Ken Kwong recalls the early days of fMRI

Gary Boas
November 28, 2014

It was early 1991 and the NMR Center at the Massachusetts General Hospital—now the Martinos Center for Biomedical Imaging—was flush with excitement. Jack Belliveau had recently performed his pioneering study with functional MRI, measuring regional brain activity with the technique using external contrast agents [1], and the possibilities of the technique seemed truly limitless.

Researchers throughout the Center were particularly inspired by the potential for brain mapping that that was evident in Belliveau’s work. They could now see, more or less in real time, changes in the brain occurring in response to particular stimuli or tasks. Kenneth Kwong wanted to take it a step further, though. The need for external contrast limited the potential of fMRI in human subjects—any medically unnecessary injection poses some degree of risk—so Kwong, then a postdoctoral fellow in the Center, set out to find a way to achieve intrinsic MR contrast.

In this, he said, he was influenced by several earlier studies. First was Keith Thulborn’s 1982 report showing the impact of deoxyhemoglobin on the MR parameter T2 [2]. (Kwong wasn’t yet aware of the results on blood oxygen level dependence (BOLD) contrast published in 1990 by Seiji Ogawa of Bell Laboratories and colleagues at the University of Minnesota [3].) Thulborn’s observations about deoxyhemoglobin and magnetic susceptibility suggested to Kwong that “gradient echo” imaging, which was known to be highly sensitive to the susceptibility contrast, might offer a means to identify changes in blood signal following sensory and cognitive stimulation.

Another piece of the puzzle: a poster presented by Robert Turner, Denis Le Bihan, Chrit Moonen and Joe Frank at the April 1991 Society of Magnetic Resonance Imaging (SMRI) meeting in Chicago. Here, the researchers showed an association between perfusion and an overshoot in the deoxyhemoglobin signal in a hypoxia experiment with a cat (see [4]). This was a significant insight, Kwong said, one that would play an important role in his interpretation of the rising MR signal response in his own experiments.

Those experiments began on the evening of May 9, in what is now Bay 3 at the Martinos Center. It’s tempting to imagine here a sense of import in the air, a knowing understanding of the significance of what was about to happen. The fact is, though, for Kwong, it was a scanning session like any other. He doesn’t recall if David Kennedy was at the console with him, but he remembers learning from Kennedy that day the anatomical location of the visual cortex of the brain. Brigitte Poncelet may also have been there; he was running her time course subtraction routine for image data analysis immediately after collecting the EPI time series.

For the experiment itself, Kwong borrowed the pair of visual stimulation goggles that Jack Belliveau had used for his early fMRI experiments—the goggles on loan from Peter Fox, who had used them in his seminal stimulation rate experiments in the early 1980s [5]. He ran a T2* weighted gradient echo EPI sequence as well as a T1 weighted inversion recovery spin echo EPI sequence while a volunteer watched the flickering stimulus in the goggles. He decided to use a block design paradigm for the latter, alternating a baseline OFF epoch with an ON epoch for a total of 70 time points.

This is significant, said Bruce Rosen, the director of the Martinos Center. While the block design seems obvious now, at the time all functional brain imaging studies were performed with single time-point injections, with analysis tools designed to identify the differences between the images acquired many minutes apart.

“Ken’s block design represented a new paradigm for activation studies, requiring different timing and analysis techniques, and remains the most commonly used functional paradigm more than 20 years later,” Rosen explained in a 2012 lecture and paper, “fMRI at 20: Has it changed the world?” [6]. The design can be seen on the left of the two pages reproduced here from Kwong’s laboratory notebook from the day of the experiment.

After a quick analysis of the data, Kwong noted a clear difference in signal due to changes in blood oxygenation in the T2* images as well as blood flow-related changes in the T1 data. He also observed that oxygen delivery, blood flow and blood volume increased during neuronal activation, while the total paramagnetic deoxyhemoglobin content decreased. Taken together, he said, this suggested that he had successfully imaged cerebral activation based on endogenous contrast.

He understood the potential significance of the results—he remembers joking with colleagues about having just demonstrated cold fusion, by then an almost mythological research goal. Still, he kept his enthusiasm in check and focused on whether the signal differences he had seen were artifacts. This question occupied him for the next couple of months as he ran further experiments and analyzed the results. Finally, he said, he was confident that they were in fact due to visual activation.

Publishing the findings; or, the third time's the charm

The next step for Kwong was of course to report the findings. This would seem to be the easy part. But as many can attest, getting the word out isn’t always as simple as it sounds.

