Image reconstruction is essential for imaging applications across the physical and life sciences, including optical and radar systems, magnetic resonance imaging, X-ray computed tomography, positron ...emission tomography, ultrasound imaging and radio astronomy. During image acquisition, the sensor encodes an intermediate representation of an object in the sensor domain, which is subsequently reconstructed into an image by an inversion of the encoding function. Image reconstruction is challenging because analytic knowledge of the exact inverse transform may not exist a priori, especially in the presence of sensor non-idealities and noise. Thus, the standard reconstruction approach involves approximating the inverse function with multiple ad hoc stages in a signal processing chain, the composition of which depends on the details of each acquisition strategy, and often requires expert parameter tuning to optimize reconstruction performance. Here we present a unified framework for image reconstruction-automated transform by manifold approximation (AUTOMAP)-which recasts image reconstruction as a data-driven supervised learning task that allows a mapping between the sensor and the image domain to emerge from an appropriate corpus of training data. We implement AUTOMAP with a deep neural network and exhibit its flexibility in learning reconstruction transforms for various magnetic resonance imaging acquisition strategies, using the same network architecture and hyperparameters. We further demonstrate that manifold learning during training results in sparse representations of domain transforms along low-dimensional data manifolds, and observe superior immunity to noise and a reduction in reconstruction artefacts compared with conventional handcrafted reconstruction methods. In addition to improving the reconstruction performance of existing acquisition methodologies, we anticipate that AUTOMAP and other learned reconstruction approaches will accelerate the development of new acquisition strategies across imaging modalities.
Purpose
Demonstrate a novel fast method for reconstruction of multi‐dimensional MR fingerprinting (MRF) data using deep learning methods.
Methods
A neural network (NN) is defined using the TensorFlow ...framework and trained on simulated MRF data computed with the extended phase graph formalism. The NN reconstruction accuracy for noiseless and noisy data is compared to conventional MRF template matching as a function of training data size and is quantified in simulated numerical brain phantom data and International Society for Magnetic Resonance in Medicine/National Institute of Standards and Technology phantom data measured on 1.5T and 3T scanners with an optimized MRF EPI and MRF fast imaging with steady state precession (FISP) sequences with spiral readout. The utility of the method is demonstrated in a healthy subject in vivo at 1.5T.
Results
Network training required 10 to 74 minutes; once trained, data reconstruction required approximately 10 ms for the MRF EPI and 76 ms for the MRF FISP sequence. Reconstruction of simulated, noiseless brain data using the NN resulted in a RMS error (RMSE) of 2.6 ms for T1 and 1.9 ms for T2. The reconstruction error in the presence of noise was less than 10% for both T1 and T2 for SNR greater than 25 dB. Phantom measurements yielded good agreement (R2 = 0.99/0.99 for MRF EPI T1/T2 and 0.94/0.98 for MRF FISP T1/T2) between the T1 and T2 estimated by the NN and reference values from the International Society for Magnetic Resonance in Medicine/National Institute of Standards and Technology phantom.
Conclusion
Reconstruction of MRF data with a NN is accurate, 300‐ to 5000‐fold faster, and more robust to noise and dictionary undersampling than conventional MRF dictionary‐matching.
We introduce a broadly applicable technique to create nuclear spin singlet states in organic molecules and other many-atom systems. We employ a novel pulse sequence to produce a spin-lock induced ...crossing (SLIC) of the spin singlet and triplet energy levels, which enables triplet-singlet polarization transfer and singlet-state preparation. We demonstrate the utility of the SLIC method by producing a long-lived nuclear spin singlet state on two strongly coupled proton pairs in the tripeptide molecule phenylalanine-glycine-glycine dissolved in D(2)O and by using SLIC to measure the J couplings, chemical shift differences, and singlet lifetimes of the proton pairs. We show that SLIC is more efficient at creating nearly equivalent nuclear spin singlet states than previous pulse sequence techniques, especially when triplet-singlet polarization transfer occurs on the same time scale as spin-lattice relaxation.
Recent years have seen a resurgence of interest in inexpensive low magnetic field (< 0.3 T) MRI systems mainly due to advances in magnet, coil and gradient set designs. Most of these advances ...have focused on improving hardware and signal acquisition strategies, and far less on the use of advanced image reconstruction methods to improve attainable image quality at low field. We describe here the use of our end-to-end deep neural network approach (AUTOMAP) to improve the image quality of highly noise-corrupted low-field MRI data. We compare the performance of this approach to two additional state-of-the-art denoising pipelines. We find that AUTOMAP improves image reconstruction of data acquired on two very different low-field MRI systems: human brain data acquired at 6.5 mT, and plant root data acquired at 47 mT, demonstrating SNR gains above Fourier reconstruction by factors of 1.5- to 4.5-fold, and 3-fold, respectively. In these applications, AUTOMAP outperformed two different contemporary image-based denoising algorithms, and suppressed noise-like spike artifacts in the reconstructed images. The impact of domain-specific training corpora on the reconstruction performance is discussed. The AUTOMAP approach to image reconstruction will enable significant image quality improvements at low-field, especially in highly noise-corrupted environments.
