Ultrafast (milliseconds), multidimensional RF pulse design with deep learning

  title={Ultrafast (milliseconds), multidimensional RF pulse design with deep learning},
  author={Mads Sloth Vinding and Birk Skyum and Ryan Sangill and Torben Ellegaard Lund},
  journal={Magnetic Resonance in Medicine},
  pages={586 - 599}
Some advanced RF pulses, like multidimensional RF pulses, are often long and require substantial computation time because of a number of constraints and requirements, sometimes hampering clinical use. However, the pulses offer opportunities of reduced‐FOV imaging, regional flip‐angle homogenization, and localized spectroscopy, e.g., of hyperpolarized metabolites. Proposed herein is a novel deep learning approach to ultrafast design of multidimensional RF pulses with intention of real‐time pulse… 

DeepControl: 2DRF pulses facilitating B1+ inhomogeneity and B0 off‐resonance compensation in vivo at 7 T

Rapid 2DRF pulse design with subject‐specific B1+ inhomogeneity and B0 off‐resonance compensation at 7 T predicted from convolutional neural networks is presented.

Multi‐task convolutional neural network‐based design of radio frequency pulse and the accompanying gradients for magnetic resonance imaging

A convolutional neural network with multi‐task learning, trained with SLR transformation pairs, that is capable of simultaneously generating RF and two channels of gradient waveforms, given the desired spatially two‐dimensional excitation profiles is presented.

Deep reinforcement learning-designed radiofrequency waveform in MRI

An artificial intelligence (AI)powered RF pulse design framework, DeepRF, is proposed, which utilizes the self-learning characteristics of deep reinforcement learning to generate a novel RF pulse.

Deep Reinforcement Learning Designed Shinnar-Le Roux RF Pulse Using Root-Flipping: DeepRFSLR

A novel approach of applying deep reinforcement learning to an RF pulse design that suggests a new way of designing an RF by applying a machine learning algorithm, demonstrating a “machine-designed” MRI sequence.

Deep Reinforcement Learning Designed Shinnar-Le Roux RF Pulse using Root-Flipping : DeepRF SLR

A novel approach of applying deep reinforcement learning to an RF pulse design that suggests a new way of designing an RF by applying a machine learning algorithm, demonstrating a “machine-designed” MRI sequence.

DeepControl: 2D RF pulses facilitating B1+ inhomogeneity and B0 off-resonance compensation in vivo at 7T

The reconstructed image data agrees well with the simulations using the acquired $B_1^+$ and $B-0$ maps and the 2D RF pulse predicted by the convolutional neural networks is as good as the conventional RF pulse obtained by optimal control.

Three‐dimensional static and dynamic parallel transmission of the human heart at 7 T

It is demonstrated in vivo that kT‐point pTx pulses are highly suitable for mitigating nominal FA heterogeneities across the entire 3D heart volume at 7 T and demonstrate a practical tradeoff between nominal FA heterogeneity mitigation and RF power.

Development, validation, and pilot MRI safety study of a high-resolution, open source, whole body pediatric numerical simulation model.

A 29-month-old male whole-body native numerical model, "MARTIN", that includes 28 head and 86 body tissue compartments, segmented directly from the high spatial resolution MRI and CT images is introduced.



Application of the limited-memory quasi-Newton algorithm for multi-dimensional, large flip-angle RF pulses at 7T

An optimal-control algorithm is proposed as a tool for designing advanced multi-dimensional, large flip-angle, radio-frequency pulses, thus enabling reduced field-of-view imaging using pulse trajectory acceleration factors up to 4.5 GHz.

Deep learning with domain adaptation for accelerated projection‐reconstruction MR

A novel deep learning approach with domain adaptation is proposed to restore high‐resolution MR images from under‐sampled k‐space data to solve the problem of streaking artifact patterns in magnetic resonance imaging.

Fast numerical design of spatial-selective rf pulses in MRI using Krotov and quasi-Newton based optimal control methods.

A Krotov-based optimal control approach is presented which as compared to earlier approaches provides very fast, monotonic convergence even without educated initial guesses, essential for in vivo MRI applications.

Designing multichannel, multidimensional, arbitrary flip angle RF pulses using an optimal control approach

The multidimensional multichannel RF pulse design is formulating as an optimal control problem with multiple controls based directly on the Bloch equation to improve the spatial localization or homogeneity of the magnetization profiles.

Spatial localization combining projection presaturation with a two‐dimensional excitation pulse

The improvement in outer volume suppression achieved by combining these two techniques was demonstrated by images collected after the application of projection presaturation alone, a 2D pulse alone, and the two combined.

A convolutional neural network to filter artifacts in spectroscopic MRI

Proton MRSI is a noninvasive modality capable of generating volumetric maps of in vivo tissue metabolism without the need for ionizing radiation or injected contrast agent. Magnetic resonance

Real-time 2D spatially selective MRI experiments: Comparative analysis of optimal control design methods.

Variable-rate selective excitation

Two‐dimensional spatially‐selective RF excitation pulses in echo‐planar imaging

Two‐dimensional spatially‐selective RF excitation pulses were developed for single‐shot echo‐planar imaging (EPI) with reduced field of view (FOV) in the phase‐encoding direction and yielded an echo train length of only 21 ms, as opposed to 102 ms for conventional EPI.

Pulseq: A rapid and hardware‐independent pulse sequence prototyping framework

This work presents an alternative programming environment that is hardware‐independent, open‐source, and promotes rapid sequence prototyping.