Paul Lauterbur and the Invention of MRI

by M. Joan Dawson

  Figure 10.5

  Elise Lauterbur, 2005.

  Paul was not feeling well in the second week of March 2007 and, seeing the lab results, his doctor told us once again that time was running out. Paul responded that he had “obeyed all Dr’s. orders except for dying.” We drove from Urbana to Oberlin, Ohio, for the weekend of March 17 to hear our daughter’s senior recital on recorder and Baroque flute, a requirement for her degree from Oberlin Conservatory. His health was good enough for him to enjoy his daughter’s music, to be a very proud papa. The family gathering was cheering to him. We reached home the following Tuesday afternoon and Paul told me, perhaps for the first time ever, that he didn’t feel like going in to the lab. Since he had told Nasrin, his secretary, we might still be traveling that day, it might not be remiss, he explained, if he worked at home for a day. He spent the rest of the day reading literature on regulatory genomics. The next day he could not rise from the bed. As he became weaker during the next few days, we realized that his light was dying, and this was truly the end. “Damn, damn, damn!” he said. Family gathered, and former students and colleagues were informed.

  Zhi-Pei visited us on Sunday and remarked to the press that right up until his final days, Lauterbur was graceful, kind, and caring. On his deathbed, with all of us feeling tragic, Paul asked Zhi-Pei the most normal thing in the world, the thing he would always ask a former student in greeting: “So, how’s your research coming along?” Paul and I talked of the wonderful times we had had together, and when he could no longer talk we held hands.

  One week after his last working day Paul went, ungentle, “into that good night.”11 The last thing I said to him was, “I love you.” The last thing he said to me was, “I love you too.”

  Figure 10.6

  Last portrait of Paul.

  Epilogue

  Paul himself summarized his life as an experimental scientist, in reminiscences about his childhood laboratory in the basement of his parents’ home on the occasion of the Kyoto Prize ceremony:

  Many strange results and invaluable experience with laboratory work, but no real discoveries, came from this work, although I still do not understand the results of some of those experiments of almost 50 years ago. And I suspect if I were to repeat them today, with 50 years more knowledge, that I would not understand them still. Because any of you who have been involved with experimental science know that experiments that you readily understand and interpret are a minority, that one is fortunate to be able to design and to carry out and to interpret. Much of the work of science, like much of the work of filmmaking [an allusion to Akira Kurosawa, who received the Kyoto Prize in Arts and Moral Sciences that day], for example, goes on behind the scenes and is never seen by those who look at the final product and admire its form and shape and ingenuity, but never know all of the false starts and the horrible errors and the confusions that went into the creation of that final product.1

  Paul’s intelligence was complex, and gathered in various and disparate learning and powers to make up and test satisfying scientific hypotheses. His study of the origins of life was perhaps the apotheosis, but his earlier work on NMR and MRI are of the same mark. I, and many of his other colleagues, witnessed with awe the wide breadth of his intellectual curiosity. By the time I met him, Paul had made himself into a formidable physiologist in order to better understand the problems that MRI could solve. Paul’s colleagues in other fields have made similar observations about Paul’s grasp of their disciplines as well. These areas of expertise covered a vast ground: all of chemistry, of course; all areas of biology except, perhaps genomics; electrical engineering; computation; biophysics and bioengineering; and diagnostic medicine and neuroscience, to name a few. He also had an excellent grasp of astronomy, natural history, space flight, archaeology, and philosophy. The philosophy of science was a particular interest. Sometimes he seemed a force of nature. Sometimes I resented it. Sometimes I loved him for it.

  Paul always found it useful “to look at problems upside down,” or “to see the problem as its own solution.”2 He looked for solutions in places where others saw only problems. An example is the magnetic field inhomogeneity that has bedeviled real working magnets since the inception of NMR in chemistry. Paul realized that the key to imaging was to deliberately make the main static magnetic field inhomogeneous so that the signal of an atom would depend in a known way on its spatial position. This insight was available since the time NMR was first understood. But, as Paul said about his invention of MRI, “It didn’t occur to people that there was anything to think about.”

