A novel background field removal method for MRI using projection onto dipole fields (PDF)

Area of Science:

• Medical Imaging
• Biophysics
• Neuroimaging

Background:

• Accurate magnetic susceptibility quantification in imaging requires isolating local magnetic fields from background fields.
• Existing background removal techniques have limitations due to model assumptions and input data accuracy.
• Background fields in susceptibility-weighted imaging originate from imperfect shimming and tissue susceptibility variations.

Purpose of the Study:

• To develop a novel, non-parametric technique for background magnetic field removal in susceptibility-weighted imaging.
• To address the limitations of previous background removal methods.
• To improve image quality and quantitative accuracy in susceptibility mapping.

Main Methods:

• Proposed a projection onto dipole fields (PDF) technique based on the orthogonality of magnetic fields from internal and external dipoles.
• Decomposed the background field within a region of interest (ROI) into fields originating from external dipoles using Hilbert space projection.
• Validated the PDF technique using numerical simulations, phantom experiments, and human brain imaging.

Main Results:

• The PDF technique demonstrated substantial improvement in background field removal compared to the high-pass filtering method.
• Successfully applied the PDF method to human brain imaging, enhancing image quality.
• Validated the effectiveness and robustness of the PDF approach.

Conclusions:

• The projection onto dipole fields (PDF) technique is a powerful and effective method for background magnetic field removal.
• PDF offers superior performance over traditional methods for susceptibility-weighted imaging.
• This technique holds significant potential for advancing quantitative susceptibility mapping and neuroimaging research.