Transformers have gained prominence in natural language processing due to their representational capabilities and performances. Transformers process natural language as a sequence on finite context windows; however, global relationships among words beyond these windows cannot be completely modeled via sequence processing only. Graph neural network (GNN) based models have been proposed to alleviate this problem, as they provide geometric extensions to neural networks, enabling models to learn associations within a text. However, regular graph-based methods ignore the sequential nature of underlying texts. In this paper, we propose EmoVis, the first generic graph-based neural network that utilizes visibility graphs, which converts classical time-series information to graph representations. We cast the problem as an emotion classification task, enabling the proposed model to learn associations between the labels and words in a sentence. Moreover, EmoVis can be used as a highly modular graph-based extension to any transformer-based model, significantly improving their performance and learning capabilities in various languages. We experimentally show that EmoVis enables transformer-based models to outperform the state-of-the-art baselines across three diverse datasets in different languages in the SemEval2018 competition datasets and the GoEmotions dataset.

Monitoring of prevalent airborne diseases such as COVID-19 characteristically involves respiratory assessments. While auscultation is a mainstream method for preliminary screening of disease symptoms, its utility is hampered by the need for dedicated hospital visits. Remote monitoring based on recordings of respiratory sounds on portable devices is a promising alternative, which can assist in early assessment of COVID-19 that primarily affects the lower respiratory tract. In this study, we introduce a novel deep learning approach to distinguish patients with COVID-19 from healthy controls given audio recordings of cough or breathing sounds. The proposed approach leverages a novel hierarchical spectrogram transformer (HST) on spectrogram representations of respiratory sounds. HST embodies self-attention mechanisms over local windows in spectrograms, and window size is progressively grown over model stages to capture local to global context. HST is compared against state-of-the-art conventional and deep-learning baselines. Demonstrations on crowd-sourced multi-national datasets indicate that HST outperforms competing methods, achieving over 90% area under the receiver operating characteristic curve (AUC) in detecting COVID-19 cases.

Magnetic particle imaging (MPI) offers unparalleled contrast and resolution for tracing magnetic nanoparticles. A common imaging procedure calibrates a system matrix (SM) that is used to reconstruct data from subsequent scans. The ill-posed reconstruction problem can be solved by simultaneously enforcing data consistency based on the SM and regularizing the solution based on an image prior. Traditional hand-crafted priors cannot capture the complex attributes of MPI images, whereas recent MPI methods based on learned priors can suffer from extensive inference times or limited generalization performance. Here, we introduce a novel physics-driven method for MPI reconstruction based on a deep equilibrium model with learned data consistency (DEQ-MPI). DEQ-MPI reconstructs images by augmenting neural networks into an iterative optimization, as inspired by unrolling methods in deep learning. Yet, conventional unrolling methods are computationally restricted to few iterations resulting in non-convergent solutions, and they use hand-crafted consistency measures that can yield suboptimal capture of the data distribution. DEQ-MPI instead trains an implicit mapping to maximize the quality of a convergent solution, and it incorporates a learned consistency measure to better account for the data distribution. Demonstrations on simulated and experimental data indicate that DEQ-MPI achieves superior image quality and competitive inference time to state-of-the-art MPI reconstruction methods.

Magnetic particle imaging (MPI) is a rapidly developing medical imaging modality that exploits the non-linear response of magnetic nanoparticles (MNPs). Color MPI widens the functionality of MPI, empowering it with the capability to distinguish different MNPs and/or MNP environments. The system function approach for color MPI relies on extensive calibrations that capture the differences in the harmonic responses of the MNPs. An alternative calibration-free x-space-based method called TAURUS estimates a map of the relaxation time constant, τ , by recovering the underlying mirror symmetry in the MPI signal. However, TAURUS requires a back and forth scanning of a given region, restricting its usage to slow trajectories with constant or piecewise constant focus fields (FFs). In this work, we propose a novel technique to increase the performance of TAURUS and enable τ map estimation for rapid and multi-dimensional trajectories. The proposed technique is based on correcting the distortions on mirror symmetry induced by time-varying FFs. We demonstrate via simulations and experiments in our in-house MPI scanner that the proposed method successfully estimates high-fidelity τ maps for rapid trajectories that provide orders of magnitude reduction in scanning time (over 300 fold for simulations and over 8 fold for experiments) while preserving the calibration-free property of TAURUS.