Abstract Purpose To evaluate the DSM-5 diagnosis of Avoidant/Restrictive Food Intake Disorder (ARFID) in children and adolescents with poor eating not associated with body image concerns. Methods A ...retrospective case-control study of 8–18-year-olds, using a diagnostic algorithm, compared all cases with ARFID presenting to seven adolescent-medicine eating disorder programs in 2010 to a randomly selected sample with anorexia nervosa (AN) and bulimia nervosa (BN). Demographic and clinical information were recorded. Results Of 712 individuals studied, 98 (13.8%) met ARFID criteria. Patients with ARFID were younger than those with AN (n = 98) or BN (n = 66), (12.9 vs. 15.6 vs. 16.5 years), had longer durations of illness (33.3 vs. 14.5 vs. 23.5 months), were more likely to be male (29% vs. 15% vs. 6%), and had a percent median body weight intermediate between those with AN or BN (86.5 vs. 81.0 and 107.5). Patients with ARFID included those with selective (picky) eating since early childhood (28.7%); generalized anxiety (21.4%); gastrointestinal symptoms (19.4%); a history of vomiting/choking (13.2%); and food allergies (4.1%). Patients with ARFID were more likely to have a comorbid medical condition (55% vs. 10% vs. 11%) or anxiety disorder (58% vs. 35% vs. 33%) and were less likely to have a mood disorder (19% vs. 31% vs. 58%). Conclusions Patients with ARFID were demographically and clinically distinct from those with AN or BN. They were significantly underweight with a longer duration of illness and had a greater likelihood of comorbid medical and/or psychiatric symptoms.
Although polyethylene (PE) and polypropylene (PP) are by far the world’s largest volume plastics, only a tiny fraction of these energy-rich polyolefins are currently recycled. Depolymerization of PE ...to its constituent monomer, ethylene, is highly endothermic and conventionally accessible only through unselective, high-temperature pyrolysis. Here, we provide experimental demonstrations of our recently proposed tandem catalysis strategy, which uses ethylene to convert PE to propylene, the commodity monomer used to make PP. The approach combines rapid olefin metathesis with rate-limiting isomerization. Monounsaturated PE is progressively disassembled at modest temperatures via many consecutive ethenolysis events, resulting selectively in propylene. Fully saturated PE can be converted to unsaturated PE starting with a single transfer dehydrogenation to ethylene, which produces a small amount of ethane (1 equiv per dehydrogenation event). These principles are demonstrated using both homogeneous and heterogeneous catalysts. While selectivity under batch conditions is limited at high conversion by the formation of an equilibrium mixture of olefins, high selectivity to propylene (≥94%) is achieved in a semicontinuous process due to the continuous removal of propylene from the reaction mixture.
The modular nature of metal–organic frameworks (MOFs) leads to a very large number of possible structures. High-throughput computational screening has led to a rapid increase in property data that ...has enabled several potential applications for MOFs, including gas storage, separations, catalysis, and other fields. Despite their rich chemistry, MOFs are typically named using an ad hoc approach, which can impede their searchability and the discovery of broad insights. In this article, we develop two systematic MOF identifiers, coined MOFid and MOFkey, and algorithms for deconstructing MOFs into their building blocks and underlying topological network. We review existing cheminformatics formats for small molecules and address the challenges of adapting them to periodic crystal structures. Our algorithms are distributed as open-source software, and we apply them here to extract insights from several MOF databases. Through the process of designing MOFid and MOFkey, we provide a perspective on opportunities for the community to facilitate data reuse, improve searchability, and rapidly apply cheminformatics analyses.
We cross-correlate the largest available mid-infrared (Wide-field Infrared Survey Explorer - WISE), X-ray (3XMM) and radio (Faint Images of the Radio Sky at Twenty centimetres+NRAO VLA Sky Survey) ...catalogues to define the MIXR sample of AGN and star-forming galaxies. We pre-classify the sources based on their positions on the WISE colour/colour plot, showing that the MIXR triple selection is extremely effective to diagnose the star formation and AGN activity of individual populations, even on a flux/magnitude basis, extending the diagnostics to objects with luminosities and redshifts from SDSS DR12. We recover the radio/mid-IR star formation correlation with great accuracy, and use it to classify our sources, based on their activity, as radio-loud and radio-quiet active galactic nuclei (AGN), low excitation radio galaxies/low ionization nuclear emission line regions, and non-AGN galaxies. These diagnostics can prove extremely useful for large AGN and galaxy samples, and help develop ways to efficiently triage sources when data from the next generation of instruments becomes available. We study bias in detail, and show that while the widely used WISE colour selections for AGN are very successful at cleanly selecting samples of luminous AGN, they miss or misclassify a substantial fraction of AGN at lower luminosities and/or higher redshifts. MIXR also allows us to test the relation between radiative and kinetic (jet) power in radio-loud AGN, for which a tight correlation is expected due to a mutual dependence on accretion. Our results highlight that long-term AGN variability, jet regulation, and other factors affecting the Q/L sub( bol) relation, are introducing a vast amount of scatter in this relation, with dramatic potential consequences on our current understanding of AGN feedback and its effect on star formation.