Paul Lauterbur and the Invention of MRI

by M. Joan Dawson

  So there you have it. Paul was the inventor of the basic method that is used in all MRI scanners worldwide. As he said, “I thought of it, I showed it would work and I did it.” A new age of medical diagnosis had begun.

  What Is MRI?

  Paul liked to describe MRI this way. The nuclei of many kinds of atoms, commonly hydrogen, are tiny magnets. In the Earth’s magnetic field they line up to some extent just as you walk around. When you walk past a piece of iron, they’ll flop around in different directions. We may think of us as having microscopic compass needles precessing (spinning on their axes like gyroscopes) in an orderly direction. To make an MR image, this tendency of the nuclei to line up in the direction of a magnetic field is manipulated and measured. Since the nuclei from different regions of the body can be made to precess at different frequencies, these frequencies yield signals that are location dependent. Computer images can be calculated, enhanced, and displayed.

  To understand the technique, let’s first look at its full name, nuclear magnetic resonance imaging, usually abbreviated to MRI. An atomic nucleus exposed to a static magnetic field resonates when a varying magnetic field is applied at the proper frequency. An image is computed from the frequency and phase (timing) of the resonance signals of the responding nuclei.

  MRI revolutionized diagnostic medicine. It images the interior of structures noninvasively and, unlike x-rays or CT, without causing harm. It is especially good at imaging the soft tissue of the body, and this is the reason it became a preeminent diagnostic technique so quickly. It is now the jewel in the crown of medical imaging techniques.

  MRI is an extremely versatile technology. Because it looks at the environment of ubiquitous atomic nuclei, depending on what information is of interest it can reveal anything from blood flow to nerve-fiber track orientation, or show in real time the beating of the heart. And, although MRI is usually thought of as a medical tool, it has important applications in the food and oil industries as well.

  MRI is a specialized subfield of NMR, a complex phenomenon of nuclear quantum mechanics governed by a simple equation. The Larmor equation, ν = k B0, shows a constant (k) relationship between signal frequency (ν, the useful data) and the magnetic field (B0). It is easy to see that a homogeneous magnetic field is required to obtain chemical information. If the magnetic field does not have exactly the same strength everywhere, then the “I am here” signals from nuclei at different positions will be shifted to different frequencies, causing a broad fuzz around the average signal.

  Variations in magnetic field or magnetic field gradients have always been important in NMR. If you were a chemist or a physicist you struggled to get rid of them because they messed up the NMR signal. But Paul turned the problem on its head, making field gradients a solution. He added a G (magnetic field gradient) to the Larmor equation, making it ν = k (B0 + G), where Gx, y, z is unique to each location. Simply put, imaging is this: If instead of placing an object in a very uniform magnetic field as everyone had been doing since NMR began, one puts it in a magnetic field that varies in strength from one physical place to another in a known and controlled way, then the numbers representing frequency, measured in Hertz (cycles per second), of an NMR spectrum can be converted into numbers that represent position, measured in centimeters. In effect, he was giving magnetic Zip Codes to the atomic nuclei.

  Let us do a puzzle. I put two simple objects in the sample holder of an NMR spectrometer; your task is to find them. Paul did this first with two capillaries of water, that fluid with the symbol H2O, surrounded by heavy water (D2O), so I will do the same. You then tune the magnet to ordinary water and put a magnetic field gradient along the horizontal axis, as shown in the first figure. You observe two humps of NMR signal in this spatial direction; this is a “one-dimensional image,” revealing where the capillaries are along this gradient path. It’s useful information, but not most people’s idea of an image. You decide to look along another direction (a perpendicular gradient would be a good idea). Now you locate the test tubes by applying gradients in both the horizontal and vertical directions so that you can draw a map showing exactly where to find these objects. Then a third gradient is applied at 45 degrees to the others. The second figure shows the first image, constructed in this simple way.

  Figure 1.1

  (a) The setup used for the first imaging experiments. (b) The hand-processed results. From the first published paper on magnetic resonance imaging, published by Lauterbur in Nature, 1973. Reproduced by permission.

