Sharing brain mapping statistical results with the neuroimaging data model

Maumet, Camille; Auer, Tibor; Bowring, Alexander; Chen, Gang; Das, Samir; Flandin, Guillaume; Ghosh, Samik; Glatard, Tristan; Gorgolewski, Krzysztof J.; Helmer, Karl G.; Jenkinson, Mark; Keator, David B.; Nichols, B. Nolan; Poline, Jean‐Baptiste; Reynolds, Richard C.; Sochat, Vanessa; Turner, Jessica A.; Nichols, Thomas E.

doi:10.1038/sdata.2016.102

Cited by 50 publications

(31 citation statements)

References 39 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Once analyses were complete, the results for each software package were exported as NIDM‐Results packs (FSL and SPM only, [Maumet et al, ]) and uploaded to a public collection on the NeuroVault (, http://neurovault.org; [Gorgolewski et al, ]) online data repository.…”

Section: Methodsmentioning

confidence: 99%

Exploring the impact of analysis software on task fMRI results

Bowring

Maumet

Nichols

2019

Human Brain Mapping

Self Cite

126

104

View full text Add to dashboard Cite

A wealth of analysis tools are available to fMRI researchers in order to extract patterns of task variation and, ultimately, understand cognitive function. However, this “methodological plurality” comes with a drawback. While conceptually similar, two different analysis pipelines applied on the same dataset may not produce the same scientific results. Differences in methods, implementations across software, and even operating systems or software versions all contribute to this variability. Consequently, attention in the field has recently been directed to reproducibility and data sharing. In this work, our goal is to understand how choice of software package impacts on analysis results. We use publicly shared data from three published task fMRI neuroimaging studies, reanalyzing each study using the three main neuroimaging software packages, AFNI, FSL, and SPM, using parametric and nonparametric inference. We obtain all information on how to process, analyse, and model each dataset from the publications. We make quantitative and qualitative comparisons between our replications to gauge the scale of variability in our results and assess the fundamental differences between each software package. Qualitatively we find similarities between packages, backed up by Neurosynth association analyses that correlate similar words and phrases to all three software package's unthresholded results for each of the studies we reanalyse. However, we also discover marked differences, such as Dice similarity coefficients ranging from 0.000 to 0.684 in comparisons of thresholded statistic maps between software. We discuss the challenges involved in trying to reanalyse the published studies, and highlight our efforts to make this research reproducible.

show abstract

Section: Methodsmentioning

confidence: 99%

Exploring the impact of analysis software on task fMRI results

Bowring

Maumet

Nichols

2019

Human Brain Mapping

Self Cite

126

104

View full text Add to dashboard Cite

show abstract

“…LORIS is committed to the standardization of ontologies, and currently adopts a practical approach where (1) all the (DICOM) fields related to imaging data are preserved and made queryable, and (2) terms used for behavioral variables and biobanking studies are defined on a study-by-study basis, while their re-utilization is also promoted across studies, compliant (where possible) with conventions such as BIDS (Gorgolewski et al, 2016) or NDAR (Hall et al, 2012). Prospectively, LORIS plans to adopt ontologies under development by the NIDM initiative to formally and uniformly describe raw data, terms, workflows and derived data (Maumet et al, 2016), as well as open data citation standards such as those developed for neuroimaging (Honor et al, 2016). Further integration of domain-specific standards, such as MIABIS 2.0 developed for biobanking data by the BBMRI-ERIC network (Merino-Martinez et al, 2016), is a priority for integration of data dissemination formats for the Open Science platform.…”

Section: Methodsmentioning

confidence: 99%

“…At the same time, emerging definitions of common data sharing standards, practices, and formats are being established via BIDS (Gorgolewski et al, 2016), the Neuro-Imaging Data Model (NIDM) (Maumet et al, 2016), FAIR principles (Wilkinson et al, 2016) and even extending to data organization and citation strategies (Honor et al, 2016). Meanwhile, governments and funding agencies in the USA (National Institutes of Health, 2014; National Institute of Mental Health, 2015), Canada (Tri-Agency Statement of Principles of Digital Data Management, 2016), Europe (Horizon 2020, The Wellcome Trust, 2016) and elsewhere encourage and increasingly require research programs to establish data management and sharing plans from the start of the research data lifecycle.…”

Section: Introductionmentioning

confidence: 99%

Cyberinfrastructure for Open Science at the Montreal Neurological Institute

Das

Glatard

Rogers

et al. 2017

Front. Neuroinform.