  Another example of Paul’s mental clarity is in his thinking about the origins of life. The result of the famous Miller-Urey experiment is not only the formation of simple organic compounds but also a gooey mess on the side of the flask in which the experiment is done. Paul wondered about the gooey mess that everyone had been throwing away; what answers lay in this mud? Might not this mess hold the key to building biological polymers? From this wondering came his musings about the role of surface catalysis in life’s origins, which led naturally to his study of molecular imprints. Once that creative step is conceived, of how self-reproducing polymers can be created, the rest of a theory of the creation of “protolife” becomes quite straightforward. Creationists like to say that a natural step from physics and chemistry to life is “inconceivable,” and even scientists often stay away from making such a leap, allowing only that “here a miracle happens.” But Paul conceived of simple known steps without miracles or divine intervention.

  Bob Shulman, a well-known NMR biochemist, describes Paul as having “a great rugged dependence upon insight and its accompanying vision [that] are in the heroic mode.”3 A staff writer of the Economist was describing the same thing when he complained that Paul’s science was “a little wild.” Paul cracked the door open now and then to the profundity of existence. The genius of Paul, maybe all genius, was to access the most ancient and shrewd parts of his being from time to time, and to bring those parts to bear on his intellectual pursuits. Yes, Paul’s great science did have a certain wildness, a well-regulated wildness that cannot be faked. Paul was always very clear-headed and methodical about his wild science, and everything he did fit coherently with the rest of the work. Other people simply couldn’t see all of those connections.

  But all of this sounds finally too adult to fit Paul. To paraphrase Arthur C. Clarke, Paul never stopped growing and he never grew up. He valued play and kept going back to his childish pastime of doing creative science. In Paul there was a generosity of spirit that needed childishness at its core. Grown-ups with a strong mental focus and a sense of purpose don’t venture far beyond the known; they are unlikely to make it through the maze of the unknown. Paul always picked big problems to work with, from his childhood interest in why there is no silicon-based biology through his expansion of NMR techniques beyond the proton, to MRI, and on to the origins of life. Sometimes this generosity, this childishness, seemed to others to be mad. But Paul was too complicated to “settle on any given madness.”4

  A corollary was courage. Most of us don’t have the courage to tackle humanity’s biggest questions. Paul’s comment was, “You can spend 18 hours a day running a hot dog stand or running GM.” He also said, “Perhaps an early fascination with the limitless hubris of Eddington and Bertrand Russell deserves some credit, or blame, for later elements of style, although there was probably a pre-disposition to infection.”5 With his childish forgetfulness and indifference to what sensible people think, Paul continued his foolish pastime, the making of real, honest, and intellectually significant science.

  Paul worked routinely in his office and lab seven days a week, a source of immense enjoyment and personal satisfaction. He quoted Percy Bridgman’s writings, “Science is doing one’s damnedest with one’s mind, no holds barred.”6 Also for Paul, science was a safe refuge, as Allegra Goodman presciently said, “as if hiding behind old tapestries.”7 The only times he did not follow a daily working schedule w
ere Christmas, Easter, and New Year’s Day. He was rarely ill before being struck by chronic disease, and even then he kept on working, on one occasion taking a manuscript into an operating room. When he broke his leg and was confined to bed, he held staff meetings in our bedroom. I had to clear off his working papers in the evening to claim my space in bed. He continued to have creative new ideas, and continued to dream of making yet another contribution to science and society. He explained to me once his ideas of a new kind of chemistry, a chemistry not of molecules and atoms but of what is left behind when a molecule is departed. It is a memory, with no mass, no atoms of its own, but various chemical properties, and it can vanish without a trace. His died too soon to finish his chemistry of departed ghosts—or to have it sufficiently started that others could take it up.

  Paul’s was a stubborn honesty, a trait he shared with other great scientists I have known. It protected him from scientific fads and dogma, and thus from the prerequisites of the NIH and NSF. To understand the nature of nature you must approach it with an open mind, seeking the truth, and not evidence to support hypotheses. His honesty and integrity were both his strength and his weakness. I have never known Paul to lie, though he did like to keep his chocolates secret from me. Also, as he got older he understood that his health was deteriorating and death was stalking, but once having acknowledged this he refused to think about it again. The classic pop psychology term “in denial” may be applicable, but if so, it was freely chosen. Still, through it all, Paul was no magician; he was a demonstrable member of the human race, a perfect husband and father for all his imperfections.

