Consensus inference in neuroimaging.

Area of Science:

• Neuroimaging
• Medical Image Analysis
• Statistical Modeling

Background:

• Neuroimaging studies often generate complex model summary images.
• Direct comparison and averaging of these images can be challenging due to variations.
• Enhancing the reliability and performance of neuroimaging analyses is crucial.

Purpose of the Study:

• To introduce and validate a novel model averaging technique for neuroimaging.
• To demonstrate the utility of histogram equalization for image comparison.
• To improve the performance of statistical models in neuroimaging analyses.

Main Methods:

• Development of a model averaging procedure for neuroimaging data.
• Application of histogram equalization to normalize image intensity distributions.
• Evaluation of the method using a simulation study and a functional MRI (fMRI) dataset.

Main Results:

• Model summary images can be directly compared and averaged after histogram equalization.
• The averaging procedure significantly enhances the receiver operating characteristic (ROC) curve in simulations.
• The method was successfully applied to an fMRI study of the motor cortex, showing improved performance.

Conclusions:

• Model averaging, combined with histogram equalization, is an effective technique for neuroimaging.
• This approach enhances the diagnostic performance of neuroimaging models.
• The method offers a valuable tool for analyzing fMRI data and other neuroimaging modalities.