Related Experiment Video
Updated: Mar 16, 2026
06:26
Meta-analysis of Voxel-Based Neuroimaging Studies using Seed-based d Mapping with Permutation of Subject Images SDM-PSI
Published on: November 27, 2019
78.7K
Improving effect size estimation and statistical power with multi-echo fMRI and its impact on understanding the
Michael V Lombardo1, Bonnie Auyeung2, Rosemary J Holt3
1Center for Applied Neuroscience, Department of Psychology, University of Cyprus, Cyprus; Autism Research Centre, Department of Psychiatry, University of Cambridge, UK.
Neuroimage
|July 16, 2016
Summary
Functional magnetic resonance imaging (fMRI) studies can be improved by removing non-BOLD artifacts using multi-echo ICA (ME-ICA). This enhances statistical power and effect sizes, particularly in brain regions like the cerebellum, enabling more robust findings.
Area of Science:
- Neuroimaging
- Cognitive Neuroscience
- Biophysics
Background:
- Functional magnetic resonance imaging (fMRI) research faces statistical power limitations due to small sample sizes and data artifacts.
- Artifacts in fMRI data can negatively impact effect size estimation and statistical power, especially in task-based studies.
- Improving statistical power is crucial for reliable fMRI findings and understanding brain function.
Purpose of the Study:
- To demonstrate how removing non-BOLD artifacts using ME-ICA can enhance effect size estimation and statistical power in task-fMRI.
- To investigate the application of ME-ICA in the domain of mentalizing/theory of mind tasks.
- To assess the impact of ME-ICA on effect sizes in both canonical and non-canonical brain regions.
Main Methods:
- Utilized multi-echo fMRI acquisition combined with independent components analysis (ME-ICA) for artifact identification and removal.
- Applied ME-ICA to task-fMRI data from two mentalizing tasks.
- Performed group-level analysis to evaluate effect size changes in specific brain regions, with and without smoothing.
Main Results:
- ME-ICA enhanced group-level effect size estimates by a median of 24% in canonical mentalizing regions without smoothing.
- Substantial effect size boosts (40-149%) were observed in non-canonical cerebellar areas after ME-ICA application.
- ME-ICA improved statistical power, enabling higher-powered studies at traditional sample sizes and revealing potentially unobservable cerebellar effects.
Conclusions:
- ME-ICA provides a principled, design-agnostic method for removing non-BOLD artifacts in task-fMRI.
- This artifact removal significantly improves effect size estimates and statistical power, particularly in cerebellar regions.
- ME-ICA can help mitigate statistical power issues in fMRI research and facilitate the discovery of under-appreciated aspects of brain organization.
