Multiple reproducibility indices for evaluation of cognitive functional MR imaging paradigms

Area of Science:

• Neuroimaging
• Cognitive Neuroscience

Background:

• Reproducibility of cognitive functional magnetic resonance imaging (fMRI) paradigms is crucial for reliable results.
• Existing indices may not fully capture activation patterns, including overlapping and extraneous regions.

Purpose of the Study:

• To evaluate various reproducibility indices for cognitive fMRI paradigms.
• To compare the reproducibility of word-generation and text-listening paradigms.

Main Methods:

• Eight volunteers underwent fMRI using word-generation and forward-backward text-listening paradigms, repeated within-session and after one week.
• Reproducibility indices were calculated based on the ratio of probability-weighted intersection volume to union volume of activation clusters.
• Indices included global, language area, extraneous activation reproducibility, and laterality.

Main Results:

• The word-generation paradigm demonstrated significantly higher global (0.75 vs 0.5) and language area (0.85 vs 0.6) reproducibility than the text-listening paradigm.
• Word-generation tasks showed larger, highly reproducible activations outside dominant language areas.
• Text-listening tasks exhibited more focal activations with less extraneous activation.

Conclusions:

• Multiple reproducibility indices are essential for selecting clinically relevant fMRI language paradigms.
• Word-generation tasks offer higher reproducibility at the cost of precise localization.
• Forward-backward text-listening paradigms provide more focal activation with reduced extraneous signals.