Magnetic particle imaging (MPI) offers exceptional contrast for magnetic nanoparticles (MNP) at high spatio-temporal resolution. A common procedure in MPI starts with a calibration scan to measure the system matrix (SM), which is then used to set up an inverse problem to reconstruct images of the MNP distribution during subsequent scans. This calibration enables the reconstruction to sensitively account for various system imperfections. Yet time-consuming SM measurements have to be repeated under notable changes in system properties. Here, we introduce a novel deep learning approach for accelerated MPI calibration based on Transformers for SM super-resolution (TranSMS). Low-resolution SM measurements are performed using large MNP samples for improved signal-to-noise ratio efficiency, and the high-resolution SM is super-resolved via model-based deep learning. TranSMS leverages a vision transformer module to capture contextual relationships in low-resolution input images, a dense convolutional module for localizing high-resolution image features, and a data-consistency module to ensure measurement fidelity. Demonstrations on simulated and experimental data indicate that TranSMS significantly improves SM recovery and MPI reconstruction for up to 64-fold acceleration in two-dimensional imaging.

Magnetic Particle Imaging (MPI) is an emerging medical imaging modality that images the spatial distribution of superparamagnetic iron oxide (SPIO) nanoparticles using their nonlinear response to applied magnetic fields. In standard x-space approach to MPI, the image is reconstructed by gridding the speed-compensated nanoparticle signal to the instantaneous position of the field free point (FFP). However, due to safety limits on the drive field, the field-of-view (FOV) needs to be covered by multiple relatively small partial field-of-views (pFOVs). The image of the entire FOV is then pieced together from individually processed pFOVs. These processing steps can be sensitive to non-ideal signal conditions such as harmonic interference, noise, and relaxation effects. In this work, we propose a robust x-space reconstruction technique, Partial FOV Center Imaging (PCI), with substantially simplified pFOV processing. PCI first forms a raw image of the entire FOV by mapping MPI signal directly to pFOV center locations. The corresponding MPI image is then obtained by deconvolving this raw image by a compact kernel, whose fully-known shape solely depends on the pFOV size. We analyze the performance of the proposed reconstruction via extensive simulations, as well as imaging experiments on our in-house FFP MPI scanner. The results show that PCI offers a trade-off between noise robustness and interference robustness, outperforming standard x-space reconstruction in terms of both robustness against non-ideal signal conditions and image quality.

Magnetic particle imaging (MPI) is a relatively new medical imaging modality, which detects the nonlinear response of magnetic nanoparticles (MNPs) that are exposed to external magnetic fields. The system matrix (SM) method for MPI image reconstruction requires a time consuming system calibration scan prior to image acquisition, where a single MNP sample is measured at each voxel position in the field-of-view (FOV). The scanned sample has the maximum size of a voxel so that the calibration measurements have relatively poor signal-to-noise ratio (SNR). In this paper, we present the coded calibration scene (CCS) framework, where we place multiple MNP samples inside the FOV in a random or pseudo-random fashion. Taking advantage of the sparsity of the SM, we reconstruct the SM by solving a convex optimization problem with alternating direction method of multipliers using CCS measurements. We analyze the effects of filling rate, number of measurements, and SNR on the SM reconstruction using simulations and demonstrate different implementations of CCS for practical realization. We also compare the imaging performance of the proposed framework with that of a standard compressed sensing SM reconstruction that utilizes a subset of calibration measurements from a single MNP sample. The results show that CCS significantly reduces calibration time while increasing both the SM reconstruction and image reconstruction performances.

