Resting-state fMRI data reflects default network activity rather than null data: A defense of commonly employed...

Area of Science:

• Neuroimaging
• Statistical Analysis
• Brain Activity

Background:

• Functional magnetic resonance imaging (fMRI) analysis requires correcting for numerous statistical tests to control familywise error.
• A recent study by Eklund et al. questioned the validity of these correction methods using resting-state fMRI data.

Discussion:

• Eklund et al. assumed resting-state fMRI data represented null data, but it contained default network activity.
• This activity artificially inflated familywise error rates in their analysis.
• Their findings do not invalidate established multiple comparison correction methods in fMRI.

Key Insights:

• The 'null data' used by Eklund et al. was not truly null, containing inherent brain activity.
• Default mode network activity can significantly impact statistical results in fMRI.
• Current methods for correcting multiple comparisons in fMRI remain valid.

Outlook:

• Further research should ensure resting-state fMRI data used for null hypothesis testing is properly controlled.
• Replication of Eklund et al.'s analysis with appropriate null data is needed.
• This study reinforces the importance of rigorous methodology in neuroimaging data analysis.