[MetaROR] Ch9 Field-Specific Challenges (MRI) — author revisions

[LukasWallrich]

Author revisions for Chapter 9 (Field-Specific Replication Challenges — MRI) from the MetaROR reviews (R2).

• Mention Jeanette Mumford's work, including the 2012 power-calculation guide and the "mumfordbrainstats" YouTube channel. (R2)
• Add at least one sentence on Eklund et al. (2016) cluster-failure paper and its field-wide impact. (R2)
• Add references to Durnez et al. (2016) on cluster-based power and Ellis et al. (2020) on simulation-based approaches. (R2)

Cross-cutting

• Ensure the chapter is clearly justified and linked to the other field specific chapters, so that it is clear why this narrow deep-dive is part of the book

[LukasWallrich]

Full extract from R2 comments

In this section, the authors provide an overview of the common problems researchers face when dealing with the replication of fMRI studies. Specifically (Section III.9."0".3), they attempt a quick review of problems in estimating power. While an extensive literature review falls outside the scope of the handbook, I believe there is room for improvement:

• Mention could be made of the extensive work by Jeanette Mumford, starting from the classic 2012 paper: A power calculation guide for fMRI studies. In a handbook that does not limit references to peer-reviewed papers, I think it would also be worth citing Mumford’s YouTube channel, "mumfordbrainstats," which serves as a quick but rigorous reference for practical solutions.
• I would spend at least one sentence citing the impact of Eklund et al. (2016) (Cluster failure: Why fMRI inferences for spatial extent have inflated false-positive rates) on the entire field (including implications for popular science, sensational titles, etc.).
• Further contributions worth including are
  • Durnez et al. (2016) (Power and sample size calculations for fMRI studies based on the prevalence of active peaks), which switches the focus from voxels to clusters, and
  • Ellis et al. (2020) (Facilitating open-science with realistic fMRI simulation: validation and application), regarding simulation-based approaches.
• I would also expand (by 1-2 sentences) the comment on Botvinik-Nezer et al. (2020), integrating the classic "70 teams, 70 different pipelines" observation, which is entirely accurate, by connecting it with the popularization of automated preprocessing pipelines (cited in 9."0".5).