A regularized, model-based approach to phase-based conductivity mapping using MRI.

Area of Science:

• Medical Imaging
• Biophysics
• Computational Science

Background:

• Phase-based conductivity mapping is crucial for various medical applications.
• Existing methods often struggle with accuracy, especially near boundaries and in noisy environments.

Purpose of the Study:

• To develop a novel regularized, model-based approach for phase-based conductivity mapping.
• To enhance the accuracy of conductivity maps by incorporating structural information.

Main Methods:

• Utilized the inverse of the 3D Laplacian operator to model phase-conductivity relationships.
• Employed a penalized weighted least-squares optimization with spatial masks derived from structural information.
• Compared the Inverse Laplacian method against a restricted Gaussian filter in simulations, phantoms, and human experiments.

Main Results:

• The Inverse Laplacian method demonstrated lower reconstruction bias and noise-induced error in simulations compared to the Gaussian filter.
• In phantom experiments, the Inverse Laplacian method yielded conductivity maps closer to measured values.
• Human brain imaging showed reduced noise in conductivity maps using the Inverse Laplacian method.

Conclusions:

• The Inverse Laplacian method provides more accurate conductivity maps with reduced noise, particularly near boundaries.
• This improved accuracy is vital for advancing the clinical and research applications of conductivity mapping.