Neuroinformatics

, Volume 10, Issue 1, pp 67–80

Publication Bias in Neuroimaging Research: Implications for Meta-Analyses

Original Article

DOI: 10.1007/s12021-011-9125-y

Cite this article as:
Jennings, R.G. & Van Horn, J.D. Neuroinform (2012) 10: 67. doi:10.1007/s12021-011-9125-y

Abstract

Neuroimaging and the neurosciences have made notable advances in sharing activation results through detailed databases, making meta-analysis of the published research faster and easier. However, the effect of publication bias in these fields has not been previously addressed or accounted for in the developed meta-analytic methods. In this article, we examine publication bias in functional magnetic resonance imaging (fMRI) for tasks involving working memory in the frontal lobes (Brodmann Areas 4, 6, 8, 9, 10, 37, 45, 46, and 47). Seventy-four studies were selected from the literature and the effect of publication bias was examined using a number of regression-based techniques. Pearson’s r correlation coefficient and Cohen’s d effect size estimates were computed for the activation in each study and compared to the study sample size using Egger’s regression, Macaskill’s regression, and the ‘Trim and Fill’ method. Evidence for publication bias was identified in this body of literature (p < 0.01 for each test), generally, though was neither task- nor sub-region-dependent. While we focused our analysis on this subgroup of brain mapping studies, we believe our findings generalize to the brain imaging literature as a whole and databases seeking to curate their collective results. While neuroimaging databases of summary effects are of enormous value to the community, the potential publication bias should be considered when performing meta-analyses based on database contents.

Keywords

Brain imaging fMRI Databases Meta-analysis Publication bias 

Copyright information

© Springer Science+Business Media, LLC (outside the USA) 2011

Authors and Affiliations

  1. 1.Department of BiostatisticsUniversity of California Los AngelesLos AngelesUSA
  2. 2.Laboratory of Neuro Imaging (LONI), Department of Neurology, David Geffen School of MedicineUniversity of California Los AngelesLos AngelesUSA

Personalised recommendations