Quantitative single point imaging with compressed sensing

Department of Chemical Engineering and Biotechnology, University of Cambridge, Cambridge CB2 3RA, UK.
Journal of Magnetic Resonance (Impact Factor: 2.51). 09/2009; 201(1):72-80. DOI: 10.1016/j.jmr.2009.08.003
Source: PubMed


A novel approach with respect to single point imaging (SPI), compressed sensing, is presented here that is shown to significantly reduce the loss of accuracy of reconstructed images from under-sampled acquisition data. SPI complements compressed sensing extremely well as it allows unconstrained selection of sampling trajectories. Dynamic processes featuring short T2* NMR signal can thus be more rapidly imaged, in our case the absorption of moisture by a cereal-based wafer material, with minimal loss of image quantification. The absolute moisture content distribution is recovered via a series of images acquired with variable phase encoding times allowing extrapolation to time zero for each image pixel and the effective removal of T2* contrast.

Download full-text


Available from: Michael Johns, Jul 25, 2014
  • Source
    • "CS works by exploiting the natural structure of MR images to reconstruct images accurately from partially sampled k-space data. CS has been applied to many systems [17] [18] [19] [20] [21] and pulse sequences but to the authors knowledge, has not yet been used with UTE. "
    [Show abstract] [Hide abstract]
    ABSTRACT: Ultrashort echo time (UTE) imaging is a well-known technique used in medical MRI, however, the implementation of the sequence remains non-trivial. This paper introduces UTE for non-medical applications and outlines a method for the implementation of UTE to enable accurate slice selection and short acquisition times. Slice selection in UTE requires fast, accurate switching of the gradient and r.f. pulses. Here a gradient "pre-equalization" technique is used to optimize the gradient switching and achieve an effective echo time of 10 mu s. In order to minimize the echo time, k-space is sampled radially. A compressed sensing approach is used to minimize the total acquisition time. Using the corrections for slice selection and acquisition along with novel image reconstruction techniques, UTE is shown to be a viable method to study samples of cork and rubber with a shorter signal lifetime than can typically be measured. Further, the compressed sensing image reconstruction algorithm is shown to provide accurate images of the samples with as little as 12.5% of the full k-space data set, potentially permitting real time imaging of short T-2* materials. (C) 2014 The Authors. Published by Elsevier Inc.
    Full-text · Article · Aug 2014 · Journal of Magnetic Resonance
  • Source
    • "In the present study, a 2D setup and small field-of-views and acquisition matrices were used to achieve adequate spatial resolution, while avoiding excessive scan times and still be able to demonstrate essential features. In future work we will try to push the limits by incorporating some of the acceleration tools that we already discussed and partly demonstrated in previous work [7] [17] [18], including multiple spin echo techniques, multi-point k-space mapping techniques, parallel imaging techniques, and sparse sampling techniques [19] [20] [21]. Finally it should be realized that the slowness of mSPI and the necessity to synchronize the acquisition with a repetitive signal disturbance both emerge from the spatial encoding part of the technique. "
    [Show abstract] [Hide abstract]
    ABSTRACT: In this paper we aim to lay down and demonstrate the use of multiple single-point imaging (mSPI) as a tool for capturing and characterizing steady-state MR signals and repetitive disturbances thereof with high temporal resolution. To achieve this goal, various 2D mSPI sequences were derived from the nearest standard 3D imaging sequences by (i) replacing the excitation of a 3D slab by the excitation of a 2D slice orthogonal to the read axis, (ii) setting the readout gradient to zero, and (iii) leaving out the inverse Fourier transform in the read direction. The thus created mSPI sequences, albeit slow with regard to the spatial encoding part, were shown to result into a series of densely spaced 2D single-point images in the time domain enabling monitoring of the evolution of the magnetization with a high temporal resolution and without interference from any encoding gradients. The high-speed capabilities of mSPI were demonstrated by capturing and characterizing the free induction decays and spin echoes of substances with long T2s (>30ms) and long and short T2*s (4->30ms) and by monitoring the perturbation of the transverse magnetization by, respectively, a titanium cylinder, representing a static disturbance; a pulsed magnetic field gradient, representing a stimulus inherent to a conventional MRI experiment; and a pulsed electric current, representing an external stimulus. The results of the study indicate the potential of mSPI for assessing the evolution of the magnetization and, when properly synchronized with the acquisition, repeatable disturbances thereof with a temporal resolution that is ultimately limited by the bandwidth of the receiver, but in practice governed by the SNR of the experiment and the magnitude of the disturbance. Potential applications of mSPI can be envisaged in research areas that are concerned with MR signal behavior, MR system performance and MR evaluation of magnetically evoked responses.
    Full-text · Article · Jun 2013 · Magnetic Resonance Imaging
  • Source
    • "As with spectroscopic imaging, many techniques have been proposed to speed up the basic SPI experiment, including turbo spin-echo SPI [20] [21] [22] [23] [24] [25]; single-point ramped imaging with T1 enhancement, a technique in which one of the phase-encode gradients is ramped in equally spaced discrete steps [26]; multipoint k-space mapping techniques [27] [28]; SPI with variable phase encoding interval [29]; and, last but not least, sparse k-space sampling [30]. The latter technique, also termed compressed sensing (CS), has already been shown to be very effective when imaging objects which possess a sparse representation in a certain domain. "
    [Show abstract] [Hide abstract]
    ABSTRACT: Lack of spatial accuracy is a recognized problem in magnetic resonance imaging (MRI) which severely detracts from its value as a stand-alone modality for applications that put high demands on geometric fidelity, such as radiotherapy treatment planning and stereotactic neurosurgery. In this paper, we illustrate the potential and discuss the limitations of spectroscopic imaging as a tool for generating purely phase-encoded MR images and parameter maps that preserve the geometry of an object and allow localization of object features in world coordinates. Experiments were done on a clinical system with standard facilities for imaging and spectroscopy. Images were acquired with a regular spin echo sequence and a corresponding spectroscopic imaging sequence. In the latter, successive samples of the acquired echo were used for the reconstruction of a series of evenly spaced images in the time and frequency domain. Experiments were done with a spatial linearity phantom and a series of test objects representing a wide range of susceptibility- and chemical-shift-induced off-resonance conditions. In contrast to regular spin echo imaging, spectroscopic imaging was shown to be immune to off-resonance effects, such as those caused by field inhomogeneity, susceptibility, chemical shift, f(0) offset and field drift, and to yield geometrically accurate images and parameter maps that allowed object structures to be localized in world coordinates. From these illustrative examples and a discussion of the limitations of purely phase-encoded imaging techniques, it is concluded that spectroscopic imaging offers a fundamental solution to the geometric deficiencies of MRI which may evolve toward a practical solution when full advantage will be taken of current developments with regard to scan time reduction. This perspective is backed up by a demonstration of the significant scan time reduction that may be achieved by the use of compressed sensing for a simple phantom.
    Full-text · Article · Aug 2012 · Magnetic Resonance Imaging
Show more