Event Abstract

Back to Event

Assessing cluster validity in coordinate-based meta-analysis for fMRI

Freya Acar1*, Ruth Seurinck1, Simon B. Eickhoff2, 3 and Beatrijs Moerkerke1
  • 1 Ghent University, Data Analysis, Belgium
  • 2 Heinrich-Heine University Düsseldorf, Institute for Systems Neuroscience, Medical Faculty, Germany
  • 3 Research Center Jülich, Institute for Neuroscience and Medicine (INM-7), Germany

Meta-analyses for fMRI are booming but remain challenging, given the complex structure of the data and the often censored reporting of results. Because many fMRI studies only report locations of peak voxels that survive a statistical threshold, coordinate-based methods (CBMA), such as ALE (Eickhoff et. al, 2009; Eickhoff et. al, 2012; Turkeltaub et. al, 2012), have been developed. Meta-analyses are in general prone to publication bias or the systematic difference between results of published and unpublished studies, with certain studies remaining in the file-drawer (Rosenthal, 1979), but fMRI studies may be particularly susceptible to small sample bias. fMRI studies with small sample sizes may tend to employ more lenient thresholding to compensate for underpowered tests. In classical meta-analyses this is assessed by regressing the observed effect sizes of studies on the sample sizes (Egger, Davey Smith, Schneider & Minder, 1997). The limited amount of information that serves as input for CBMA prohibits the use of classical techniques to assess and correct for publication bias. In this study, we aim to assess the validity of clusters resulting from an ALE meta-analysis. We first propose a method to verify the robustness of the clusters and assess the file-drawer problem by computing the Fail-Safe N, which is the amount of null studies (i.e. studies that do not contribute to the activation of the cluster) that can be added before the cluster is no longer statistically significant. In a second stage we provide a method to assess small sample bias and illustrate observed patterns under small sample bias scenario's. The presence of small sample bias in the form of lenient thresholding alters the ALE results and reveals distinct patterns in the regression of sample sizes on study contribution, with smaller slopes in meta-analyses that suffer from small sample bias.

Acknowledgements

Authors FA, RS and BM would like to acknowledge the Research Foundation Flanders (FWO) for financial support (Grant G.0149.14).

References

Eickhoff SB, Laird AR, Grefkes C, Wang LE, Zilles K, and Fox PT (2009). Coordinate- based activation likelihood estimation meta-analysis of neuroimaging data: A random-effects approach based on empirical estimates of spatial uncertainty. Human Brain Mapping, 30, 2907– 2926.
Eickhoff SB, Bzdok D, Laird AR, Kurth F, and Fox PT (2012). Activation likelihood esti- mation revisited. NeuroImage, 59, 2349–2361.
Turkeltaub PE, Eickhoff SB, Laird AR, Fox M, Wiener M, and Fox P (2012). Minimiz- ing within-experiment and within-group effects in activation likelihood estimation meta-analyses. Human Brain Mapping, 33, 1–13.
Rosenthal R (1979). The File Drawer Problem and Tolerance for Null Results. Psychological Bulletin, 86 (3), 638–641.
Egger M, Davey Smith G, Schneider M, and Minder C (1997). Bias in meta-analysis de- tected by a simple, graphical test. British Medical Journal, 315, 629–634.

Keywords: fMRI methods, Meta-analysis, Publication Bias, coordinate-based meta-analysis, Validity and Reliability, activation likelihood estimation (ALE) meta-analysis, fMRI

Conference: 12th National Congress of the Belgian Society for Neuroscience, Gent, Belgium, 22 May - 22 May, 2017.

Presentation Type: Poster Presentation

Topic: Novel Methods and Technology Development

Citation: Acar F, Seurinck R, Eickhoff SB and Moerkerke B (2019). Assessing cluster validity in coordinate-based meta-analysis for fMRI. Front. Neurosci. Conference Abstract: 12th National Congress of the Belgian Society for Neuroscience. doi: 10.3389/conf.fnins.2017.94.00042

Copyright: The abstracts in this collection have not been subject to any Frontiers peer review or checks, and are not endorsed by Frontiers. They are made available through the Frontiers publishing platform as a service to conference organizers and presenters.

The copyright in the individual abstracts is owned by the author of each abstract or his/her employer unless otherwise stated.

Each abstract, as well as the collection of abstracts, are published under a Creative Commons CC-BY 4.0 (attribution) licence (https://creativecommons.org/licenses/by/4.0/) and may thus be reproduced, translated, adapted and be the subject of derivative works provided the authors and Frontiers are attributed.

For Frontiers’ terms and conditions please see https://www.frontiersin.org/legal/terms-and-conditions.

Received: 27 Apr 2017; Published Online: 25 Jan 2019.

* Correspondence: Miss. Freya Acar, Ghent University, Data Analysis, Gent, [Select a State], 9000, Belgium, freya.acar@ugent.be