  The human body is mostly water, so the water-filled tubes are good, simple models of it. But in the complicated human body you won’t get much of a picture with just two gradients. The same principle applies; you just need many, many more gradients to make a detailed picture. But the mathematics required to make an image is not trivial, and Paul worked on it for years. When other people became involved in developing Paul’s idea, many different methods of obtaining images were demonstrated. Today, modern MR radiologists take advantage of the fact that not just the frequency but also the phase (direction) of the NMR signal can be varied spatially using a time-dependent magnetic gradient. A combination of signal frequency and phase gives the best speed and resolution of an MR image.

  Paul believed that all his years of using NMR for chemical studies, years of his own hands-on efforts to optimize magnetic fields for NMR and to find very small signals within suboptimal fields, in some way prepared the insight from which MRI sprang. The trick in chemical NMR is to minimize the unwanted gradients in order to produce the clearest possible signal in the largest possible volume. At first this “shimming” was done with a large crowbar-like lever with a long pole attached to increase torque. Strong guys could move the big iron magnets a few parts per million with this lever, but smaller people were seen dangling off its end while trying to adjust the magnetic field. Success was greatly improved with the invention of electronic “shim coils.”

  Paul wrote a single-page note, dated March 22, 1970, a year and a half before he invented MRI, showing how to impose magnetic field gradients to cancel out the natural ones and make the field more homogeneous.4 This insight was shared by many people. But it is interesting that this idea is exactly the opposite of, and in Paul’s case a specific precursor to, using gradients to make an image. His mind was prepared for the big breakthrough.

  Zeugma What?

  Paul named his baby “zeugmatographic imaging” or “zeugmatographic magnetic resonance imaging,” from the Greek ζενγμα, “that which is used for joining,” to emphasize that he had found an entirely new way of making images, unknown to physics at the time. He did not want to call it simply MRI because his idea would work with so many other modalities as well. Zeugmatography (MRI) is fundamentally different from all other imaging techniques. In classic image formation, the resolution obtained from an electromagnetic wave is no better than half its wavelength, which is why light microscopy has less spatial resolution than electron microscopy. Light microscopy, electron microscopy, infrared and ultraviolet imaging, and the human eye all work in this same way. So the use of radio frequencies with wavelengths measured in meters or kilometers to image the human body by magnetic resonance is, on the face of it, absurd.

  For an unfinished paper, Paul wrote:

  This strange word was introduced into the scientific literature for a reason. Magnetic resonance imaging was made possible not by physics analogous to that of optics, with rays of particles or waves passing through a small region of an object (as in x-ray computed tomography [CT] for example) but by a type of physics not previously employed in imaging. That type channels the image forming process into two parts, localizing a volume (determining its coordinates) and measuring some physical property associated with it. Together the two responses give a spatial representation of the distribution of those properties in space, or, in short, an image (which may be one, two or three-dimensional, or, by extending the definition of a coordinate, even more dimensions).5

  It is the n
ovel method Paul thought of for coupling two fields of long wavelength and low energy that makes MRI possible. It was his original insight that gave birth to all the specialized ways of doing MRI that followed. In effect, he was using the magnetic field gradient to encode spatial information—the magnetic Zip Codes—as spectroscopic frequency and then to reconvert the frequency back into an image. There is no diffraction or any of the traditional complications of optics to go with it. Sometime during the eventful night in September 1971 Paul understood that the imaging technique he had invented was not limited to MRI. In principle, this method of noninvasive imaging by coupling two fields is useful in any modality for which any kind of signal can be made to vary spatially. Paul emphasized this point in his first paper.

  It was hard to grasp that Paul had discovered a whole new principle of imaging, one that was not subject to ordinary imaging rules. Much of what drove Paul intellectually at this time was the realization that he had invented a totally new imaging concept. Mercifully, the name zeugmatography has died out. Paul was disappointed when scientists and physicians tossed out the name for his brainchild. Paul said he felt like “a parent that had given his kid a strange name—when the kid grows up, he won’t use it.” But really, he couldn’t blame us. “MRI” is so much simpler to say and remember.

