Photoacoustic 3D brain phantom and multi speed of sound reconstruction

Photoacoustic tomography (PAT) is emerging as a popular non-radioactive imaging modality in cerebral imaging research. However, skull-induced aberrations pose significant challenges for achieving clear cerebral vascular distribution in PAT cerebral imaging. Additionally, the limitation of available photoacoustic data hampers the application of data-driven correction methods. In this study, we developed a comprehensive three-dimensional numerical human brain phantom dataset, derived from clinical T1/T2-weighted magnetic resonance imaging (MRI) data and corresponding magnetic resonance angiography (MRA) data. A realistic and feasible experimental setup for Monte Carlo-based optical simulation and k-wave-based acoustic simulation was designed to generate photoacoustic data. We employed multi-speed-of-sound PAT image reconstruction using two conventional methods to test cerebral imaging. The evaluation results indicate that the reconstructed image quality is significantly enhanced through multi-speed-of-sound image reconstruction.