Copyright © 1999 Massachusetts Institute of Technology
Magnetic resonance imaging (MRI) is based on the phenomenon of nuclear magnetic resonance (NMR), first described in landmark papers over fifty years ago (Rabi et al. 1938; Rabi, Millman, and Kusch 1939; Purcell et al. 1945; Bloch, Hansen, and Packard 1946). In the presence of an external magnetic field, atomic nuclei with magnetic moments, such as 1H, 13C, and 31P nuclei, encounter a separation in the energy levels of their quantum mechanically allowed orientations relative to the external field. Transitions between these orientations can be induced with electromagnetic radiation typically in the radiofrequency range. The discrete frequency associated with such a transition is proportional to the external magnetic field strength and to two parameters that are determined by the intrinsic properties of the nucleus and its chemical environments, the nuclear gyromagnetic ratio, and the chemical shift, respectively. Based on the ability to obtain discrete resonances sensitive to the chemical environment, NMR has evolved rapidly to become an indispensable tool in chemical and biological research focused on molecular composition, structure, and dynamics.
In 1973, a novel concept of using NMR of the hydrogen atoms in the human body as an imaging modality was introduced (Lauterbur 1973). While all NMR applications of the time were singularly concerned with eliminating inhomogeneities in magnetic field magnitude over the sample, this new concept embraced them and proposed to utilize them to extract spatial information. Consequently, today magnetic resonance is solidly established as a noninvasive imaging technique suitable for use with humans.
MRI is essentially based on two fundamental ideas, "spatial encoding" and "contrast." The former is the means by which the NMR data contain information on the spatial origin of the NMR signal. The latter must provide the ability to distinguish and visualize different structures or processes occurring within the imaged object and is ultimately translated into gray scale or color coding for presentation. Spatial encoding is accomplished by external magnetic fields whose magnitude depend linearly on spatial coordinates in any one of three orthogonal directions. Contrast is achieved either based on the regional density of the nuclear species imaged or by the impact of the chemical and biological environment on parameters that determine the behavior or relaxation of a population of spins from a nonequilibrium state toward thermal equilibrium. Improving contrast and spatial encoding strategies are central to developments in MRI as the field strives to image faster, with higher resolution and structural detail, and image not only anatomy but also physiological processes such as blood flow, perfusion, organ function, and intracellular chemistry.
An avidly pursued new dimension in the acquisition of physiological and biochemical information with MRI is mapping human brain function, referred to as fMRI. The first fMRI image of the human brain was based on measurements of task-induced blood volume change assessed with intravenous bolus injection of an MRI contrast agent, a highly paramagnetic substance, into the human subject and tracking the bolus passage through the brain with consecutive, rapidly acquired images (Belliveau et al. 1991). However, this method was quickly rendered obsolete with the introduction of totally noninvasive methods of fMRI. Of the two current ways of mapping alterations in neuronal activation noninvasively, the most commonly used method relies on the weak magnetic interactions between the nuclear spins of water protons in tissue and blood, and the paramagnetic deoxyhemoglobin molecule, termed BOLD (blood oxygen level - dependent) contrast, first described for the brain by Ogawa (Ogawa et al. 1990a, 1990b; Ogawa and Lee 1990) and is similar to the effect described for blood alone by Thulborn et al. (1982). The presence of paramagnetic deoxyhemoglobin, compartmentalized in red blood cells and in blood vessels, generates local magnetic field inhomogeneities surrounding these compartments which are dynamically (due to rapid diffusion) or statically averaged over the smallest volume element in the image and lead to signal loss when a delay is introduced between signal excitation and subsequent sampling.
