Noise and interference in measured NMR images

Abstract

This article describes the measurement and evaluation of noise and interference in experiments with nuclear magnetic resonance (NMR). Their presence in experimental results causes irreversible artefacts in measured images or errors in the calculated values of investigated quantities. Noise in the signal also reduces the sensitivity of the measuring instrument. From the point of view of results, there is no significant difference between noise and interference. Both signals are multi-frequency or have a changing frequency, meaning that they must be characterised using their power or root mean square (RMS) values of voltage or current. The RMS voltage at the output of the NMR receiver with a receive coil in its input was measured using an NMR console and compared to thermal noise voltages of known sources, e.g. resistors with known resistivity in the input of their receiver. The comparison indicates the possible inherence of interference. The noise measurement using the console was performed in two ways: using an acquisition only and using a complete measuring NMR sequence, which resulted in an image. The results served to test the quality of the hardware used for NMR experiments. It has been proven that the presented method is suitable for testing and improving the signal-to-noise ratio of NMR scanners without the need for the accumulation of seldom-used expensive instrumentation.