  Paul held out for a long time, continuing to use “zeugmatography” in the title of his papers for the next fifteen years. Despite the generality of the principle, zeugmatographic techniques did not immediately cross into other fields. Paul showed in the late 1970s that zeugmatography could be used for electron spin resonance (ESR) imaging, another magnetic resonance technique. ESR imaging has since been well explored by others. Additional obvious applications include electrical field gradients and optical transitions. But the idea did not resurface for these uses until many years later.

  A Complete Vision

  Paul’s discovery of MRI is well known. Not so well known is the completeness of his vision at the time. Even the citation for his Nobel Prize in 2003 reads, “In the early 1970s Paul Lauterbur discovered the possibility to create a two-dimensional image by introducing gradients in the magnetic field.” In fact, Paul’s original notebook refers to imaging in three dimensions.6 He always thought that two-dimensional imaging was a stopgap until true three-dimensional imaging became practical, and this it turned out to be.

  In “The Notebook,” Paul cited contrast by density, relaxation or signal decay times, and diffusion. He suggested spectroscopic imaging and isotope exchange imaging. He indicated that imaging could be done by time-dependent and time-independent methods, and within the next days and weeks he described imaging flow and the use of magnetic contrast agents, analogous to the dyes used in x-ray and radioisotope examinations.

  Thunder of Objections

  Paul’s ideas were met with vehement objections. Why? The basic idea behind MRI is so simple that once the creative leap was made, it is hard to understand why it hadn’t been done much earlier. How could outstanding scientists have used NMR for more than thirty years without thinking of using the data to obtain pictures? “Mental block,” said Paul. Some people just couldn’t get over the classic relationship between wavelength and spatial resolution. “The principles involved are completely different,” Paul more than once patiently explained. It was really hard for many physicists (“who were thinking about physics in too narrow a way”) to wrap their heads around these ideas.7

  The mental block Paul described was demonstrated in a particularly irksome way when Paul gave one of his first lectures on MRI at the Bell Laboratories. Saul Meiboom, a highly respected NMR physicist, spoke from the audience: “I don’t know what you’re doing but it can’t be right. It violates the Heisenberg uncertainty principle.” This was a public accusation of fraud: Paul had shown the images, and if he had not got them the way he described, he must be lying. Another time, at a meeting at Morzeen, a beautiful region in alpine France, Paul sat on a bus next to a member of Anatole Abragam’s laboratory. This man told of the disdain that Abragam felt for Paul’s work. Abragam was known to be arrogant, but he was brilliantly productive and the author of “Abragam’s Bible,” the textbook everyone used to learn NMR. “You have to have some psychological armor, “Paul said, “to stand up to that sort of hostility.”

  Dave Kramer, a student of Paul’s in the early days of imaging, remembers the opposition from those days. “There were many skeptics; perhaps rightly so. At the time, X-ray CT was the darling of medical imaging and everything was compared first to that benchmark. Many very influential people thought that NMR had too weak of a signal to ever get the spatial resolution of CT and their interest would wane.”8 One of those skeptics came to Paul years later at a meeting in Houston, Texas. “I used to think you were a charlatan,” he said. “Now I know better.”

  Friends, and even students, also had their doubts. Paul tells the story of coming into the lab one morning to find a student agitated, tense, and weary. “It doesn’t work! MRI doesn’t work!” The student had been up all night proving that the physics was wrong, which he demonstrated with complex scribbled equations. Paul never forgot this demonstration of how old assumptions refuse to die, for the student had been doing MRI for months. To another longtime friend and collaborator, Hal Swartz, Paul’s new “zeugmatography” looked a bit of a stretch: “[Paul] had recently published on a new and strange approach, called zeugmatography. The idea, I thought, clearly was not going to amount to very much except perhaps being of some instrumental and theoretical interest.”9

  John Waugh of MIT, a scientist whom Paul greatly liked and admired, illustrated another rather amusing skepticism when he visited the Stony Brook campus to talk about his highly respected work in solid state NMR. As they were chatting over drinks at a bar, Paul mentioned that he had been trying for some time to actually do Peter Mansfield’s “crystal” experiment, the experiment that started Peter (later Sir Peter, co-laureate with Paul in 2003) toward his work in MRI but which he was never able to accomplish. John answered, “You’ll never succeed. If you did, you would get a Nobel Prize, and I don’t think you have one in you!”