In the original papers describing the BOLD effect, functional mapping in the human brain using BOLD was anticipated (Ogawa et al. 1990a) based on data documenting regional elevation in blood flow and glucose metabolism without a commensurate increase in oxygen consumption rate during increased neuronal activity (Fox and Raichle 1985; Fox et al. 1988); these data would predict a task-induced decrease in deoxyhemoglobin content in the human brain and a consequent alteration in MRI signal intensity when the signal intensity difference between two states, for example, in the absence and presence of a mental task or sensory stimulation, is examined. This was demonstrated and the first BOLD-based fMRI images of the human brain were published in 1992 by three groups in papers submitted within five days of each other (Bandettini et al. 1992; Kwong et al. 1992; Ogawa et al. 1992). Initial functional brain mapping studies were focused on simple sensory stimulation and regions of the brain that are relatively well understood. These studies were aimed at demonstrating and evaluating the validity of the technique rather than addressing the plethora of as yet unanswered questions concerned with aspects of brain function. In the short period of time since its introduction, however, BOLD fMRI has been used to map functions in the whole brain, including subcortical nuclei, with a few millimeters resolution and has been shown to display specificity at the level of ocular dominance columns in humans at the high magnetic field of 4 tesla (Menon, Ogawa, and Ugurbil 1996; Menon et al. 1997). At present field strengths, the sensitivity and hence the spatial resolution attainable, however, is at the margin of what is required for visualizing human ocular dominance columns which are approximately 1 x 1 mm in cross-sectional dimensions.
A second approach to generating functional maps of the brain with fMRI relies on the task-induced increase in regional blood flow alone. This method is analogous to the POSITRON EMISSION TOMOGRAPHY (PET) - based functional brain mapping using water labeled with a positron emitter (H215O). In the noninvasive MRI approach, however, the label is simply the collective spins of the water molecules whose net bulk magnetization is inverted or nulled (saturated) either within a slice to be imaged or outside of the slice to be imaged. For example, if the slice to be imaged is inverted, the inverted magnetization must relax back to its thermal equilibrium value and does so in a few seconds; in the absence of flow, this occurs with what is termed spin- lattice relaxation mechanisms. However, when flow is present, apparent relaxation occurs because of replacement of inverted spins by unperturbed spins coming from outside the inversion slice. Such flow-based fMRI methods were first demonstrated in 1992 (Kwong et al. 1992), and significantly refined subsequently (Edelman et al. 1992; Kim 1995). While flow-based techniques have some advantages over BOLD methods, such as simplicity of interpretation and ability to poise the sensitivity to perfusion as opposed to macrovascular flow, rapid imaging of large sections of the brain or whole brain is not yet possible.
fMRI techniques rely on secondary and tertiary responses, metabolic and hemodynamic, to increased neuronal activity. Hence, they are subject to limitations imposed by the temporal characteristics and spatial specificity of these responses. Current data suggest that BOLD images, when designed with appropriate paradigms, may have spatial specificity down to the millimeter to submillimiter scale (e.g., ocular dominance columns) presumably because the spatial extent of altered oxygen consumption, hence deoxyhemoglobin alterations, coupled to neuronal activity, is confined accurately to the region of elevated neuronal activity; this scale may be coarser, possibly in the range of several millimeters, for perfusion images if blood flow response extends beyond the region of increased activity (Malonek and Grinvald 1996). With respect to temporal resolution, the sluggish metabolic response and even more sluggish hemodynamic response to changes in neuronal activity suggest that better than approximately 0.5-sec time resolution may not be achievable with current fMRI techniques even though image acquisition can be accomplished in as little as 20 to 30 msec. While this excludes a very large temporal domain of interest, the plethora of mental processes accomplished in the seconds domain by the human brain remains accessible to fMRI.
Bandettini, A., E. C. Wang, R. S. Hinks, R. S. Rikofsky, and J. S. Hyde. (1992). Time course EPI of human brain function during task activation. Magn. Reson. Med. 25: 390.
Belliveau, J. W., D. N. Kennedy, R. C. McKinstry, B. R. Buchbinder, R. M. Weisskoff, M. S. Cohen, J. M. Vevea, T. J. Brady, and B. R. Rosen. (1991). Functional mapping of the human visual cortex by magnetic resonance imaging. Science 254:716-719.
