Coordinate based meta-analysis of functional neuroimaging data; false discovery control and diagnostics.

Area of Science:

• Neuroimaging
• Cognitive Neuroscience
• Statistical Analysis

Background:

• Coordinate based meta-analysis (CBMA) identifies consistent brain activation regions across fMRI studies.
• Current CBMA methods involve numerous statistical tests, increasing the risk of Type 1 errors.

Purpose of the Study:

• To reduce the number of statistical tests in CBMA.
• To introduce a novel statistical control method, the false cluster discovery rate (FCDR).
• To enhance the interpretability and accuracy of CBMA findings.

Main Methods:

• Developed a novel clustering scheme to reduce statistical tests.
• Introduced the false cluster discovery rate (FCDR) for statistical control.
• Implemented a data diagnostic scheme for quality assurance.

Main Results:

• The new method significantly reduces the number of statistical tests compared to voxel-wise approaches.
• False Cluster Discovery Rate (FCDR) offers interpretable control over Type 1 errors in CBMA.
• Numerical experiments demonstrate superior control of false clusters compared to the ALE method.
• The clustering scheme improves the reporting of task-relevant brain structures.

Conclusions:

• The proposed CBMA approach with FCDR provides a more accurate and interpretable method for analyzing neuroimaging data.
• This method enhances the reliability of findings in functional neuroimaging meta-analyses.
• The data diagnostic scheme ensures the quality and validity of the analysis.