A Bayesian Spatial Model to Predict Disease Status Using Imaging Data From Various Modalities

Area of Science:

• Neuroimaging
• Computational Neuroscience
• Biostatistics

Background:

• Neurological and psychiatric disorders exhibit complex brain alterations.
• Integrating functional and structural neuroimaging data can enhance clinical relevance.
• Predicting disease status from imaging requires sophisticated analytical models.

Purpose of the Study:

• To develop and validate a Bayesian hierarchical model for predicting disease status using multimodal brain imaging data.
• To identify specific brain regions associated with disease status through voxel-level prediction.
• To improve the clinical significance of neuroimaging studies by accurately relating imaging data to disease.

Main Methods:

• A two-stage whole-brain parcellation into 282 subregions was employed.
• A Bayesian hierarchical model was utilized, accounting for correlations between brain regions.
• Markov Chain Monte Carlo (MCMC) methods and importance sampling were used for parameter estimation and computation reduction.
• Leave-one-out cross-validation was performed to assess prediction accuracy.

Main Results:

• The proposed model demonstrated high prediction accuracy for disease status.
• Voxel-level prediction successfully identified key brain regions associated with the disease.
• Significant regions included the caudate, putamen, fusiform gyrus, and sensory system areas.
• The model was successfully applied to multimodal brain imaging data from Parkinson's disease patients.

Conclusions:

• The Bayesian hierarchical model is effective for predicting disease status from multimodal neuroimaging data.
• The model provides insights into the specific neuroanatomical correlates of neurological disorders.
• This approach enhances the clinical utility of neuroimaging in diagnosing and understanding diseases like Parkinson's.