Reconstruction strategy for echo planar spectroscopy and its application to partially undersampled imaging.

Area of Science:

• Magnetic Resonance Imaging
• Spectroscopy
• Signal Processing

Background:

• Echo planar spectroscopy commonly uses oscillating gradients during readout, leading to non-Cartesian data sampling.
• Reconstructing this non-Cartesian data efficiently is crucial for advanced magnetic resonance applications.

Purpose of the Study:

• To present a fast gridding algorithm for echo planar spectroscopy data acquired with oscillating gradients.
• To demonstrate the algorithm's effectiveness in reconstructing partially undersampled data.

Main Methods:

• Development of an optimal gridding algorithm for non-Cartesian echo planar spectroscopy data.
• Application of the Fast Fourier Transform (FFT) for efficient signal reconstruction.
• Demonstration using data partially undersampled in the time domain.

Main Results:

• The presented gridding method achieves optimal performance, comparable to the full discrete Fourier transform for band-limited signals.
• Reconstruction of partially undersampled data is demonstrated to be feasible and effective.
• Artifacts from undersampling were found to be acceptable for spectroscopy with long echo times.

Conclusions:

• The developed fast gridding algorithm offers an efficient solution for echo planar spectroscopy data reconstruction.
• This method supports undersampling, enabling lower hardware requirements or faster acquisitions.
• The technique is particularly beneficial for high-field scanning and systems with limited gradient performance.