Elsevier

NeuroImage

Volume 40, Issue 4, 1 May 2008, Pages 1561-1566
NeuroImage

The molecular basis for gray and white matter contrast in phase imaging

https://doi.org/10.1016/j.neuroimage.2008.01.061Get rights and content

Abstract

Direct magnetic resonance phase images acquired at high field have been shown to yield superior gray and white matter contrast up to 10-fold higher compared to conventional magnitude images. However, the underlying contrast mechanism is not yet understood. This study demonstrates that the water resonance frequency is directly shifted by water–macromolecule exchange processes (0.040 ppm/mM for bovine serum albumin) and might be a major source of contribution to in vivo phase image contrast. Therefore, magnetic resonance phase imaging based on the proposed contrast mechanism could potentially be applied for in vivo studies of pathologies on a macromolecular level.

Introduction

The phase is an essential property of the signal-generating transverse magnetization in magnetic resonance imaging (MRI). However, in most conventional MRI studies, only the magnitude data representing tissue spin densities, relaxation properties, etc. are preserved, and the phase information is usually discarded. Exceptions do exist, e.g., in cardiac imaging where phase difference maps contain information about blood flow patterns or tissue motion. Such methods utilize the additionally accumulated phase of macroscopically mobile spins traveling inside a linear magnetic field gradient. However, little attention is paid to the intrinsic local frequency or phase differences in tissue, and most often, such frequency shifts are attributed to magnetic field inhomogeneity effects and removed.

Recently, increased interest in the image phase arose in order to enhance the image contrast, e.g., for susceptibility weighted imaging (SWI) (Abduljalil et al., 2003, Haacke et al., 2004). In addition, direct phase imaging at 1.5 T also demonstrated high contrast containing information different from magnitude images (Rauscher et al., 2005). Such phase contrast was greatly enhanced with the availability of ultra-high field. Phase images acquired at 7 T (Duyn et al., 2007) and in this study (Fig. 1) showed superior gray matter (GM) and white matter (WM) contrast with a contrast-to-noise ratio (CNR) gain of up to 10 compared to conventional magnitude images. It was tentatively suggested that the bulk tissue susceptibility accounts for the GM/WM phase contrast. Possible sources of susceptibility differences between GM and WM, such as blood deoxyhemoglobin (Duyn et al., 2007, Haacke et al., 2005), tissue myelin content (Annese et al., 2004, Duyn et al., 2007, O'Brien and Sampson, 1965), and tissue iron content (Bizzi et al., 1990, Drayer et al., 1986, Duyn et al., 2007, Haacke et al., 2005, Morris et al., 1992, Ogg et al., 1999, Schenck, 2003, Schenck and Zimmerman, 2004, Zhou et al., 2001) have been discussed. However, none of those factors can fully explain the observed in vivo phase difference between GM and WM. For example, the maximum estimated frequency shift from blood deoxyhemoglobin is around 2 Hz (Duyn et al., 2007) and is much smaller than the observed 6-Hz in vivo GM/WM phase separation. Therefore, either a detailed model integrating all potential sources of susceptibility contrast is needed to verify that susceptibility alone can explain the observed data or an alternate contrast generating mechanism is needed. In this work, we propose a novel source of frequency differences, which can plausibly explain the observed GM/WM phase contrast, namely the microscopic chemical exchange processes between free water and macromolecules. It is well known from magnetization transfer studies (Balaban and Ceckler, 1992, Bryant, 1996, Gochberg and Gore, 2007, Gochberg et al., 1998, Henkelman et al., 2001, Kennan et al., 1996, Koenig and Brown, 1993, Liepinsh and Otting, 1996, Sled et al., 2004, van Zijl et al., 2003) as well as from water–bovine serum albumin (BSA) measurements (Gallier et al., 1987, Hills et al., 1989, Olechnowicz et al., 1999) that the spin relaxation times T1 and T2 are closely related to the macromolecule concentration via exchange between bulk tissue water and the hydrophilic groups on the surface of macromolecules. Such exchange processes are typically fast (on the order of 10 10 s) but can result in a small frequency shift of the water resonance due to the different chemical environment between free water and the hydrophilic groups on the macromolecule surface. Since the exchange is not rate limited (Koenig and Brown, 1993, Liepinsh and Otting, 1996), the frequency shift will mainly depend on the macromolecule concentration and the macromolecule type. Therefore, a phase difference between GM and WM is to be expected due to their different macromolecule content (Mader et al., 2002b, McLean and Barker, 2006, O'Brien and Sampson, 1965). Surprisingly, such a direct frequency effect has not yet been assessed in water–macromolecule exchange studies despite the overwhelming number of publications in this field. The lack of such a direct frequency measurement might be due to the fact that most exchange studies focus on the water relaxation properties. In addition, most MR scanners are not equipped with an internal frequency lock to differentiate any minute frequency shift that might arise from water–macromolecule exchange, and often the water resonance frequency was used as a reference (4.73 ppm).

