Statistical testing and power analysis for brain-wide association study.

Area of Science:

• Neuroimaging
• Statistical Genetics
• Computational Neuroscience

Background:

• Connexel-wise analysis in brain-wide association studies (BWAS) identifies functional connectivities between voxels.
• Existing methods for connexel-wise analysis lack robust multiple correction and power analysis, hindering research in mental disorders like schizophrenia, autism, and depression.

Purpose of the Study:

• To develop a rigorous statistical framework for connexel-wise significance testing using Gaussian random field theory.
• To enable accurate control of the family-wise error rate (FWER) and perform power/sample size calculations for connexel-wise studies.
• To provide an efficient alternative to non-parametric permutation for multiple comparison correction in large fMRI datasets.

Main Methods:

• Developed a statistical framework based on Gaussian random field theory for connexel-wise significance testing.
• Incorporated topological inference methods for FWER control.
• Validated the framework using resting-state fMRI datasets and compared it with Bonferroni correction and FDR.

Main Results:

• The framework accurately controls the false-positive rate.
• It reduces false-positive rates and increases statistical power compared to Bonferroni correction and FDR by leveraging spatial information.
• The method efficiently handles large, high-resolution fMRI datasets without non-parametric permutation.
• Successfully identified altered functional connectivities in a major depressive disorder dataset where existing methods failed.

Conclusions:

• The developed statistical framework provides a powerful and efficient tool for connexel-wise association studies.
• It offers improved accuracy and statistical power for detecting functional connectivity changes in mental disorders.
• The available software package facilitates the application of this novel methodology in neuroimaging research.