Quiet echo planar imaging for functional and diffusion MRI

    loading  Checking for direct PDF access through Ovid


While acoustic noise reduction is a general aim for MRI examinations to enhance patient comfort or to avoid unwanted activation in functional MRI (fMRI) studies 1, it is of particular importance for the success of fetal examinations. The vulnerability of the unborn human to excessive acoustic noise is postulated to contribute in the extreme case to high frequency hearing loss, shortened gestation, and decreased birth weight [for a review see reference 2]. While no incident involving MRI has been reported to date, and indeed retrospective studies of human subjects have shown no detectable long term effects of noise in fetal MRI 3, it puts a particular emphasis on adequate levels of protection for fetal MRI scans. Sound levels are typically expressed on a dB (A) scale, including a weighting to match the perceived relative loudness in the human ear (termed A‐weighting). A number of studies 4 state and discuss regulations for safe exposure, all stating the limits in dB (A). Studies evaluating the acoustic environment and level of protection provided by the maternal torso surrounding the fetus have been performed in sheep models 5, and an analogue has been studied in humans by placing a hydrophone placed in the fluid filled stomach of an adult male 6, which demonstrated a typical attenuation of 30 dB, although conditions in‐utero have been found to depend on the position of the fetus as well as the frequency of the sound 7. The thickness of the amniotic fluid layer contributes only marginally to sound attenuation in general 5.
External noise protection techniques for the fetus in‐utero are either not feasible (ear plugs, headphones) or not generally applicable (wrapping the mother's abdomen). Therefore, reducing the acoustic noise output of the scanner is particularly desirable in this subject group. In clinical practice, this is typically achieved by imposing constraints on gradient slew rate and amplitude. Control of acoustic output in this way is associated with decreased temporal and spatial resolution as well as limited scan efficiency for most sequences, it specifically restricts the performance of echo‐planar‐imaging (EPI) sequences.
Single‐shot EPI is widely used for advanced applications such as fMRI and diffusion MRI (dMRI), and as such is a key tool for connectome type studies 8 of the fetal brain and perhaps other applications involving pregnant subjects. The EPI readout critically relies on rapid switching of gradient polarity and fast gradient rises are commonly employed to shorten other EPI sequence components. Such fast single shot sequences provide an acoustic challenge that is not well addressed by simply de‐rating gradient performance. In addition, novel acquisition techniques such as multiband (MB) imaging 9 introduce extra gradient blips that further contribute to acoustic EPI noise. Finally, the emergence of novel analysis pipelines requiring high angular coverage for multi‐shell diffusion sequences and high temporal resolution for fMRI studies puts additional emphasis on the efficiency of the EPI acquisition. Obtaining this data in acceptable acquisitions times thus further motivates the use of highly accelerated and efficient EPI sequences.
The acoustic noise in the scanner is mainly generated by the gradient system, particularly if there is extremely rapid switching of gradient amplitude and polarity 10. The time varying currents, I, driven through the gradient coils by the gradient amplifiers lead to interactions with the static magnetic field B and thus to Lorentz forces ( JOURNAL/mrim/04.02/01445475-201803000-00023/math_23MM1/v/2018-01-24T161827Z/r/image-png ). These forces work against the coil stiffness and lead to the generation of sound pressure approximately proportional to the velocity of the coil former surface. These effects depend on the geometry and material properties of the individual scanner setup.

Related Topics

    loading  Loading Related Articles