Behavior, sensitivity, and power of activation likelihood estimation characterized by massive empirical simulation

Area of Science:

• Neuroimaging
• Cognitive Neuroscience
• Data Science

Background:

• Quantitative meta-analyses are crucial for extracting knowledge on brain-behavior associations from growing neuroimaging data.
• Activation Likelihood Estimation (ALE) is a widely adopted method for coordinate-based neuroimaging meta-analyses.
• Determining optimal statistical inference and sample size is vital for robust ALE meta-analyses.

Purpose of the Study:

• To identify the most appropriate statistical thresholding method for ALE meta-analyses.
• To determine the minimum number of experiments required for reliable ALE meta-analyses.
• To provide evidence-based recommendations for future ALE study designs.

Main Methods:

• Simulated over 120,000 ALE meta-analysis datasets using empirical parameters from the BrainMap database.
• Varied parameters included number of subjects, reported foci, and distribution of activation foci.
• Analyzed the behavior of ALE analyses under different conditions to estimate statistical power.

Main Results:

• Cluster-level family-wise error (FWE) correction was identified as the most appropriate method for statistical inference in ALE.
• Voxel-level FWE correction is valid but more conservative; uncorrected and false-discovery rate methods should be avoided.
• A minimum of 20 experiments is recommended for ALE meta-analyses to achieve sufficient power for detecting moderate effects.

Conclusions:

• Cluster-level FWE correction is the preferred method for statistical inference in ALE meta-analyses.
• Researchers should include at least 20 experiments in ALE studies for adequate statistical power.
• These recommendations are specific to the ALE method and may not apply to other meta-analysis approaches.