Self Cite

View full text Add to dashboard Cite

Data sharing is becoming more of a requirement as technologies mature and as global research and communications diversify. As a result, researchers are looking for practical solutions, not only to enhance scientific collaborations, but also to acquire larger amounts of data, and to access specialized datasets. In many cases, the realities of data acquisition present a significant burden, therefore gaining access to public datasets allows for more robust analyses and broadly enriched data exploration. To answer this demand, the Montreal Neurological Institute has announced its commitment to Open Science, harnessing the power of making both clinical and research data available to the world (Owens, 2016a,b). As such, the LORIS and CBRAIN (Das et al., 2016) platforms have been tasked with the technical challenges specific to the institutional-level implementation of open data sharing, including: Comprehensive linking of multimodal data (phenotypic, clinical, neuroimaging, biobanking, and genomics, etc.)Secure database encryption, specifically designed for institutional and multi-project data sharing, ensuring subject confidentiality (using multi-tiered identifiers).Querying capabilities with multiple levels of single study and institutional permissions, allowing public data sharing for all consented and de-identified subject data.Configurable pipelines and flags to facilitate acquisition and analysis, as well as access to High Performance Computing clusters for rapid data processing and sharing of software tools.Robust Workflows and Quality Control mechanisms ensuring transparency and consistency in best practices.Long term storage (and web access) of data, reducing loss of institutional data assets.Enhanced web-based visualization of imaging, genomic, and phenotypic data, allowing for real-time viewing and manipulation of data from anywhere in the world.Numerous modules for data filtering, summary statistics, and personalized and configurable dashboards.Implementing the vision of Open Science at the Montreal Neurological Institute will be a concerted undertaking that seeks to facilitate data sharing for the global research community. Our goal is to utilize the years of experience in multi-site collaborative research infrastructure to implement the technical requirements to achieve this level of public data sharing in a practical yet robust manner, in support of accelerating scientific discovery.

show abstract

“…While these tools focus on planning an experiment, there is also an increased effort to standardize the organization and storage of 1 http://neuropowertools.org obtained fMRI data. Examples are the Brain Imaging Data Structure (Gorgolewski et al, 2016) or the Neuroimaging Data Model which facilitates representation of meta-data. Combined, these protocols enable easy sharing of data between research groups.…”

Section: Introductionmentioning

confidence: 99%

Standardized Effect Sizes and Image-Based Meta-Analytical Approaches for fMRI Data

Bossier

Nichols

Moerkerke

2019

Preprint

Self Cite

View full text Add to dashboard Cite

Scientific progress is based on the ability to compare opposing theories and thereby develop consensus among existing hypotheses or create new ones. We argue that data aggregation (i.e. combine data across studies or research groups) for neuroscience is an important tool in this process.An important prerequisite is the ability to directly compare fMRI results over studies. In this paper, we discuss how an observed effect size in an fMRI dataanalysis can be transformed into a standardized effect size. We demonstrate how these enable direct comparison and data aggregation over studies. Furthermore, we also discuss the influence of key parameters in the design of an fMRI experiment (such as number of scans and the sample size) on (statistical)properties of standardized effect sizes. In the second part of the paper, we give 1 an overview of two approaches to aggregate fMRI results over studies. The first corresponds to extending the two-level general linear model approach as is typically used in individual fMRI studies with a third level. This requires the parameter estimates corresponding to the group models from each study together with estimated variances and meta-data. Unfortunately, there is a risk of running into unit mismatches when the primary studies use different scales to measure the BOLD response. To circumvent, it is possible to aggregate (unitless) standardized effect sizes which can be derived from summary statistics. We discuss a general model to aggregate these and different approaches to deal with between-study heterogeneity. Furthermore, we hope to further promote the usage of standardized effect sizes in fMRI research.

show abstract

Sharing brain mapping statistical results with the neuroimaging data model

Cited by 50 publications

References 39 publications

Exploring the impact of analysis software on task fMRI results

Exploring the impact of analysis software on task fMRI results

Cyberinfrastructure for Open Science at the Montreal Neurological Institute

Standardized Effect Sizes and Image-Based Meta-Analytical Approaches for fMRI Data

Contact Info

Product

Resources

About