Abstract
Data analysis workflows in many scientific domains have become increasingly complex and flexible. Here we assess the effect of this flexibility on the results of functional magnetic resonance imaging by asking 70 independent teams to analyse the same dataset, testing the same 9 ex-ante hypotheses1. The flexibility of analytical approaches is exemplified by the fact that no two teams chose identical workflows to analyse the data. This flexibility resulted in sizeable variation in the results of hypothesis tests, even for teams whose statistical maps were highly correlated at intermediate stages of the analysis pipeline. Variation in reported results was related to several aspects of analysis methodology. Notably, a meta-analytical approach that aggregated information across teams yielded a significant consensus in activated regions. Furthermore, prediction markets of researchers in the field revealed an overestimation of the likelihood of significant findings, even by researchers with direct knowledge of the dataset2–5. Our findings show that analytical flexibility can have substantial effects on scientific conclusions, and identify factors that may be related to variability in the analysis of functional magnetic resonance imaging. The results emphasize the importance of validating and sharing complex analysis workflows, and demonstrate the need for performing and reporting multiple analyses of the same data. Potential approaches that could be used to mitigate issues related to analytical variability are discussed.
| Original language | English (US) |
|---|---|
| Pages (from-to) | 84-88 |
| Number of pages | 5 |
| Journal | Nature |
| Volume | 582 |
| Issue number | 7810 |
| DOIs | |
| State | Published - Jun 4 2020 |
Funding
Acknowledgements Neuroimaging data collection, performed at Tel Aviv University, was supported by the Austrian Science Fund (P29362-G27), the Israel Science Foundation (ISF 2004/15 to T. Schonberg) and the Swedish Foundation for Humanities and Social Sciences (NHS14-1719:1). Hosting of the data on OpenNeuro was supported by a National Institutes of Health (NIH) grant (R24MH117179). We thank M. C. Frank, Y. Assaf and N. Daw for comments on an earlier draft; the Texas Advanced Computing Center for providing computing resources for preprocessing of the data; the Stanford Research Computing Facility for hosting the data; and D. Roll for assisting with data processing. T. Schonberg thanks The Alfredo Federico Strauss Center for Computational Neuroimaging at Tel Aviv University; A.D. thanks the Knut and Alice Wallenberg Foundation and the Marianne and Marcus Wallenberg Foundation (A.D. is a Wallenberg Scholar), the Austrian Science Fund (FWF, SFB F63) and the Jan Wallander and Tom Hedelius Foundation (Svenska Handelsbankens Forskningsstiftelser); F. Holzmeister, J. Huber and M. Kirchler thank the Austrian Science Fund (FWF, SFB F63); D.W. was supported by the Research Foundation Flanders (FWO) and the European Union\u2019s Horizon 2020 research and innovation programme (https://ec.europa.eu/programmes/horizon2020/en) under the Marie Sk\u0142odowska-Curie grant agreement no. 665501; L. Tisdall was supported by the University of Basel Research Fund for Junior Researchers; C.B.C. was supported by grant 12O7719N from the Research Foundation Flanders; E.L. was supported by grant 12T2517N from the Research Foundation Flanders and Marie Sk\u0142odowska-Curie Actions under COFUND grant agreement 665501; A. Eed was supported by a predoctoral fellowship La Caixa-Severo Ochoa from Obra Social La Caixa and also acknowledges Comunidad de C\u00E1lculo Cient\u00EDfico del CSIC for the high-performance computing (HPC) use; C.L. was supported by the Vienna Science and Technology Fund (WWTF VRG13-007) and Austrian Science Fund (FWF P 32686); A.B.L.V. was supported by the Vienna Science and Technology Fund (WWTF VRG13-007); L.Z. was supported by the Vienna Science and Technology Fund (WWTF VRG13-007), the National Natural Science Foundation of China (no. 71801110), MOE (Ministry of Education in China) Project of Humanities and Social Sciences (no. 18YJC630268) and China Postdoctoral Science Foundation (no. 2018M633270); D.P. is currently supported by the Deutsche Forschungsgemeinschaft (DFG, German Research Foundation) under Germany\u2019s Excellence Strategy \u2018Science of Intelligence\u2019 (EXC 2002/1, project number 390523135); P.H. was supported in part by funding provided by Brain Canada, in partnership with Health Canada, for the Canadian Open Neuroscience Platform initiative; J.-B.P. was partially funded by the NIH (NIH-NIBIB P41 EB019936 (ReproNim), NIH-NIMH R01 MH083320 (CANDIShare) and NIH RF1 MH120021 (NIDM)) and the National Institute Of Mental Health of the NIH under award number R01MH096906 (Neurosynth), as well as the Canada First Research Excellence Fund, awarded to McGill University for the Healthy Brains for Healthy Lives initiative and the Brain Canada Foundation with support from Health Canada; S.B.E. was supported by the European Union\u2019s Horizon 2020 Research and Innovation Programme under grant agreement no. 785907 (HBP SGA2); G.M. was supported by the Max Planck Society; S. Heunis has received funding from the Dutch foundation LSH-TKI (grant LSHM16053-SGF); J.F.G.M. was supported by a Graduate Research Fellowship from the NSF and T32 Predoctoral Fellowship from the NIH; B.M. was supported by the Deutsche Forschungsgemeinschaft (grant CRC1193, subproject B01); A.R.L. was supported by NSF 1631325 and NIH R01 DA041353; M.E.H., T.J. and D.J.W. were supported by the Australian National Imaging Facility, a National Collaborative Research Infrastructure Strategy (NCRIS) capability; P.M.I. was supported by VIDI grant 452-17-013 from the Netherlands Organisation for Scientific Research; B.M.B. was supported by the Max Planck Society; J.P.H. was supported by a grant from the Swedish Research Council; R.W.C. and R.C.R. were supported by NIH IRP project number ZICMH002888; D.M.N., R.W.C., and R.C.R. used the computational resources of the National Institutes of Health High Performance Computing Biowulf cluster (http://hpc.nih.gov); D.M.N. was supported by NIH IRP project number ZICMH002960; C.F.C. was supported by the Tianqiao and Chrissy Center for Social and Decision Neuroscience Center Leadership Chair; R.G.B. was supported by the Max Planck Society; R.M.W.J.B. was supported by the Max Planck Society; M.B., O.C. and R.G. were supported by the Belgian Excellence of Science program (EOS project 30991544) from the FNRS-Belgium; O.C. is a research associate at the FRS-FNRS of Belgium; A.D.L. was supported by grant R4195 \u201CRepimpact\u201D of EraNET Neuron; Q.S. was funded by grant no. 71971199,71602175 and 71942004 from the National Natural Science Foundation of China and no. 16YJC630103 of the Ministry of Education of Humanities and Social Science; and T.E.N. was supported by the Wellcome Trust award 100309/Z/12/Z.
ASJC Scopus subject areas
- General