Kwong planned to submit an abstract describing “work in progress” movies of brain activation to the 10th annual meeting of the Society for Magnetic Resonance in Medicine (SMRM), to be held in San Francisco in August 1991. As was the custom in those pre-online submission days, he found himself racing to the FedEx office minutes before the midnight deadline. He made it. Somehow, though, tragically, the package was lost in the mail. This left the announcement of his groundbreaking findings to a mention by Bernice Hoppel during a paper presentation and a short video in a plenary lecture by Thomas Brady, the founding director of the NMR Center.

Kwong naturally would have preferred to present his findings in full, but even this brief glimpse created quite a stir. Brady had shown for the very first time in public cortex activation in response to photic stimulation. “It was certainly not Hollywood quality,” Rosen said of the video, “but many people in the audience appreciated its potential immediately; and for some, it changed the course of their careers” [6].

Among them: Peter Bandettini, then a graduate student at the Medical College of Wisconsin. Bandettini recalls seeing a “jaw-dropping movie” in San Francisco and then returning home and initiating similar experiments using Kwong’s method to measure changes in brain activity [7]. He would report his findings the following June, in his Magnetic Resonance in Medicine paper, “Time course EPI of human brain function during task activation” [8].

Meanwhile, Kwong and colleagues wrote a “broad and comprehensive” paper reporting the fMRI phenomenon. They submitted it to Nature in October of 1991. A few months later, it was rejected. Here, one of the reviewers commented: “If the point of this paper is that MRI can be used to map the brain, this point has been made in the Science paper [by Belliveau et al.]. If the point of this paper is that MRI can shed new light on the regulation of cerebral hemodynamics and metabolism by neural activity, I am not yet convinced.”

The authors were disappointed, to be sure. The reviewer seemed to have missed one of the most important advances of the study: dynamic mapping of the brain using only endogenous contrast. Nevertheless, they soldiered on. They expanded the scope and depth of the report and submitted it to the Proceedings of the National Academy of Sciences (PNAS) in 1992. As the saying goes, the third time’s the charm. This paper, “Dynamic magnetic resonance imaging of human brain activity during primary sensory stimulation,” was accepted and published in June [9].

With these experiments, Kwong and colleagues demonstrated that imaging of cerebral activation was possible using only endogenous contrast, that they could observe changes in the brain following sensory stimulation without having to inject the subject with any kind of external agent. It remained now to explore more fully the potential of the technique, to discover what about the complexities of the brain and the human body as a whole could be learned in applying it. As it turned out, investigators around the world were eager to do just that.


  1. Belliveau, J. W., D. N. Kennedy, Jr., 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(5032): 716-719.
  2. Thulborn, K. R., J. C. Waterton, P. M. Matthews and G. K. Radda (1982). "Oxygenation dependence of the transverse relaxation time of water protons in whole blood at high field." Biochim Biophys Acta 714(2): 265-270.
  3. Ogawa, S., T. M. Lee, A. R. Kay and D. W. Tank (1990). "Brain magnetic resonance imaging with contrast dependent on blood oxygenation." Proc Natl Acad Sci U S A 87(24): 9868-9872.
  4. Turner, R., D. Le Bihan, C. T. Moonen, D. Despres and J. Frank (1991). "Echo-planar time course MRI of cat brain oxygenation changes." Magn Reson Med 22(1): 159-166.
  5. Fox, P. T., M. A. Mintun, M. E. Raichle and P. Herscovitch (1984). "A noninvasive approach to quantitative functional brain mapping with H2 (15)O and positron emission tomography." J Cereb Blood Flow Metab 4(3): 329-333.
  6. Rosen, B. R. and R. L. Savoy (2012). "fMRI at 20: has it changed the world?" Neuroimage 62(2): 1316-1324.
  7. Bandettini, P. A. (2012). "Twenty years of functional MRI: the science and the stories." Neuroimage 62(2): 575-588.
  8. Bandettini, P. A., E. C. Wong, R. S. Hinks, R. S. Tikofsky and J. S. Hyde (1992). "Time course EPI of human brain function during task activation." Magn Reson Med 25(2): 390-397.
  9. Kwong, K. K., J. W. Belliveau, D. A. Chesler, I. E. Goldberg, R. M. Weisskoff, B. P. Poncelet, D. N. Kennedy, B. E. Hoppel, M. S. Cohen, R. Turner and et al. (1992). "Dynamic magnetic resonance imaging of human brain activity during primary sensory stimulation." Proc Natl Acad Sci U S A 89(12): 5675-5679.