Accelerated MRI protocols routinely involve a predefined sampling pattern that undersamples the k-space. Finding an optimal pattern can enhance the reconstruction quality, however this optimization is a challenging task. To address this challenge, we introduce a novel deep learning framework, AutoSamp, based on variational information maximization that enables joint optimization of sampling pattern and reconstruction of MRI scans. We represent the encoder as a non-uniform Fast Fourier Transform that allows continuous optimization of k-space sample locations on a non-Cartesian plane, and the decoder as a deep reconstruction network. Experiments on public 3D acquired MRI datasets show improved reconstruction quality of the proposed AutoSamp method over the prevailing variable density and variable density Poisson disc sampling for both compressed sensing and deep learning reconstructions. We demonstrate that our data-driven sampling optimization method achieves 4.4dB, 2.0dB, 0.75dB, 0.7dB PSNR improvements over reconstruction with Poisson Disc masks for acceleration factors of R =5, 10, 15, 25, respectively. Prospectively accelerated acquisitions with 3D FSE sequences using our optimized sampling patterns exhibit improved image quality and sharpness. Furthermore, we analyze the characteristics of the learned sampling patterns with respect to changes in acceleration factor, measurement noise, underlying anatomy, and coil sensitivities. We show that all these factors contribute to the optimization result by affecting the sampling density, k-space coverage and point spread functions of the learned sampling patterns.

The Shinnar-Le-Roux (SLR) algorithm is widely used to design frequency selective pulses with large flip angles. We improve its design process to generate pulses with lower energy (by as much as 26%) and more accurate phase profiles. Concretely, the SLR algorithm consists of two steps: (1) an invertible transform between frequency selective pulses and polynomial pairs that represent Cayley-Klein (CK) parameters and (2) the design of the CK polynomial pair to match the desired magnetization profiles. Because the CK polynomial pair is bi-linearly coupled, the original algorithm sequentially solves for each polynomial instead of jointly. This results in sub-optimal pulses. Instead, we leverage a convex relaxation technique, commonly used for low rank matrix recovery, to address the bi-linearity. Our numerical experiments show that the resulting pulses are almost always globally optimal in practice. For slice excitation, the proposed algorithm results in more accurate linear phase profiles. And in general the improved pulses have lower energy than the original SLR pulses.

To enable wireless MRI receive arrays, per-channel power consumption must be reduced by a significant factor. To address this, a low-power SiGe alternative to industry standard MRI pre-amplifier blocks has been proposed and its impact on imaging performance evaluated in a benchtop environment. The SiGe amplifier reduces power consumption 28x, but exhibits increased non-linearity and reduced dynamic range relative to industry standard amplifiers. This distorts the images, causing reduced contrast and a blurring of fine features. In conjunction with the amplifier, a semi-blind calibration and compensation framework has been proposed to remove artifacts caused by this non-linearity. Requiring the knowledge of the calibration signal bandwidth, the associated peak transmit powers, and the distorted baseband signals, a second non-linearity is constructed that when cascaded with the receive chain produces a linear response. This method was evaluated for both knee and phantom image datasets of peak input power -20dBm with a -40dBm peak input power image as reference. In the benchtop environment, industry standard amplifiers produced input normalized RMSEs of 0.0199 and 0.0310 for phantom and knee datasets, respectively. The low-power SiGe amplifier resulted in RMSEs of 0.0869 and 0.1130 which were reduced to 0.0158 and 0.0168 following compensation, for phantom and knee images respectively. The ability to effectively compensate for this reduced dynamic range encourages further investigation of low-power SiGe amplifiers for power limited MRI receive arrays.

Reliable MRI is crucial for accurate interpretation in therapeutic and diagnostic tasks. However, undersampling during MRI acquisition as well as the overparameterized and non-transparent nature of deep learning (DL) leaves substantial uncertainty about the accuracy of DL reconstruction. With this in mind, this study aims to quantify the uncertainty in image recovery with DL models. To this end, we first leverage variational autoencoders (VAEs) to develop a probabilistic reconstruction scheme that maps out (low-quality) short scans with aliasing artifacts to the diagnostic-quality ones. The VAE encodes the acquisition uncertainty in a latent code and naturally offers a posterior of the image from which one can generate pixel variance maps using Monte-Carlo sampling. Accurately predicting risk requires knowledge of the bias as well, for which we leverage Stein's Unbiased Risk Estimator (SURE) as a proxy for mean-squared-error (MSE). A range of empirical experiments is performed for Knee MRI reconstruction under different training losses (adversarial and pixel-wise) and unrolled recurrent network architectures. Our key observations indicate that: 1) adversarial losses introduce more uncertainty; and 2) recurrent unrolled nets reduce the prediction uncertainty and risk.