  Some early objections to MRI were downright comical. Several NMR spectroscopists confused secondary details of the procedure with the fundamental facts of physics. For example, it is usual in chemical NMR studies to slowly spin the test tube containing the sample. This has the effect of removing (averaging out) inhomogeneities in the magnetic field and sharpening the signals. So in the early days of MRI, Paul was often asked, “But how do you spin the patient?”

  Paul’s response to all this criticism was, “When you know you’re right, you just have to ignore the criticism and counter it when necessary.”10 When Paul was asked, “How were you able to stand up to the tremendous criticism you received?” his response was, “It is like those pictures, where something is hidden on a page of seemingly meaningless dots. Once you see the image in the dots, you can’t unsee it. Also, I knew the answer to every criticism I was given. I understood the problems in thinking that had led to an erroneous conclusion. That gives one a great deal of confidence in one’s own thinking.”

  So who was Paul Lauterbur, and where did he come from?

  2

  Portrait of a Scientist as a Young Man

  For the child is father to the man.

  —William Wordsworth

  Science is conducted in as many ways as there are scientists. What may be Paul’s most important gift is illustrated by an experience during his freshman year in high school. He was looking at a chemistry book and came across a description of how the carbon content of a substance being burned determines the color of the flame. He was mortally chagrined, not because he had not known this previously but because he had never even thought to ask! How could he ever make anything of himself, he worried, if he was really such a dull boy! The capacity and courage to ask questions others ignored would become Paul’s most important scientific resource.

  Paul himself was very skeptical of
efforts to understand the process of scientific creativity. “Most descriptions of the tangled processes of scientific discovery are cartoons, so oversimplified as to be useless for any purpose other than providing apparent evidence for preconceived notions, polemical disputes, and the justification of budgets.”1 With this in mind I set out here to describe the boy, Paul Lauterbur, and his “tangled processes of scientific discovery.”

  The Fathers of the Man

  They were German American. They were Roman Catholic. They were small-town, midwestern, middle class. Paul’s father, Edward Joseph Lauterbur (1899–1967), was an engineer with the Hobart Company in Troy, Ohio. Paul was raised in nearby Sidney, a small town twenty-seven miles northwest of Dayton. Sidney is located in the valley of the Great Miami River, the village spreading from its origin in a little hollow on the river’s west side. When Paul was a child, Sidney numbered about nine thousand souls, big enough for a cultural life but small enough to have midwestern country values. He was born to the virtues of these hard-working, self-sufficient, and very productive people, virtues that became the bedrock of his life.

  The Lauterburs were one tough family, who worked hard and believed in their own capabilities. James, the eldest of Paul’s uncles, started the Lauterbur Manufacturing Company, where among other things he sold and repaired cars. Frank, or FX (for Francis Xavier), the third boy, was president of the Peerless Manufacturing Company. He died at a young forty-four years of age, and his wife, Wilhelmina (Aunt Billie), then married his brother Leo, causing much tongue wagging. Among them, the brothers held about sixty patents. The Lauterbur sisters, Anna and Mary, never wed, and unloosed their maternal instincts on the nieces and nephews; both worked for a living, unusual in those days. Mary ran the secretarial pool at Peerless, and Anna taught at Ball State Teachers College in Muncie, Indiana. Aunt Anna was the sweet one who always had time for a child. Aunt Mary was “respectable” and religious; she could tell Jesus Christ to stop any nonsense.