Systems and methods for magnetic resonance fingerprinting with increased volumetric coverage using deep learning reconstruction

Abstract

Presented herein are systems and methods for magnetic resonance fingerprinting with increased volumetric coverage using deep learning reconstruction. Fingerprinting techniques enable obtaining quantitative molecular information using a scanner in a non-invasive way. However, existing techniques for slice acquisition result in limited spatial coverage. In a single size approach, the pulse sequence can include a full saturation train. The pulse saturation train can excite a single slice. After a delay to allow the magnetization to recover, the pulse saturation train excites the magnetization again with another saturation pulse with different saturation powers. The magnetization is repeated that over a certain length of acquisitions. The magnetization can generate the fingerprint from which to extract the underlying quantitative parameters. Instead of exciting a single slice and allowing it to recover, the present disclosure can excite multiple slices to increase the spatial coverage without increasing the scan time.

Dependencies are inferred from the claim text.