Signal, noise, and contrast in nuclear magnetic resonance (NMR) imaging.

Abstract

Calculations of the sensitivity of the saturation recovery and inversion recovery pulse sequences used in nuclear magnetic resonance imaging show the former to be superior in discriminating between tissues with the same proton density but different T1's. Two other pulse sequences, which are combinations of the above, have also been analyzed. These have lower T1 discrimination sensitivity, but other considerations, such as self-normalization, may still make them attractive. The calculations are only valid for selective excitation pulse sequences in which the selected slice profiles are approximately rectangular, and thus a sin(bt)/t radiofrequency excitation is desirable. In order to ensure that the saturation recovery sequence gives valid results for pulse repetition times comparable to or shorter than T2, it is necessary to destroy the coherence between pulse applications. For this purpose we use a series of "spoiler" gradient pulses between pulse trains. The saturation recovery pulse sequence also has the advantage that, by the correct choice of interpulse spacing, sensitivity close to the optimum T1 discrimination can be achieved over a wide range of T1 values. This has the potential advantage to the clinician of simplifying his choice of parameters for imaging.