Lack of ground-truth MR images impedes the common supervised training of neural networks for image reconstruction. To cope with this challenge, this article leverages unpaired adversarial training for reconstruction networks, where the inputs are undersampled k-space and naively reconstructed images from one dataset, and the labels are high-quality images from another dataset. The reconstruction networks consist of a generator which suppresses the input image artifacts, and a discriminator using a pool of (unpaired) labels to adjust the reconstruction quality. The generator is an unrolled neural network - a cascade of convolutional and data consistency layers. The discriminator is also a multilayer CNN that plays the role of a critic scoring the quality of reconstructed images based on the Wasserstein distance. Our experiments with knee MRI datasets demonstrate that the proposed unpaired training enables diagnostic-quality reconstruction when high-quality image labels are not available for the input types of interest, or when the amount of labels is small. In addition, our adversarial training scheme can achieve better image quality (as rated by expert radiologists) compared with the paired training schemes with pixel-wise loss.

Undersampled magnetic resonance image (MRI) reconstruction is typically an ill-posed linear inverse task. The time and resource intensive computations require tradeoffs between accuracy and speed. In addition, state-of-the-art compressed sensing (CS) analytics are not cognizant of the image diagnostic quality. To address these challenges, we propose a novel CS framework that uses generative adversarial networks (GAN) to model the (low-dimensional) manifold of high-quality MR images. Leveraging a mixture of least-squares (LS) GANs and pixel-wise ℓ1/ℓ2cost, a deep residual network with skip connections is trained as the generator that learns to remove the aliasing artifacts by projecting onto the image manifold. The LSGAN learns the texture details, while the ℓ1/ℓ2cost suppresses high-frequency noise. A discriminator network, which is a multilayer convolutional neural network (CNN), plays the role of a perceptual cost that is then jointly trained based on high-quality MR images to score the quality of retrieved images. In the operational phase, an initial aliased estimate (e.g., simply obtained by zero-filling) is propagated into the trained generator to output the desired reconstruction. This demands a very low computational overhead. Extensive evaluations are performed on a large contrast-enhanced MR dataset of pediatric patients. Images rated by expert radiologists corroborate that GANCS retrieves higher quality images with improved fine texture details compared with conventional Wavelet-based and dictionary-learning-based CS schemes as well as with deep-learning-based schemes using pixel-wise training. In addition, it offers reconstruction times of under a few milliseconds, which are two orders of magnitude faster than the current state-of-the-art CS-MRI schemes.

Recovering high-resolution images from limited sensory data typically leads to a serious ill-posed inverse problem, demanding inversion algorithms that effectively capture the prior information. Learning a good inverse mapping from training data faces severe challenges, including: (i) scarcity of training data; (ii) need for plausible reconstructions that are physically feasible; (iii) need for fast reconstruction, especially in real-time applications. We develop a successful system solving all these challenges, using as basic architecture the repetitive application of alternating proximal and data fidelity constraints. We learn a proximal map that works well with real images based on residual networks with recurrent blocks. Extensive experiments are carried out under different settings: (a) reconstructing abdominal MRI of pediatric patients from highly undersampled k-space data and (b) super-resolving natural face images. Our key findings include: 1. a recurrent ResNet with a single residual block (10-fold repetition) yields an effective proximal which accurately reveals MR image details. 2. Our architecture significantly outperforms conventional non-recurrent deep ResNets by 2dB SNR; it is also trained much more rapidly. 3. It outperforms state-of-the-art compressed-sensing Wavelet-based methods by 4dB SNR, with 100x speedups in reconstruction time.

