SPIRiT: Iterative self-consistent parallel imaging reconstruction from arbitrary k-space.

Area of Science:

• Magnetic Resonance Imaging (MRI)
• Image Reconstruction
• Medical Imaging Technology

Background:

• Parallel imaging reconstruction methods require accurate coil sensitivity maps.
• Existing autocalibration techniques can be sensitive to undersampling and noise.
• A generalized framework is needed for robust coil-by-coil reconstruction.

Purpose of the Study:

• To present a generalized, self-consistency-based framework for autocalibrating parallel imaging reconstruction.
• To develop a reconstruction approach that is robust to arbitrary k-space sampling patterns.
• To incorporate advanced image priors, such as off-resonance correction and compressed sensing regularization.

Main Methods:

• Formulation of the reconstruction problem as an optimization problem based on self-consistency.
• Iterative solution strategies in both image and k-space domains using projection over convex sets and conjugate gradient algorithms.
• Flexible incorporation of image priors including off-resonance correction and nonlinear l(1)-wavelet regularization.

Main Results:

• Demonstration of efficient reconstructions from undersampled Cartesian and spiral trajectories using phantom and in vivo data.
• Successful reconstruction of images with incorporated off-resonance correction.
• Validation of nonlinear l(1)-wavelet regularization for improved reconstruction quality.

Conclusions:

• The proposed self-consistency framework provides a robust and generalized approach to autocalibrating parallel imaging.
• The method accurately reconstructs images from undersampled data and flexibly integrates advanced priors.
• This technique has the potential to improve the efficiency and quality of MRI scans.