Recovering Weak Signals in Magnetic Resonance Spectroscopy with SNRs of Order Unity
Madhur Srivastava
Cornell University, Ithaca, United States

Noise impedes experimental studies by reducing signal resolution and/or suppressing weak peaks, especially at low SNRs. We developed a wavelet transform-based approach to effectively remove noise from the spectroscopic data. We demonstrate its power in extracting signals from noisy spectra on a variety of signal types ranging from hyperfine lines to overlapped peaks to weak peaks overlaid on strong ones. The results show that one can accurately extract details of complex spectra, including retrieval of very weak ones. It accurately recovers signals at SNR ~ 1 and improves SNR by about three orders of magnitude with high fidelity. The method referred to as NERD (Noise Elimination and Reduction via Denoising) can be accessed via denoising.cornell.edu

Non-Cartesian Trajectories for Magnetic Resonance Imaging and Spectroscopy
Uzay Emir
Purdue University, West Lafayette, United States

This talk will introduce and demonstrate non-cartesian trajectories for Magnetic Resonance Imaging and Spectroscopy, providing flexible structural and molecular contrast to probe human organs noninvasively within a clinically feasible acquisition time at 3T. The methods have been validated in a series of phantom experiments, and their feasibility was assessed in healthy volunteers in challenging applications such as cartilage, lung, and cerebellum imaging. Experiments qualitatively demonstrate the advantage of the proposed methods in terms of their improved resolution, reduced contamination from neighboring voxels, and further acceleration with novel reconstruction techniques.

"SPICY" MRSI: Ultrafast MR Spectroscopic Imaging with Learned Subspaces
Zhi-Pei Liang
University of Illinois at Urbana-Champaign, Urbana, United States

MR spectroscopic imaging (MRSI) has long been recognized as a potentially powerful tool for noninvasive, label-free molecular imaging. This talk will discuss our recent advances in ultrafast MRSI using a new technology known as SPICE (SPectroscopic Imaging by exploiting spatiospectral CorrElation). SPICE uses a subspace mathematical framework to effectively integrate rapid scanning, sparse sampling, constrained image reconstruction, quantum simulation, and machine learning. Experimental results show an unprecedented capability for simultaneous mapping of brain structures, function and metabolism using intrinsic spin signals from multiple molecules. In this talk, I’ll will give an overview of SPICE and also show some “SPICY” experimental results we have obtained.