Numerical investigation of optimal transmission-reception conditions for aliasing-free ultrasound localization microscopy

초록

Ultrasound localization microscopy (ULM) can surpass the diffraction-limited resolution of conventional ultrasound by localizing individual microbubbles with sub-pixel precision. However, if the point-spread function (PSF) of the imaging system is insufficiently sampled, aliasing artifacts arise and degrade both microbubble localization and motion correction accuracy. In this study, we derive and validate a set of transmit-receive conditions that ensure artifact-free PSFs without unnecessary computational overhead. We demonstrate that by appropriately selecting the plane-wave steering angles, transmit pulse cycle count, and receive aperture (Fnumber) in relation to the pixel spacing, high spatial frequencies in the PSF remain within the Nyquist limit for all imaging depths. Through a series of simulations, wire phantom tests, and custom flow phantom experiments, we compare undersampled, oversampled, and balanced parameters. The balanced configurations, where the maximum frequency of the PSF matches the beamforming grid, consistently mitigate aliasing artifacts, eliminate grid-like patterns, and preserve microbubble trajectories under sub-pixel translations. In a flow phantom study, we further confirm that ULM images obtained with these optimized settings retain fine details without incurring the substantial computational costs of an overly fine sampling grid. Our findings highlight the importance of analysing PSF bandwidth in both lateral and axial dimensions and offer a straightforward method to align transmit pulse width, receive aperture, and grid spacing. Ultimately, this approach provides a pathway toward efficient, high-fidelity ULM, with significant implications for real-time super-resolution imaging in clinical and preclinical environments.

키워드