Related Experiment Video
Updated: Sep 27, 2025
14:27
Identification of Disease-related Spatial Covariance Patterns using Neuroimaging Data
Published on: June 26, 2013
15.8K
CLEAN: Leveraging spatial autocorrelation in neuroimaging data in clusterwise inference.
1Department of Statistical Sciences and Department of Psychology, University of Toronto, Toronto, ON M5S, Canada.
Neuroimage
|April 10, 2022
Summary
We introduce CLEAN, a new statistical method for neuroimaging that accounts for spatial autocorrelation to improve reproducibility. This approach enhances sensitivity in clusterwise inference for brain imaging analysis.
Area of Science:
- Neuroimaging
- Statistical analysis
- Brain imaging
Background:
- Clusterwise inference is popular in neuroimaging for improved sensitivity.
- Current methods fail to account for spatial autocorrelation, potentially reducing reproducibility.
- High-resolution MRI data exhibit significant spatial autocorrelation.
Purpose of the Study:
- To propose a novel, powerful, and fast statistical method for neuroimaging.
- To address limitations in current clusterwise inference methods regarding spatial autocorrelation.
- To enhance the reproducibility of neuroimaging studies.
Main Methods:
- Developed CLEAN (Clusterwise inference Leveraging spatial Autocorrelations in Neuroimaging).
- CLEAN computes multivariate test statistics modeling brain-wise spatial autocorrelations.
- Employs a refitting-free resampling approach for false positive control.
Main Results:
- CLEAN was validated using simulations and real data from the Human Connectome Project.
- The method effectively models spatial autocorrelations in neuroimaging data.
- Demonstrated improved sensitivity and reproducibility in clusterwise inference.
Conclusions:
- CLEAN offers a significant advancement in neuroimaging statistical analysis.
- The method is well-suited for analyzing high-resolution MRI data with spatial autocorrelation.
- Provides a new direction for robust and reproducible neuroimaging research.
