Estimation for the bivariate quantile varying coefficient model with application to diffusion tensor imaging data...

Area of Science:

• Neuroimaging
• Biomedical Data Analysis
• Statistical Modeling

Background:

• Joint modeling of multiple functional responses, like diffusion properties in neuroimaging, lacks robust statistical methods.
• Existing methods struggle to capture the complex relationships within biomedical data.

Purpose of the Study:

• To propose a novel varying coefficient quantile regression model for bivariate functional responses.
• To enable joint distribution modeling of functional variables across clinical covariates for enhanced biomedical insights.

Main Methods:

• Utilized a varying coefficient quantile regression model for bivariate functional responses.
• Employed alternating direction method of multipliers and propagation separation algorithms for estimation.
• Incorporated B-spline basis and L2 smoothness penalty for interpretability.

Main Results:

• The proposed model effectively handles joint analysis of functional fractional anisotropy and mean diffusivity.
• Demonstrated performance through simulation studies and application to neurodevelopmental data.
• Provided insights into the association of diffusion properties with gestational age and sex.

Conclusions:

• The developed statistical method offers a robust approach for joint modeling of diffusion MRI data.
• This technique enhances understanding of neurodevelopmental trajectories by analyzing multiple functional responses simultaneously.
• The model provides interpretable results, facilitating innovative biomedical research.