Reducing Individual Variation for fMRI Studies in Children by Minimizing Template Related Errors

Area of Science:

• Neuroimaging
• Developmental Neuroscience

Background:

• Spatial normalization is crucial for group comparisons in functional MRI (fMRI) studies.
• Normalization errors can occur, especially in pediatric, elderly, or clinical populations, and in brain regions with high variability.
• Study-specific templates can minimize errors, but large sample sizes are often infeasible for pediatric research.

Purpose of the Study:

• To evaluate the performance of different templates in pediatric fMRI studies with limited sample sizes.
• To investigate methods for reducing individual variation and enhancing group comparison sensitivity in fMRI.
• To demonstrate training-related brain activation changes in children.

Main Methods:

• Compared three templates: study-specific, Chinese children's, and adult MNI, in a working memory task with 29 children.
• Analyzed differences in the right orbitofrontal cortex across templates.
• Introduced and tested a method incorporating the spatial normalization factor (SFN) as a covariate in group statistics.

Main Results:

• Study-specific and Chinese children's templates detected more between-group differences in the orbitofrontal cortex than the MNI template.
• A correlation was found between BOLD contrast size and SFN, indicating SFN's utility in characterizing individual variation.
• Including SFN as a covariate significantly improved detection power in group-level tests.
• Demonstrated enhanced right orbitofrontal activation correlated with behavioral improvements in the trained group.

Conclusions:

• Child-specific templates are more sensitive than adult templates for pediatric fMRI group comparisons.
• Incorporating SFN as a covariate is an effective strategy to reduce individual variation and boost statistical power in fMRI.
• The study successfully identified training-related neuroplasticity in the orbitofrontal cortex of children.