Eliminating electromagnetic interference for RF shielding-free MRI via k-space convolution: Insights from MR parallel...

Area of Science:

• Magnetic Resonance Imaging (MRI)
• Biomedical Engineering
• Signal Processing

Background:

• Ultra-low field MRI offers potential for democratizing MRI applications.
• Active sensing and elimination of electromagnetic interference (EMI) are key to shielding-free imaging in ULF MRI.

Purpose of the Study:

• To review analytical approaches for EMI estimation and elimination in ULF MRI.
• To investigate the theoretical basis and relationship of EMI methods with parallel imaging reconstruction.
• To propose an adaptive method for EMI estimation using tailored convolutional kernels.

Main Methods:

• Review of analytical EMI estimation/elimination techniques.
• Formulation of EMI estimation as k-space convolution or spectrum-space multiplication.
• Development and application of tailored convolutional kernels for adaptive EMI fitting.
• Evaluation using simulation studies and human brain imaging.

Main Results:

• Existing EMI approaches are analyzed and related to parallel imaging.
• A novel adaptive EMI estimation method using tailored convolutional kernels is proposed.
• The proposed method demonstrates robust performance against system and acquisition imperfections in simulations and human brain imaging.

Conclusions:

• The study provides a deeper understanding of EMI estimation and elimination in ULF MRI.
• Tailored convolutional kernels offer an adaptive and robust solution for EMI compensation.
• This work contributes to the advancement of shielding-free ULF MRI technology.