A central limitation of multiple-acquisition magnetic resonance imaging (MRI) is the degradation in scan efficiency as the number of distinct datasets grows. Sparse recovery techniques can alleviate this limitation via randomly undersampled acquisitions. A frequent sampling strategy is to prescribe for each acquisition a different random pattern drawn from a common sampling density. However, naive random patterns often contain gaps or clusters across the acquisition dimension that, in turn, can degrade reconstruction quality or reduce scan efficiency. To address this problem, a statistically segregated sampling method is proposed for multiple-acquisition MRI. This method generates multiple patterns sequentially while adaptively modifying the sampling density to minimize k-space overlap across patterns. As a result, it improves incoherence across acquisitions while still maintaining similar sampling density across the radial dimension of k-space. Comprehensive simulations and in vivo results are presented for phase-cycled balanced steady-state free precession and multi-echo [Formula: see text]-weighted imaging. Segregated sampling achieves significantly improved quality in both Fourier and compressed-sensing reconstructions of multiple-acquisition datasets.

The compressed sensing (CS) framework leverages the sparsity of MR images to reconstruct from the undersampled acquisitions. CS reconstructions involve one or more regularization parameters that weigh sparsity in transform domains against fidelity to acquired data. While parameter selection is critical for reconstruction quality, the optimal parameters are subject and dataset specific. Thus, commonly practiced heuristic parameter selection generalizes poorly to independent datasets. Recent studies have proposed to tune parameters by estimating the risk of removing significant image coefficients. Line searches are performed across the parameter space to identify the parameter value that minimizes this risk. Although effective, these line searches yield prolonged reconstruction times. Here, we propose a new self-tuning CS method that uses computationally efficient projections onto epigraph sets of the${\ell }_{{1}}$and total-variation norms to simultaneously achieve parameter selection and regularization. In vivo demonstrations are provided for balanced steady-state free precession, time-of-flight, and T1-weighted imaging. The proposed method achieves an order of magnitude improvement in computational efficiency over line-search methods while maintaining near-optimal parameter selection.

Magnetic particle imaging (MPI) is a novel imaging modality with important potential applications, such as angiography, stem cell tracking, and cancer imaging. Recently, there have been efforts to increase the functionality of MPI via multi-color imaging methods that can distinguish the responses of different nanoparticles, or nanoparticles in different environmental conditions. The proposed techniques typically rely on extensive calibrations that capture the differences in the harmonic responses of the nanoparticles. In this paper, we propose a method to directly estimate the relaxation time constant of the nanoparticles from the MPI signal, which is then used to generate a multi-color relaxation map. The technique is based on the underlying mirror symmetry of the adiabatic MPI signal when the same region is scanned back and forth. We validate the proposed method via simulations, and via experiments on our in-house magnetic particle spectrometer setup at 10.8 kHz and our in-house MPI scanner at 9.7 kHz. Our results show that nanoparticles can be successfully distinguished with the proposed technique, without any calibration or prior knowledge about the nanoparticles.

Magnetic particle imaging (MPI) shows promise for medical imaging, particularly in angiography of patients with chronic kidney disease. As the first biomedical imaging technique that truly depends on nanoscale materials properties, MPI requires highly optimized magnetic nanoparticle tracers to generate quality images. Until now, researchers have relied on tracers optimized for MRI T2(∗) -weighted imaging that are sub-optimal for MPI. Here, we describe new tracers tailored to MPI's unique physics, synthesized using an organic-phase process and functionalized to ensure biocompatibility and adequate in vivo circulation time. Tailored tracers showed up to 3 × greater signal-to-noise ratio and better spatial resolution than existing commercial tracers in MPI images of phantoms.

Magnetic Particle Imaging (MPI) is a promising tracer imaging modality that employs a kidney-safe contrast agent and does not use ionizing radiation. MPI already shows high contrast and sensitivity in small animal imaging, with great potential for many clinical applications, including angiography, cancer detection, inflammation imaging, and treatment monitoring. Currently, almost all clinically relevant imaging techniques can be modeled as systems with linearity and shift invariance (LSI), characteristics crucial for quantification and diagnostic utility. In theory, MPI has been proven to be LSI. However, in practice, high-pass filters designed to remove unavoidable direct feedthrough interference also remove information crucial to ensuring LSI in MPI scans. In this work, we present a complete theoretical and experimental description of the image artifacts from filtering. We then propose and validate a robust algorithm to completely restore the lost information for the x-space MPI method. We provide the theoretical, simulated, and experimental proof that our algorithm indeed restores the LSI properties of MPI.

