Quantifying and comparing region-of-interest activation patterns in functional brain MR imaging: methodology...

Area of Science:

• Neuroimaging
• Cognitive Neuroscience
• Biostatistics

Background:

• Functional brain magnetic resonance imaging (fMRI) is crucial for understanding brain activity during tasks.
• Accurate quantification of brain activation is essential for reliable fMRI study outcomes.
• Existing methods for evaluating fMRI analysis techniques have limitations in comparing activation quantification.

Purpose of the Study:

• To present a quantitative evaluation technique for functional imaging data analysis methods.
• To compare two distinct methods for quantifying localized brain activation: pixel count vs. sum of t-values.
• To examine the sensitivity of analysis of variance (ANOVA) results to user-defined parameters like threshold and region of interest (ROI) size.

Main Methods:

• Developed an ANOVA-based technique to evaluate activation quantification metrics across statistical thresholds.
• Compared the pixel count method with the sum of t-values method for quantifying brain activation.
• Utilized both simulated and real fMRI data to demonstrate the evaluation technique and parameter sensitivity.

Main Results:

• The ANOVA-based technique effectively reveals the sensitivity of different activation quantification metrics to noise and variables.
• The study provides a comparative analysis of the pixel count and sum of t-values approaches for activation quantification.
• Sensitivity analysis demonstrates the impact of user-defined parameters (threshold, ROI size) on fMRI analysis outcomes.

Conclusions:

• The proposed ANOVA-based technique offers a robust framework for quantitatively evaluating fMRI analysis methods.
• Understanding the performance differences between quantification methods and parameter sensitivity is vital for accurate fMRI research.
• This work contributes to improving the reliability and reproducibility of functional brain imaging studies.