compared to the control cohort.
We obtained seven networks namely default-mode (DMN), sensorimotor (SMN), visual (VSN), subcortical (SCN), frontoparietal (FPN), dorsal attention (DAN), and ventral attention (VAN) networks. Our results demonstrated increased functional connectivity in the FPN networks in the LTLE and the RTLE cohorts compared to HC. Increased FC has been observed between DMN, FPN, DAN, VAN, and VSN in the LTLE cohort and between the DMN and FPN networks in the RTLE cohort.
The proposed method has obtained promising results within a range of SNR and properly overlapped with the well-known functional networks using the Hausdorff distance. The consistent alteration patterns in within-and between-network FC for LTLE and RTLE patient cohorts would reflect the reliability of identification of large-scale brain networks in patients with mTLE. Different pattern of alterations in LTLE and RTLE compare with HC groups my be usefull for laterality porpose.
Machine learning for filtering out false positive grey matter atrophies in single subject voxel based morphometry: A simulation based study
2021, Journal of the Neurological Sciences
Single subject VBM (SS-VBM), has been used as an alternative tool to standard VBM for single case studies. However, it has the disadvantage of producing an excessively large number of false positive detections. In this study we propose a machine learning technique widely used for automated data classification, namely Support Vector Machine (SVM), to refine the findings produced by SS-VBM. A controlled set of experiments was conducted to evaluate the proposed approach using three-dimensional T1 MRI scans from control subjects collected from the publicly available IXI dataset. The scans were artificially atrophied at different locations and with different sizes to mimic the behavior of neurological disorders. Results empirically demonstrated that the proposed method is able to significantly reduce the amount of false positive clusters (p < 0.05), with no statistical differences in the true positive findings (p > 0.05). This evidence was observed to be consistent for different atrophied areas and sizes of atrophies. This approach could be potentially be applied to alleviate the intensive manual analysis that radiologists and clinicians have to perform to filter out miss-detections of SS-VBM, increasing its usability for image reading.
Is Physical Frailty a Neuromuscular Condition?
2019, Journal of the American Medical Directors Association
Statistical challenges of big brain network data
2018, Statistics and Probability Letters
We explore the main characteristics of big brain network data that offer unique statistical challenges. The brain networks are biologically expected to be both sparse and hierarchical. Such unique characterizations put specific topological constraints onto statistical approaches and models we can use effectively. We explore the limitations of the current models used in the field and offer alternative approaches and explain new challenges.
False positive rates in surface-based anatomical analysis
2018, NeuroImage
The false positive rates (FPR) for surface-based group analysis of cortical thickness, surface area, and volume were evaluated for parametric and non-parametric clusterwise correction for multiple comparisons for a range of smoothing levels and cluster-forming thresholds (CFT) using real data under group assignments that should not yield significant results. For whole cortical surface analysis, thickness showed modest inflation in parametric FPRs above the nominal level (10% versus 5%). Surface area and volume FPRs were much higher (20–30%). In the analysis of interhemispheric thickness asymmetries, FPRs were well controlled by parametric correction, but FPRs for surface area and volume asymmetries were still inflated. In all cases, non-parametric permutation adequately controlled the FPRs. It was found that inflated parametric FPRs were caused by violations in the parametric assumptions, namely a heavier-than-Gaussian spatial correlation. The non-Gaussian spatial correlation originates from anatomical features unique to individuals (e.g., a patch of cortex slightly thicker or thinner than average) and is not a by-product of scanning or processing. Thickness performed better than surface area and volume because thickness does not require a Jacobian correction.

View all citing articles on Scopus

View full text

Regular ArticleDistributional Assumptions in Voxel-Based Morphometry

Abstract

Neuroimage

Neuroimage

Voxel based morphometry—The methods

Neuroimage

Spatial Registration and Normalisation of Images

Hum. Brain Map.

Regular Article
Distributional Assumptions in Voxel-Based Morphometry