Bloch, F., W. W. Hansen, and M. Packard. (1946). The nuclear induction experiment. Physical Review 70:474-485.
Edelman, R. E., B. Siewer, D. G. Darby, V. Thangaraj, A. C. Nobre, M. M. Mesulam, and S. Warach. (1992). Qualitative mapping of cerebral blood flow and functional localization with echo-planar MR imaging and signal targeting with alternating radio frequency. Radiology 192:513-520.
Fox, P. T., and M. E. Raichle. (1985). Stimulus rate determines regional brain blood flow in striate cortex. Ann. Neurol. 17:303-305.
Fox, P. T., M. E. Raichle, M. A. Mintun, and C. Dence. (1988). Nonoxidative glucose consumption during focal physiologic neural activity. Science 241:462-464.
Kim, S.-G. (1995). Quantification of relative cerebral blood flow change by flow-sensitive alternating inversion recovery (FAIR) technique: Application to functional mapping. Magn. Reson. Med. 34:293-301.
Kwong, K. K., J. W. Belliveau, D. A. Chesler, I. E. Goldberg, R. M. Weisskoff, B. Poncelet, D. N. Kennedy, B. E. Hoppel, M. S. Cohen, R. Turner, H. -M. Cheng, T. J. Brady, and B. R. Rosen. (1992). Dynamic magnetic resonance imaging of human brain activity during primary sensory stimulation. Proc. Natl. Acad. Sci. U. S. A. 89:5675-5679.
Lauterbur, P. C. (1973). Image formation by induced local interaction: Examples employing nuclear magnetic resonance. Nature 242:190-191.
Malonek, D., and A. Grinvald. (1996). Interactions between electrical activity and cortical microcirculation revealed by imaging spectroscopy: Implication for functional brain mapping. Science 272:551-554.
Menon, R. S., S. Ogawa, and K. Ugurbil. (1996). Mapping ocular dominance columns in human V1 using fMRI. Neuroimage 3: S357.
Menon, R. S., S. Ogawa, J. P. Strupp, and K. Ugurbil. (1997). Ocular dominance in human V1 demonstrated by functional magnetic resonance imaging. J. Neurophysiol. 77(5):2780-2787.
Ogawa, S., and T.-M. Lee. (1990). Magnetic resonance imaging of blood vessels at high fields: In vivo and in vitro measurements and image simulation. Magn. Reson. Med. 16:9-18.
Ogawa, S., T. -M. Lee, A. R. Kay, and D. W. Tank. (1990a). Brain magnetic resonance imaging with contrast dependent on blood oxygenation. Proc. Natl. Acad. Sci. U. S. A. 87:9868-9872.
Ogawa, S., T. -M. Lee, A. S. Nayak, and P. Glynn. (1990b). Oxygenation-sensitive contrast in magnetic resonance image of rodent brain at high magnetic fields. Magn. Reson. Med. 14:68-78.
Ogawa, S., D. W. Tank, R. Menon, J. M. Ellermann, S. -G. Kim, H. Merkle, and K. Ugurbil. (1992). Intrinsic signal changes accompanying sensory stimulation: Functional brain mapping with magnetic resonance imaging. Proc. Natl. Acad. Sci. U. S. A. 89:5951-5955.
Purcell, E. M., H. C. Torrey, and R. V. Pound. (1945). Resonance absorption by nuclear magnetic moments in a solid. Physical Review 69: 37.
Rabi, I. I., S. Millman, and P. Kusch. (1939). The molecular beam resonance method for measuring nuclear magnetic moments. Physical Review 55:526-535.
Rabi, I. I., J. R. Zacharias, S. Millman, and P. Kusch. (1938). A new method of measuring nuclear magnetic moment. Physical Review 53: 318.
Thulborn, K. R., J. C. Waterton, P. M. Mathews, and G. K. Radda. (1982). Oxygenation dependence of the transverse relaxation time of water protons in whole blood at high field. Biochem. Biophys. Acta 714:265-270 .
Copyright © 1999 Massachusetts Institute of Technology