False positives in neuroimaging genetics using voxel-based morphometry data

Department of Mathematics, Imperial College London, London, UK.
NeuroImage (Impact Factor: 6.36). 01/2011; 54(2):992-1000. DOI: 10.1016/j.neuroimage.2010.08.049
Source: PubMed


Voxel-wise statistical inference is commonly used to identify significant experimental effects or group differences in both functional and structural studies of the living brain. Tests based on the size of spatially extended clusters of contiguous suprathreshold voxels are also widely used due to their typically increased statistical power. In “imaging genetics”, such tests are used to identify regions of the brain that are associated with genetic variation. However, concerns have been raised about the adequate control of rejection rates in studies of this type. A previous study tested the effect of a set of ‘null’ SNPs on brain structure and function, and found that false positive rates were well-controlled. However, no similar analysis of false positive rates in an imaging genetic study using cluster size inference has yet been undertaken.

Download full-text


Available from: Thomas Nichols
  • Source
    • "Scarpazza et al. [13] for example used freely available anatomical images from 396 healthy controls [6] to investigate the validity of parametric statistical methods for voxel based morphometry [14], when comparing a single subject to a group. Silver et al. [15] instead used image and genotype data from 181 subjects in the Alzheimer's disease neuroimaging initiative (ADNI) [10] [11]. The data were used to evaluate statistical methods common in imaging genetics, where the goal is to find genes that can explain variation in brain structure or function. "
    [Show abstract] [Hide abstract]
    ABSTRACT: The most widely used task fMRI analyses use parametric methods that depend on a variety of assumptions. While individual aspects of these fMRI models have been evaluated, they have not been evaluated in a comprehensive manner with empirical data. In this work, a total of 2 million random task fMRI group analyses have been performed using resting state fMRI data, to compute empirical familywise error rates for the software packages SPM, FSL and AFNI, as well as a standard non-parametric permutation method. While there is some variation, for a nominal familywise error rate of 5% the parametric statistical methods are shown to be conservative for voxel-wise inference and invalid for cluster-wise inference; in particular, cluster size inference with a cluster defining threshold of p = 0.01 generates familywise error rates up to 60%. We conduct a number of follow up analyses and investigations that suggest the cause of the invalid cluster inferences is spatial auto correlation functions that do not follow the assumed Gaussian shape. By comparison, the non-parametric permutation test, which is based on a small number of assumptions, is found to produce valid results for voxel as well as cluster wise inference. Using real task data, we compare the results between one parametric method and the permutation test, and find stark differences in the conclusions drawn between the two using cluster inference. These findings speak to the need of validating the statistical methods being used in the neuroimaging field.
    Full-text · Article · Nov 2015
  • Source
    • "Due to structural images displaying local variation in smoothness , a non-stationary cluster extent correction was then applied when calculating family-wise error (FWE-cluster) corrected p values (p < .05) using the NS toolbox (Hayasaka et al., 2004; Meisenzahl et al., 2008; Silver et al., 2011; Cullen et al., 2013). For completeness, we also report the p-value for a family-wise error correction at the voxel level (FWE-voxel) (Bennett et al., 2009). "
    [Show abstract] [Hide abstract]
    ABSTRACT: Female participants have been underrepresented in previous structural magnetic resonance imaging reports on attention-deficit/hyperactivity disorder (ADHD). In this study, we used optimized voxel-based morphometry to examine grey matter volumes in a sample of 33 never-medicated children with combined-type ADHD and 27 typically developing (TD) children. We found a gender-by-diagnosis interaction effect in the ventral anterior cingulate cortex (ACC), whereby boys with ADHD exhibited reduced volumes compared with TD boys, while girls with ADHD showed increased volumes when compared with TD girls. Considering the key role played by the ventral ACC in emotional regulation, we discuss the potential contribution of these alterations to gender-specific symptoms' profiles in ADHD. Copyright © 2015 The Authors. Published by Elsevier Ltd.. All rights reserved.
    Full-text · Article · Jun 2015 · Developmental Cognitive Neuroscience
  • Source
    • "The images were modulated with the individual Jacobian determinants to preserve the local amount of gray matter (GM) and white matter (WM) (Keller et al., 2004). For whole-brain analyses, tables and maps were thresholded at p ¼ 0.001 and cluster-size of 10 (Silver et al., 2011). Significant clusters were identified by nonstationary cluster extent correction using random fields (Hayasaka et al., 2004) as implemented using the NS-toolbox ( "
    [Show abstract] [Hide abstract]
    ABSTRACT: Social anxiety disorder (SAD) has received relatively little attention in neurobiological studies. We sought to identify neuro-anatomical changes associated with successful treatment for the disorder. Fourteen patients (31 years; 57% female) with DSM-IV generalized SAD were imaged before and after 8-weeks of paroxetine treatment on a 1.5T GE Signa MRI scanner. Symptoms were assessed by a clinician using the Liebowitz Social Anxiety Scale (LSAS). Longitudinal changes in voxel based morphometry (VBM) were determined using the VBM8 Toolbox for SPM8. Symptom severity decreased by 46% following treatment (p<0.001). At week 8, significant gray matter reductions were detected in bilateral caudate and putamen, and right thalamus, and increases in the cerebellum. Gray matter decreases in left thalamus were correlated with clinical response. This is the first study to our knowledge to identify treatment related correlates of symptom improvement for SAD. Replication in larger samples with control groups is needed to confirm these findings, as well as to test their specificity and temporal stability. Copyright © 2015 Elsevier Ireland Ltd. All rights reserved.
    Full-text · Article · Jan 2015 · Psychiatry Research: Neuroimaging
Show more