Meta-analysis of functional neuroimaging data: current and future directions

Area of Science:

• Neuroimaging
• Cognitive Neuroscience
• Biostatistics

Background:

• Meta-analysis is crucial for synthesizing neuroimaging research, addressing limitations like inadequate multiple comparison correction.
• Published neuroimaging studies may contain 10-20% false positive activations due to statistical thresholds.
• Meta-analysis aids in establishing consensus on functional brain region locations and developing structure-function hypotheses.

Purpose of the Study:

• To summarize popular meta-analytic approaches and their limitations in neuroimaging.
• To introduce a revised multilevel approach for improved cross-study consistency.
• To explore multivariate methods for testing brain region co-activity hypotheses and predicting psychological states.

Main Methods:

• Review and critique of existing meta-analytic techniques in neuroimaging.
• Outline of a novel, revised multilevel meta-analytic approach.
• Discussion of multivariate statistical methods for neuroimaging data.

Main Results:

• Identification of limitations in current meta-analytic methods.
• Proposal of an enhanced multilevel approach for greater validity in neuroimaging.
• Exploration of multivariate techniques for hypothesis testing and prediction.

Conclusions:

• Meta-analysis is a valuable tool for neuroimaging, enhancing reliability and hypothesis generation.
• The proposed multilevel approach offers increased validity for cross-study consistency.
• Meta-analyses can significantly contribute to predicting psychological states from brain activity patterns.