Go figure: transparency in neuroscience images preserves context and clarifies interpretation.

Abstract

Visualizations are vital for communicating scientific results. Historically, brain images in neuroscience figures have only depicted regions that surpass a given statistical threshold. This practice substantially biases interpretation of the results and subsequent meta-analyses, particularly toward non-reproducibility. We advocate for 'transparent thresholding' that not only highlights statistically significant regions but also includes subthreshold locations, retaining key experimental context. This balances the dual needs of distilling modeling results and enabling informed interpretations of neuroimaging results. Transparent thresholding removes ambiguity, decreases hypersensitivity to nonphysiological features, helps catch potential artifacts, improves cross-study comparisons, reduces non-reproducibility biases and clarifies interpretations. We address considerations raised by researchers in the field and highlight the many software packages implementing transparent thresholding, several of which were added or streamlined as part of this work. We hope that by showing how transparent thresholding can meaningfully improve the interpretation (and reproducibility) of neuroimaging findings, we will encourage more researchers to adopt this method.