Thank you for visiting nature.com. You are using a browser version with limited support for CSS. To obtain
the best experience, we recommend you use a more up to date browser (or turn off compatibility mode in
Internet Explorer). In the meantime, to ensure continued support, we are displaying the site without styles
Neuroimaging results altered by varying analysis pipelines
Seventy laboratories that analysed the same neuroimaging data each produced different results. This finding highlights the potential consequences of a lack of standardized pipelines for processing complex data.
For most types of big data, from genome sequences to medical images, there is no single ‘best’ way to process the data. This issue is exemplified by the substantial differences in how individual laboratories preprocess and analyse data from functional magnetic resonance imaging (fMRI) experiments, which generate information about brain activity. Indeed, a survey of fMRI studies found that nearly every study used a different analysis pipeline1. Writing in Nature, Botvinik-Nezer et al.2 provide further evidence of this variability, highlighting how analytical choices made by individual researchers can greatly influence the findings gleaned from an fMRI data set. The work is bound to spark lively discussion.