Improved total sensitivity estimation for multiple receive coils in MRI using ratios of first-order statistics

Area of Science:

• Magnetic Resonance Imaging (MRI)
• Medical Physics
• Image Processing

Background:

• Spatial variation in MRI receive coil sensitivity profiles causes signal amplitude scaling issues across images.
• Current corrections involve calibration, fitting, or image/profile constraints, while summation strategies prioritize SNR or encoding, not unscaled signal estimation.
• Existing methods often fail to accurately estimate the true signal amplitude, unscaled by coil sensitivity.

Purpose of the Study:

• To present novel methods for correcting spatial variations in MRI coil sensitivity profiles.
• To estimate the unscaled signal amplitude at any position within an MRI.
• To improve image uniformity in MRI data acquired with coil arrays.

Main Methods:

• Utilized ratios of first-order statistics to approximate the unscaled signal value.
• Proposed two approaches for scaling the mean signal: using the mode of normalized signals and a derived expression based on mean and mean-of-squares.
• Validated methods through simulations with idealized coil arrays and real data from an 8-channel coil array on a uniform 13C phantom and Hyperpolarized 13C pyruvate brain MRI.

Main Results:

• Demonstrated improved image uniformity compared to a standard homomorphic filter.
• The proposed ratio-based statistical methods effectively correct for spatial sensitivity variations.
• Acknowledged that these novel approaches exhibit higher sensitivity to noise.

Conclusions:

• Introduced simple, effective methods for correcting spatial sensitivity variations in MRI coil arrays.
• These techniques offer a valuable initial or adjunct step in MRI data post-processing.
• The presented methods contribute to more accurate signal amplitude estimation and improved image quality.