  There was something allegorical about it all: the boy from little Sidney, Ohio, who challenged the greatness of the universe.

  Appendix A: The Notebook, September 1971

  Page 1

  Spatially Resolved Nuclear Magnetic Resonance Experiments

  The distribution of magnetic nuclei, such as protons, and their relaxation times and diffusion coefficients, may be obtained by imposing magnetic field gradients (ideally, a complete set of orthogonal spherical harmonies) on a sample, such as an organism or a manufactured object, and measuring the intensities and relaxation behavior of the resonances as functions of the applied magnetic field. Additional spatial discrimination may be achieved by the application of time-dependent gradient patterns so as to distinguish,

  (Signature of Paul C. Lauterbur, Sept. 2, 1971)

  (Signature of Donald Vickers, Sept. 3, 1971)

  Page 2

  for example, protons that lie at the intersection of the zero-field (relative to the main static field) lines of three linear gradients.

  The experiments proposed above can be done most conveniently and accurately by measurements of the Fourier transforms of the pulse response of the system. They should be capable of providing a detailed three-dimensional map of the distributions of particular classes of nuclei (classified by nuclear species and relaxation times) within a living organism. For example, the distribution of mobile protons in

  (Signature of Paul C. Lauterbur, Sept. 2, 1971)

  (Signature of Donald Vickers, Sept. 3, 1971)

  Page 3

  tissues, and the differences in relaxation times that appear to be characteristic of malignant tumors [R. Damadian, Science, 171, 1151 (1971)], should be measurable in an intact organism.

  (Signature of Paul C. Lauterbur, Sept. 2, 1971)

  (Signature of Donald Vickers, Sept. 3, 1971)

  Page 4

  Classification of Spatially Resolved Nuclear Magnetic Resonance Experiments

  1. Time-independent methods: based upon the combination of independently measured spectra in a distinct set of field gradients.

  A) Orthogonal gradients: in two dimensions, a pair of orthogonal first-order gradients acting on a set of n elements can produce as many as n2 images. The degree to which an image of the whole array free from significant false elements can be generated by a particular truncated set of gradients is not yet known.

  (Signature of Paul C. Lauterbur, Sept. 6, 1971)

  Page 5

  B) non-orthogonal gradients: examination of simple test cases suggests that additional linear gradients quickly eliminate many false aspects of the two-dimensional image generated by a pair of orthogonal gradients. No vigorous evaluation of the efficiency or limits of such a process has yet been carried out.

  2. Pulsed gradient methods: the rapid application of a sequence of gradients provides a unique history and modulation pattern for each point in the sample volume. Cross-correlation of the resonance response with a function derived

  (Signature of Paul C. Lauterbur, Sept. 6, 1971)

  Page 6

  from the modulation sequences could generate an image.

  Appendix B: Magnetography, October 1971

  Appendix C: Draft Disclosure, August 1972

  Paul C. Lauterbur August 8, 1972

  Magnetography

  Abstract of the Disclosure

  There is disclosed a technique and apparatus for the analysis of matter which depends upon the local fields to which the matter is subjected, whereby it is possible to selectively alter the fields in a known way to locate components or constituents in an object.

  Specification

  This invention relates to those fields of magnetic resonance spectroscopy in which the properties of substances dispose in a magnetic field are investigated by the application of energy in the radio frequ3ncy or microwave frequency range: such as nuclear magnetic resonance (NMR) and electron spin resonance (ESR) respectively (Marginal note: ESR can be radio frequency at low fields.) These techniques and apparatus are described in numerous texts, articles and other publications. In the conventional techniques here to fore practiced the observed signal, for example, the nuclear induction signal in pulsed NMR spectroscopy, represents the resultant of the signals derived from the signals derived from the excited nuclei of the sample under investigation which have been subjected to the applied magnetic and r.f. fields and subsequently detected without regard otherwise to the location or distribution within the sample of the nuclei contributing to the signal response.

