A spectroscopy tool, not yet an image
Before the 1970s, nuclear magnetic resonance was an established technique for analysing the composition of a sample as a whole: certain atomic nuclei, placed in a strong magnetic field and hit with a radio pulse at the right frequency, absorb and then re-emit energy in a way that reveals what kind of atoms are present and how they are chemically bonded. What the technique could not do was say where within a sample any particular signal originated, which meant it functioned as a chemistry tool rather than an imaging one. The claim behind magnetic resonance imaging is that this limitation was solvable: that with the right modification to the magnetic field, the same underlying physics could be made to produce a genuine two-dimensional picture of an object’s interior, non-invasively and without ionising radiation.
Lauterbur’s gradient idea
Paul Lauterbur, working at Stony Brook University, proposed the solution in a 1973 paper in Nature titled ‘Image Formation by Induced Local Interaction; Examples Employing Nuclear Magnetic Resonance,’ a paper that was initially rejected before publication on concerns about image quality. His method added gradients across the magnetic field, so that the resonance frequency of nuclei varied slightly depending on their position, which meant the position that a given signal came from could be worked out from its frequency. Applying gradients in different directions and combining the results let him reconstruct a genuine two-dimensional image rather than a single aggregate reading, which is the specific technical step that separates imaging nuclear magnetic resonance from the older spectroscopic version of the same physics.
Mansfield’s faster maths
Peter Mansfield, working separately, developed the mathematical and technical methods needed to make Lauterbur’s basic principle fast and precise enough for practical use, applying Fourier transform techniques to reconstruct images from the raw signal far more efficiently than earlier approaches allowed. This combination of Lauterbur’s gradient principle and Mansfield’s reconstruction methods is what let magnetic resonance imaging move from a proof of concept into a workable clinical instrument. The physical basis of the resulting images rests on the fact that different tissues relax back to equilibrium after the radio pulse at different rates, described by two characteristic times called T1 and T2, and it is the difference in these relaxation rates between tissue types that produces the contrast seen in an MRI scan, generally without the need for an injected contrast agent.
What the scanner actually measures
The technique’s core value holds up clearly: it produces excellent soft-tissue contrast without exposing the patient to ionising radiation, unlike CT or PET scanning, which is why it remains the preferred tool for many neurological and soft-tissue investigations. It does have real limits. Sensitivity is comparatively low, and acquisition times are longer than for CT, while the strong magnetic field rules out scanning patients with certain ferromagnetic implants. Some proposed diagnostic uses have not held up under scrutiny; research applying MRI to diagnosing attention-deficit hyperactivity disorder, for example, has shown considerable variability between studies, and MRI cannot be reliably used on its own to make that clinical diagnosis, a caution worth carrying into any claim that scanning alone can settle a diagnosis it was not originally validated for.
A prize for two, a dispute from a third
The 2003 Nobel Prize in Physiology or Medicine went to Lauterbur and Mansfield for this work, but the decision produced a public and lasting dispute. Raymond Damadian, who had published a 1971 paper in Science reporting measurable differences in T1 and T2 relaxation times between cancerous and healthy tissue, argued that this earlier work had demonstrated MRI’s medical potential and that his exclusion from the prize was unjust; he took out full-page newspaper advertisements pressing the case. The disagreement is over recognition and priority rather than over the physics itself, since Damadian’s tissue-difference observation and Lauterbur’s imaging method addressed related but distinct questions, one about what MRI could detect, the other about how to turn detection into a picture at all.
Whether the argument matters to the picture
This is worth understanding because it separates two things people tend to conflate: the physical principle that makes MRI possible, which is well established and traceable to a specific, identifiable insight in Lauterbur’s 1973 paper, and the question of who deserves credit for the technology as a whole, which remains genuinely contested. The dispute over Damadian’s exclusion is not a minor footnote; it reflects a real ambiguity about which contribution, detecting a medically relevant difference or inventing the method to visualise it spatially, matters more for a prize meant to reward a single discovery. Readers who want to understand how a routine hospital scan actually produces its image, and why the people credited with inventing it are still argued about, will find both threads laid out clearly here.