Background: The acoustic noise in magnetic resonance imaging (MRI) potentially depends on the measurement position and presence of a patient inside the scanner bore.
Purpose: To analyze the spatial characteristics of the acoustic noise by using the gradient-pulse-to-acoustic-noise transfer function (GPAN-TF) with and without a human-body phantom on the examination table.
Material and methods: Acoustic noise waveforms were acquired at 80 and 110 measurement positions with and without a phantom. The GPAN-TFs µPa/(mT/m) in the coils were calculated by deconvolution. The phantom effect on the spatial distribution of the acoustic noise was assessed using the peak sound pressure levels (SPLs), mean values, peak values, and peak frequencies of the GPAN-TFs.
Results: The peak SPLs in all positions for the X-, Y-, and Z-gradient coils were increased by 11.1 dB, 1.4 dB, and 6.1 dB, respectively, compared with the peak SPL of the magnetic isocenter. The maximum peak SPLs among all positions of the X-, Y-, and Z-gradient coils with the phantom were increased by 4.9 dB, 7.4 dB, and 6.9 dB, respectively, relative to those without the phantom. However, the peak SPLs decreased at some positions with the phantom placed on the table (X-gradient coil = 4.6 dB, Y-gradient coil = 5.0 dB, Z-gradient coil = 8.4 dB). The most common peak frequencies were in the range of 2000-3000 Hz.
Conclusion: "Hotspot" areas with and without the phantom were associated with acoustic noise sources in the clinical MRI scanner and were enhanced by the phantom's presence.
Keywords: Magnetic resonance imaging; acoustic noise; frequency analysis; phantom; spatial analysis; transfer function.