Transcranial ultrasound imaging with speed of sound-based phase correction: A numerical study

This paper presents a numerical study for ultrasound transcranial imaging. To correct for the phase aberration from the skull, two critical steps are needed prior to brain imaging. In the first step, the skull shape and speed of sound are acquired by either CT scans or ultrasound scans. In the ultrasound scan approach, phased array and double focusing technique are utilized, which are able to estimate the thickness of the skull with a maximum error of around 10% and the average speed of sound in the skull is underestimated by less than 2%. In the second step, the fast marching method is used to compute the phase delay based on the known skull shape and sound speed from the first step, and the computation can be completed in seconds for 2D problems. The computed phase delays are then used in combination with the conventional delay-and-sum algorithm for generating B-mode images. Images of wire phantoms with CT or ultrasound scan-based phase correction are shown to have much less artifact than the ones without correction. Errors of deducing speed of sound from CT scans are also discussed regarding its effect on the transcranial ultrasound images. Assuming the speed of sound grows linearly with the density, this study shows that, the CT-based phase correction approach can provide clear images of wire phantoms even if the speed of sound is overestimated by 400 m s^-1, or the linear coefficient is overestimated by 40%. While in this study, ultrasound scan-based phase correction performs almost equally well with the CT-based approach, potential problems are identified and discussed.