Radon-full-waveform inversion for suppressing scalp reverberation and skull-induced aberration in transcranial...

Photoacoustic computed tomography is promising for noninvasive imaging of cerebral function but is limited by skull-induced acoustic aberrations. Moreover, strong superficial signals from the scalp often produce reverberations within the skull, which interfere with cortical signals and introduce artifacts in the cortical imaging region. Conventional universal back-projection (UBP) and full-wave-based reconstruction algorithms fail to adequately mitigate these effects, resulting in degraded image fidelity. To address these challenges, we propose a hybrid framework, termed Radon Transform-Full Waveform Inversion (RT-FWI), to simultaneously suppress scalp artifacts and correct for skull-induced acoustic aberrations. The method begins with a high-resolution linear Radon transform to isolate scalp-related multiple reflections without relying on prior velocity models. The separated signal components are used to reconstruct a cleaned wavefield and to formulate a spatially adaptive regularization term that constrains the subsequent elastic FWI process, thereby preventing artifact reinforcement. Numerical simulations and ex vivo experiments demonstrate that RT-FWI outperforms both UBP and conventional FWI, improving the axial spatial resolution from 1.14 mm to 0.28 mm and the structural similarity index from 0.415 to 0.875. The proposed approach constitutes an advancement toward transcranial photoacoustic brain imaging.