First Functional NMR Imaging of the Human Brain

The early success with what we now know as fMRI was the result of a confluence of factors at the Massachusetts General Hospital, including considerable work with susceptibility contrast and the availability of clinical echo planar imaging (EPI). Just as important, said Kenneth Kwong, one of the originators of the technique, was the “free-wheeling and fertile intellectual environment” structured and encouraged by Thomas Brady and Bruce Rosen at the MGH NMR Center (now the Martinos Center for Biomedical Imaging). The MGH NMR Center was established in 1989 with Brady as founding Director. Back then, of course, the Center was considerably smaller than it is now. When it opened, it had one MRI scanner for human studies, three small animal scanners, and an MR spectroscopy system. The remaining space was undesignated and contained a large open space for students and fellows. Rosen recalls walking into this space for the first time and being astonished to see that it was equipped with 35–40 desks. At the time, the NMR research group consisted of a mere handful of staff, including: Rosen; Van Wedeen, MD; Jerome Ackerman, PhD; Mark S. Cohen, PhD; Randall Lauffer, PhD; E. Douglas Lewandowski, PhD; Robert Weisskoff, PhD; Keith R. Thulborn, PhD; Kenneth K. Kwong, PhD; and R. Gilberto Gonzalez, MD PhD. The space seemed far larger than they needed but Brady was confident that they would soon fill it. Later the same year, Advanced NMR Systems, of Wilmington, Mass., with whom Brady and Rosen already had a strong relationship, installed the Instascan, the first-ever human MRI scanner with high-speed echoplanar imaging capability. This scanner boasted beat-to-beat imaging, which the researchers believed would improve its ability to perform MR angiography. Without the MGH administration’s willingness to place a big bet on the department of radiology, the echo-planar system, in all probability, would have gone elsewhere, foreclosing a major opportunity to the institution. It was against this backdrop that Jack Belliveau, a graduate student in the MGH NMR Research Program, performed the first ever functional MRI (fMRI) experiments. MOVE TO GETTING THE WORD OUT SECTION? His results were first published in his Harvard University PhD thesis in 1990(1) and subsequently in Science, which displayed his 3D image of activation of the visual cortex of a human volunteer in its issue of November 1, 1991 (2). What did he set out to do? Belliveau was aware that experiments dating back to the nineteenth century (3) had established that local alterations in neuronal activity occur during task performance and induce local changes in perfusion. He also knew that the hemodynamic state in the non-activated brain is stable over time and that it was possible to use subtraction methods to demonstrate changes in perfusion due to brain activation. Moreover, a number of imaging techniques had already been used to measure changes in cerebral blood flow, including SPECT, CT, and PET (4). For example, Peter Fox and his colleagues in Washington University, St Louis measured increases in blood flow of 30-50% due to activation of the visual cortex, using PET and [15O]-H¬2O to measure blood flow (5). Belliveau realized that MRI would have a number of advantages over other imaging techniques, most of which required the use of radioisotopes in human subjects, thus limiting their applicability. MRI had a much higher spatial resolution than PET or SPECT, had high soft-tissue contrast that revealed anatomical detail, and could take advantage of a readily available isotopically stable contrast agent – gadolinium diethylenetriamine pentaacetic acid (GdDTPA) – that had recently been approved for clinical use. The MGH also had a singular advantage in that it possessed the Instascan, the first scanner in the world capable of “ultra-fast” imaging of human subjects, acquiring 2D images in 50 ms (6). This made it fast enough to use dynamic susceptibility contrast imaging to image changes in signal as intravenously injected GdDTPA made its first pass through the cerebral vasculature. Using this ultra-fast method, Belliveau and his colleagues demonstrated that injection of injection of GdDTPA or DyDTPA leads to a significant (>50%), transient decrease in T2 signal intensity when using pulse sequences with long echo times – changes much greater than those due to T-1 relaxivity changes. They demonstrated that the alterations in signal intensity versus time were consistent with those detected by dynamic CT imaging and that they were proportional to blood volume and blood flow (7). Belliveau predicted that the ability to temporally resolve functional interactions in the brain would permit detection and analysis of distributed neural networks and lead to greater understanding of structure-function correlates of neural information processing (7). To test this hypothesis, he used ultra-fast dynamic susceptibility imaging to measure changes in perfusion during stimulation of the human visual cortex. Volunteers wore light-proof Grass goggles kindly provided by Fox – the same goggles he had used in his earlier PET experiments. First, the volunteers were given a bolus of Gd-DTPA and the first set of 60 images were acquired in 45 seconds in darkness. Then, a second bolus was given and the experiment repeated while the volunteers watched a flashing checkerboard pattern. On a voxel-by-voxel basis, Belliveau used subtraction to show differences in blood volume (averaging 32±10%) during periods of rest and activation in the region of the visual cortex. In presenting his data, he used a color overlay of functional changes over the anatomical images, and three-dimensional reconstruction techniques to show the distribution of activation (2). His method of using color coded difference images overlying anatomical maps remains the standard to this day. [REFERENCES]