Reduction of laser fluctuations in real time optoacoustic volumetric mouse neuroimaging

The brain represents one of the less known yet most studied organs of the body. Scientists use imaging devices to better understand how the brain works, pursuing the hope of understanding and explaining the coupling mechanisms behind healthy structures and principles of neural dysfunctions. Among various techniques currently used for neuroimaging, optoacoustic (OA) imaging emerges as a promising hybrid modality imaging system that allows non-invasive visualization and 5D investigation of biological structures. This technology combines high sensitivity and specificity of selective light absorption from tissues, with low scattering typical of ultrasound waves, which results in an imaging tool able of achieving optical contrast with the high ultrasonic resolution for imaging depth. However, despite the promising abilities, OA appears to be non-immune to some limitations. In particular, the technique strongly relies on short laser-emitted light pulses to illuminate the target, which causes non-uniform light exposure and fluctuations of the illumination pattern both in time and space. In biological terms, variations in the light source would hamper the correct monitoring of slight changes in neuronal activity, hindering ongoing processes or enhancing unreal changes. The work proposed in this thesis aims to investigate the optical nature of OAI, tackling the light fluctuations from a hardware and software perspective to reduce the alterations and obtain a more homogeneous and stable illumination pattern at the target interface. At first, characterization of the laser beam has been performed, analyzing the variations both at the input and target interface through a camera and computation of statistical parameters. Then, the results for the original setup have been compared with the ones relative to corrective hardware implementations, such as additional water filtering methods and optical corrections, evaluating the fluctuations variations and the temporal correlation over the batch of acquisition. Afterward, the same hardware corrections have been implemented and tested on the full OAI system. The first test has been conducted imaging homogeneous phantoms, to test the efficiency of the hardware optimizations. Furthermore, given the multispectral nature of the acquired information, also software correction employing frame averaging has been implemented and compared to single frame reconstructions. Eventually, pre-clinical in vivo OA neuroimaging has been performed to evaluate the results achieved through the proposed corrections. The optimizations tested within this project proved to be valid allies in the assembly of the OA neuroimaging setup. In particular, water filtering and degassing allow preventing nuisance due to spurious signals generation and system deterioration. In this way, it is possible to reduce the spatial fluctuations present in the light distribution at the target interface, as proved by the strong temporal correlation present among different ROIs randomly selected in the illumination surface. When employed in preclinical applications, that would improve the reliability of imaging of superficial structures, to which the photons are delivered with very little previous scattering. Software correction tested on multispectral OA data proved to be able to reduce the temporal presence of artifacts and random noises, without losing the information related to ongoing hemodynamic changes. To achieve a more stable control on temporal fluctuations, future works could focus on the investigation of temporal correlation among different voxels in raw or reconstructed OA data. Thus, identifying related regions, it would be possible to correct for the temporal variations affecting biologically informative voxels.