  For convenience in expression, the present invention will be described in connection with NMR but it should be understood that it is applicable to other forms of magnetic resonance techniques.

  The present invention provides a technique and means whereby these limitations are overcome so that the derived signals are indicative, for example, of the location and distribution of the excited nuclei and moreover may be utilized to construct the size and shape of objects which are hidden from view, which can not be done with conventional techniques, but the utility of which is apparent in a variety of applications such as medical diagnosis. In this latter sense the invention is somewhat analogous to X-ray and ultrasonic techniques.

  Because the invention provides a means for graphic display of magnetic properties of objects the term “magnetography” has been coined and adopted as the most accurate general description of the technique. (Marginal note: “Magnetograph” has been used as a name for a recording magnetometer. Is that a problem, patentwise?)

  The basic principle of magnetography can best be described if one first understands the basic phenomena of magnetic resonance.

  The nuclei of some isotopes of most elements will give nuclear magnetic resonance (NMR) signals if placed in a magnetic field and exposed to radio-frequency radiation. The frequency at which the phenomenon occurs is directly proportional to the strength of the magnetic field. For example, protons, the nuclei of ordinary hydrogen atoms, give a magnetic resonance signal at 100 MHz (in the FM radio band) in a magnetic field of 23.487 gauss, a field strength readily reached by laboratory electro-magnets or permanent magnets. In a field half as strong, 11,744 gauss, the resonance frequency would be 50 MHz, and in the earth’s field of about 0.5 gauss, resonance occurs at about 2000 Hz. This proportionality between the NMR frequency and the magnetic field provides the basis for the various forms of magnetography a
nd magnetoscopy. If a magnetically homogeneous object is placed in a non-uniform magnetic field, one, for example that decreases linearly with distance across the object, the single resonance ordinarily observed is replaced by a band of resonances, each representing a particular magnetic field and therefore a particular portion of the sample. The intensity of the signal at each frequency is simply proportional to the number of nuclei in the corresponding magnetic field region. Repetition of the experiment in several differently oriented field gradients provides enough information to construct a two- or three-dimensional projection of the shape and interior structure of the object. In three dimensions, the object may be considered to contain planes of constant magnetic field.

  The intensity of the resonance at each frequency is proportional to the number of nuclei in a plane of constant magnetic field. A three-dimensional image may be constructed from spectra with differently-oriented magnetic field gradients.

  Two-dimensional images may be formed more directly from data collected by special pulse sequences. A strip (in two dimensions) or “slice” (in three dimensions) may be selectively excited in one field gradient, and the resonances of the excited nuclei analyzed in another (usually perpendicular) gradient. The procedure may then be repeated by exciting another strip (or slice) and analyzing, to build up a complete two-dimensional projection of the object. A three-dimensional image may then be formed stereoscopically, as mentioned above. (Marginal note: More complex pulse sequences may, in principle be used to obtain three-dimensional data directly.)

  A potentially much more direct method, in which all two-dimensional information is collected in a single spectrum, may be possible if “spin-echo” techniques are used. (Marginal note: Tricky—maybe should be left out.)

  Many of the potential applications of magnetography result from the fact that much more than the spatial distribution of an element can be obtained by this technique Nuclear magnetic resonance signals respond to changes in the main magnetic field or the radio-frequency field in times of microseconds to seconds, depending on the nuclear surroundings, rather than instantaneously. These differences in relaxation times may be used to distinguish between different parts of an object, even when the nuclear concentrations are identical. For example, in a CW NMR spectrum the signal intensity depends upon the product of the two relaxation times T1 and T2. As the radio-frequency power level is increased, the signal from nuclei with the largest value of T1 * T2 will saturate, or decrease in intensity. The result is that the relaxation time differences permit clear distinctions to be made between sample regions that might appear identical in a simple magnetograph or in an X-ray or optical photograph. Experiments may also be divised [sic] that distinguish signals by either their T1 or T2 values, giving great flexibility to the technique.