In this study, we first determined the in vivo GM/WM phase contrast dependency on field strength and echo time (TE) to verify that the observed phase contrast is a pure frequency effect that could originate either from susceptibility contrast or from a microscopic water–macromolecule exchange model as proposed in this work. Subsequently, we performed direct water frequency shift measurements with different macromolecule concentrations in order to assess the interaction strength of water–macromolecule exchange and compared this to the in vivo situation. The water–macromolecule exchange as a major source of contribution to in vivo GM/WM phase contrast is discussed.

Section snippets

MRI data acquisition

Experiments were carried out on three scanners: Siemens Sonata (1.5 T), Trio (3 T) and MAGNETOM 7 T. Two volunteers were scanned at all field strengths. Informed written consent was provided prior to each scan following the guidelines of the institutional review board. Eight-channel head coils were used to acquire all images at different field strengths. A modified echo-shifted 3D FLASH sequence (Chung and Duerk, 1999) was used to acquire the phase maps. Image matrix size was 208 × 256 × 80 with

Results and discussion

The in vivo data presented are from a single subject although both subjects gave similar results. Data averaging was not performed in this study since the local field effects depend on tissue geometry and coregistration induced errors are difficult to avoid.

Acknowledgments

We thank Dr. Michael Mueller, Dr. Constantin von zur Muehlen, Dr. Dominik Paul, Dr. Ute Ludwig, Anja Kurutsch, Volker Brecht, and Irene Neuendorfer for their assistance with the experiments, and Dr. Valerij Kiselev for the helpful discussion. This study is supported by grant #13N9208 in the project INUMAC supported by the BMBF (Federal Ministry for Education and Research, Germany).

References (39)

  • AshtM.M. et al.

    Biological correlates of diffusivity in brain abscess

    Magn. Reson. Med.

    (2005)
  • BalabanR.S. et al.

    Magnetization transfer contrast in magnetic resonance imaging

    Magn. Reson. Q.

    (1992)
  • BeharK.L. et al.

    Characterization of macromolecule resonances in the 1H NMR spectrum of rat brain

    Magn. Reson. Med.

    (1993)
  • BeharK.L. et al.

    Analysis of macromolecule resonances in 1H NMR spectra of human brain

    Magn. Reson. Med.

    (1994)
  • BizziA. et al.

    Role of iron and ferritin in MR imaging of the brain: a study in primates at different field strengths

    Radiology

    (1990)
  • BryantR.G.

    The dynamics of water–protein interactions

    Annu. Rev. Biophys. Biomol. Struct.

    (1996)
  • ChungY.C. et al.

    Signal formation in echo-shifted sequences

    Magn. Reson. Med.

    (1999)
  • DrayerB. et al.

    MRI of brain iron

    AJR. Am. J. Roentgenol.

    (1986)
  • DuynJ.H. et al.

    High-field MRI of brain cortical substructure based on signal phase

    Proc. Natl. Acad. Sci. U.S.A.

    (2007)
  • Cited by (113)

    • Towards in vivo ground truth susceptibility for single-orientation deep learning QSM: A multi-orientation gradient-echo MRI dataset

      2022, NeuroImage
      Citation Excerpt :

      Another limitation of our dataset is that both STI and COSMOS maps were calculated based on the assumption that the frequency shifts entirely originate from bulk susceptibility. However, evidence indicates that frequency shifts can be induced not only by tissue susceptibility but also by non-susceptibility effects such as chemical shift, chemical exchange, white matter microstructure, and tissue compartmentalization (Kuchel et al., 2003; Leutritz et al., 2013; Luo et al., 2010; Shmueli et al., 2011; Shmueli et al., 2017; Wharton and Bowtell, 2012, 2015; Zhong et al., 2008). Therefore, a perfect gold standard susceptibility map for single-orientation QSM reconstruction remains to be further explored, which may depend on a better understanding of MRI phase contrast.

    • Influence of water based embedding media composition on the relaxation properties of fixed tissue

      2020, Magnetic Resonance Imaging
      Citation Excerpt :

      On the other hand, with the increase of CE rate, the water line exhibits both line broadening and position shift. Since the carrier frequency is set locally with respect to the water signal maximum, this shift is normally not noticed in MRI but contributes to contrast in phase contrast imaging [56,57]. Of more interest is the water line width which is directly related to the measured T2 (see Fig. S1).

    • Larmor frequency in heterogeneous media

      2019, Journal of Magnetic Resonance
    • Pros and cons of ultra-high-field MRI/MRS for human application

      2018, Progress in Nuclear Magnetic Resonance Spectroscopy
    View all citing articles on Scopus
    View full text