Chapter 2 — MRI Signal, Contrast, and Reconstruction

MRI intensity is a protocol-dependent measurement

Chapter 1 ended with a reconstructed CT image: a computed object whose values depend on the measurements and choices that produced it. MRI begins with a different kind of measurement, but the same discipline applies. An MR image is reconstructed from a radio-frequency signal produced by proton magnetisation after magnetic-field and RF manipulations. It is not a photograph of tissue, and an ordinary MR intensity value is not a universal physical scale that can be carried from one scan to another.

That distinction is easy to miss because the images look familiar. The same knee, for example, can be shown with fat bright in one series, fluid bright in another, and a suppressed background in a third. Those differences are not three conflicting reports of one scalar tissue property. They are consequences of how the sequence prepared the magnetisation, when it sampled the signal, how the acquisition encoded spatial information, and how the data were reconstructed and displayed.

The chapter therefore follows one causal chain. A static field establishes a net magnetisation. An RF pulse changes its orientation, and the magnetisation then evolves through simultaneous longitudinal recovery and transverse loss of coherence. Sequence timing selects which of those changes contributes most to the measured signal. Inversion and diffusion preparation can deliberately suppress or sensitise the signal. Gradients encode the resulting signal in k-space, and a reconstruction turns sampled k-space into the image that is stored and analysed. Field strength, hardware, protocol, sampling, and reconstruction remain part of that image’s provenance at the end of the chain.

The point is not to catalogue every pulse sequence. It is to give the reader a model that explains why contrast is relative, how a quantitative-looking map is made from multiple measurements, and why an image can change when its acquisition changes. The same model will later explain acquisition-driven domain shift and why a synthetic image may contain an estimate rather than a measurement.

The first distinction is the most important one: relative intensity and quantitative limits.

Sources

• Berger A. Magnetic resonance imaging. BMJ. 2002;324(7328):35. doi:10.1136/bmj.324.7328.35. (Full text)
• Ridgway JP. Cardiovascular magnetic resonance physics for clinicians: part I. J Cardiovasc Magn Reson. 2010;12:71. doi:10.1186/1532-429X-12-71. (Full text)
• Wiltgen T, Voon C, Van Leemput K, Wiestler B, Muehlau M. Intensity scaling of conventional brain magnetic resonance images avoiding cerebral reference regions: A systematic review. PLoS One. 2024;19:e0298642. doi:10.1371/journal.pone.0298642. (Full text)