This paper demonstrates a robust diffusion-weighted single-shot fast spin echo (SS-FSE) sequence in the presence of significant off-resonance, which includes a variable-density acquisition and a self-calibrated reconstruction as improvements. A non-Carr-Purcell-Meiboom-Gill (nCPMG) SS-FSE acquisition stabilizes both the main and parasitic echo families for each echo. This preserves both the in-phase and quadrature components of the magnetization throughout the echo train. However, nCPMG SS-FSE also promotes aliasing of the quadrature component, which complicates reconstruction. A new acquisition and reconstruction approach is presented here, where the field-of-view is effectively doubled, but a partial k-space and variable density sampling is used to improve scan efficiency. The technique is presented in phantom scans to validate SNR and robustness against rapidly varying object phase. In vivo healthy volunteer examples and the clinical cases are demonstrated in abdominal imaging. This new approach provides comparable SNR to previous nCPMG acquisition techniques as well as providing more uniform apparent diffusion coefficient maps in phantom scans. In vivo scans suggest that this method is more robust against motion than previous approaches. The proposed reconstruction is an improvement to the nCPMG sequence as it is auto-calibrating and is justified to accurately treat the signal model for the nCPMG SS-FSE sequence.

SS-FSE is a fast technique that does not suffer from off-resonance distortions to the degree that EPI does. Unlike EPI, SS-FSE is ill-suited to diffusion weighted imaging (DWI) due to the Carr-Purcell-Meiboom-Geill (CPMG) condition. Non-CPMG phase cycling does accommodate SS-FSE and DWI but places constraints on reconstruction, which are resolved here through parallel imaging. Additionally, improved echo stability can be achieved by using short duration and highly selective DIVERSE radiofrequency pulses. Here, signal-to-noise ratio (SNR) comparisons between EPI and nCPMG SS-FSE acquisitions and reconstruction techniques give similar values. Diffusion imaging with nCPMG SS-FSE gives similar SNR to an EPI acquisition, though apparent diffusion coefficient values are higher than seen with EPI. In vivo images have good image quality with little distortion. This method has the ability to capture distortion-free DWI images near areas of significant off-resonance as well as preserve adequate SNR. Parallel imaging and DIVERSE refocusing RF pulses allow shorter ETL compared to previous implementations and thus reduces phase encode direction blur and SAR accumulation.

Q-spoiling is the process of decoupling an MRI receive coil to protect the equipment and patient. Conventionally, Q-spoiling is performed using a PIN diode switch that draws significant current. In this work, a Q-spoiling technique using a depletion-mode Gallium Nitride HEMT device was developed for coil detuning at both 1.5 T and 3 T MRI. The circuits with conventional PIN diode Q-spoiling and the GaN HEMT device were implemented on surface coils. SNR was measured and compared for all surfaces coils. At both 1.5 T and 3 T, comparable SNR was achieved for all coils with the proposed technique and conventional Q-spoiling. The GaN HEMT device has significantly reduced the required power for Q-spoiling. The GaN HEMT device also provides useful safety features by detuning the coil when unpowered.

We present l₁-SPIRiT, a simple algorithm for auto calibrating parallel imaging (acPI) and compressed sensing (CS) that permits an efficient implementation with clinically-feasible runtimes. We propose a CS objective function that minimizes cross-channel joint sparsity in the wavelet domain. Our reconstruction minimizes this objective via iterative soft-thresholding, and integrates naturally with iterative self-consistent parallel imaging (SPIRiT). Like many iterative magnetic resonance imaging reconstructions, l₁-SPIRiT's image quality comes at a high computational cost. Excessively long runtimes are a barrier to the clinical use of any reconstruction approach, and thus we discuss our approach to efficiently parallelizing l₁-SPIRiT and to achieving clinically-feasible runtimes. We present parallelizations of l₁-SPIRiT for both multi-GPU systems and multi-core CPUs, and discuss the software optimization and parallelization decisions made in our implementation. The performance of these alternatives depends on the processor architecture, the size of the image matrix, and the number of parallel imaging channels. Fundamentally, achieving fast runtime requires the correct trade-off between cache usage and parallelization overheads. We demonstrate image quality via a case from our clinical experimentation, using a custom 3DFT spoiled gradient echo (SPGR) sequence with up to 8× acceleration via Poisson-disc undersampling in the two phase-encoded directions.