For magnetic particle imaging (MPI), specific absorption rate (SAR) and more critically magnetostimulation (i.e., dB/dt) safety limits will determine the optimal scan parameters, such as the drive field strength and frequency. These parameters will impact the scanning speed, field-of-view (FOV) and signal-to-noise ratio in MPI. Understanding the potential safety hazards of the drive field is critical for scaling MPI for human use. In this work, we demonstrate that magnetostimulation is the primary magnetic safety consideration in MPI, and we describe the first human-subject magnetostimulation threshold experiments for MPI using homogeneous coils. Our experiments, performed on the arm and leg, indicate that magnetostimulation thresholds monotonically decrease with increasing frequency. Additionally, we show for the first time that a strong inverse correlation exists between the threshold and the body part size. The chronaxie time, on the other hand, did not vary with body part size. We conclude with an estimation of the magnetostimulation thresholds for a full-body MPI scanner: a mean asymptotic threshold of 14.3 mT-pp (peak-to-peak) with a mean chronaxie time of 289 μs, which correspond to a magnetostimulation threshold of about 15 mT-pp for frequencies between 25 and 50 kHz. These findings will have a great impact on the optimization of MPI parameters, especially in determining the number of partial FOVs required to cover a region of interest.

Projection magnetic particle imaging (MPI) can improve imaging speed by over 100-fold over traditional 3-D MPI. In this work, we derive the 2-D x-space signal equation, 2-D image equation, and introduce the concept of signal fading and resolution loss for a projection MPI imager. We then describe the design and construction of an x-space projection MPI scanner with a field gradient of 2.35 T/m across a 10 cm magnet free bore. The system has an expected resolution of 3.5 × 8.0 mm using Resovist tracer, and an experimental resolution of 3.8 × 8.4 mm resolution. The system images 2.5 cm × 5.0 cm partial field-of views (FOVs) at 10 frames/s, and acquires a full field-of-view of 10 cm × 5.0 cm in 4 s. We conclude by imaging a resolution phantom, a complex "Cal" phantom, mice injected with Resovist tracer, and experimentally confirm the theoretically predicted x-space spatial resolution.

Magnetization-prepared acquisitions offer a trade-off between image contrast and scan efficiency for magnetic resonance imaging. Because the prepared signals gradually decay, the contrast can be improved by frequently repeating the preparation, which in turn significantly increases the scan time. A common solution is to perform the data collection progressing from low- to high-spatial-frequency samples following each preparation. Unfortunately, this leads to loss of spatial resolution, and thereby image blurring. In this work, a new technique is proposed that first corrects the signal decay in high-frequency data to mitigate the resolution loss and improve the image contrast without reducing the scan efficiency. The proposed technique then employs a sparsity-based nonlinear reconstruction to further improve the image quality. In addition to reducing the amplified high-frequency noise, this reconstruction extrapolates missing k-space samples in the case of undersampled compressed-sensing acquisitions. The technique is successfully demonstrated for noncontrast-enhanced flow-independent angiography of the lower extremities, an application that substantially benefits from both the signal compensation and the nonlinear reconstruction.

Optimizing the diffusion-weighted imaging (DWI) parameters (i.e., the b-value and the number of image averages) to the tissue of interest is essential for producing high-quality apparent diffusion coefficient (ADC) maps. Previous investigation of this optimization was performed assuming Gaussian noise statistics for the ADC map, which is only valid for high signal-to-noise ratio (SNR) imaging. In this work, the true statistics of the noise in ADC maps are derived, followed by an optimization of the DWI parameters as a function of the imaging SNR. Specifically, it is demonstrated that the optimum b-value is a monotonically increasing function of the imaging SNR, which converges to the optimum b-value from previously proposed approaches for high-SNR cases, while exhibiting a significant deviation from this asymptote for low-SNR situations. Incorporating the effects of T(2) weighting further increases the SNR dependence of the optimal parameters. The proposed optimization scheme is particularly important for high-resolution DWI, which intrinsically suffers from low SNR and therefore cannot afford the use of the conventional high b-values. Comparison scans were performed for high-resolution DWI of the spinal cord, demonstrating the improvements in the resulting images and the ADC maps achieved by this method.