Joint reconstruction of initial pressure distribution and acoustic skull parameters in transcranial photoacoustic computed tomography
The development of photoacoustic computed tomography (PACT) for neuroimaging in humans will fill an important void left by available imaging techniques. However, due to the presence of the skull, accurate image reconstruction in transcranial PACT remains challenging. Variations in the shear and longitudinal wave speed distributions due to the skull can induce strong aberrations in the measured photoacoustic wavefields. To mitigate these artifacts, image reconstruction methods in transcranial PACT require knowledge of these acoustic properties. However, such information may be difficult to obtain in practice. To circumvent this, we developed a joint reconstruction (JR) method for transcranial PACT where the longitudinal and shear speed distributions are reconstructed concurrently with the sought-after initial pressure distribution. The joint estimation of the initial pressure, longitudinal speed, and shear speed distributions from PACT data alone is unstable. To overcome this instability, we propose to incorporate prior information about the acoustic properties of the skull. Specifically, a low-dimensional parameterized acoustic representation of the skull is established with the aid of adjunct CT data. The use of a low-dimensional representation of the acoustic skull parameters effectively overcomes the instability of the JR problem and allows stable reconstruction of the acoustic skull parameters and the initial pressure distribution concurrently. To validate the proposed method, we conducted 3D numerical studies based on realistic human skull models derived from adjunct CT data. The efficacy of the proposed JR method was demonstrated through accurate reconstruction of the initial pressure, longitudinal speed, and shear speed distributions from PACT measurement data alone.