Non-invasive observation of spatio-temporal neural activity of large neural populations distributed over the entire brain of complex organisms is a longstanding goal of neuroscience [1,2]. Recently, genetically encoded calcium indicators (GECIs) have revolutionized neuroimaging by enabling mapping the activity of entire neuronal populations in vivo . Visualization of these powerful sensors with fluorescence microscopy has however been limited to superficial regions while deep brain areas have so far remained unreachable . We have developed a volumetric multispectral optoacoustic tomography platform for imaging neural activation deep in scattering brains . The developed methodology can render 100 volumetric frames per second across scalable fields of view ranging between 50-1000 mm3 with respective spatial resolution of 35-150µm. Experiments performed in immobilized and freely swimming larvae and in adult zebrafish brains expressing the genetically-encoded calcium indicator GCaMP5G demonstrated, for the first time, the fundamental ability to directly track neural dynamics using optoacoustics while overcoming the depth barrier of optical imaging in scattering brains .
It was further possible to monitor calcium transients in a scattering brain of a living adult transgenic zebrafish expressing GCaMP5G calcium indicator . Fast changes in optoacoustic traces associated to GCaMP5G activity were detectable in the presence of other strongly absorbing endogenous chromophores, such as hemoglobin. The results indicate that the optoacoustic signal traces generally follow the GCaMP5G fluorescence dynamics and further enable overcoming the longstanding optical-diffusion penetration barrier associated to scattering in biological tissues .
The new functional optoacoustic neuroimaging method can visualize neural activity at penetration depths and spatio-temporal resolution scales not covered with the existing neuroimaging techniques. Thus, in addition to the well-established capacity of optoacoustics to resolve vascular anatomy and multiple hemodynamic parameters deep in scattering tissues, the newly developed methodology offers unprecedented capabilities for functional whole brain observations of fast calcium dynamics.
Scanning optoacoustic microscopy operates in two distinct regimes optical resolution microscopy relies on a focused illumination and acoustic resolution microscopy that forms images by focusing the received acoustic field. Recently, a number of approaches have been proposed that combine those two modes of operation to create a highly scalable technique that can image at multiple penetration scales by gradually exchanging microscopic optical resolution in superficial tissues with ultrasonic resolution at diffuse (macroscopic) depths. However, scanning microscopy schemes commonly employ acquisition geometries that impede the use of synthetic aperture techniques to achieve meaningful images due to non-stationary illumination patterns and strong non-uniformity of the excitation light field.
Here we present a Weighted Synthetic Aperture Focusing Technique (W-SAFT) as a universal framework that effectively accounts for the non-uniform distribution of both the excitation light field and spatial sensitivity field of the detector. As a result, W-SAFT maintains optical resolution performance at superficial depths while improving the acoustic resolving capacity for deeper tissues. The dynamic range of the optoacoustic data is compressed using a general fluence decay term applied to the W-SAFT operator, allowing a more uniform visualization of the entire imaged volume. Our three-dimensional algorithm makes use of the sample's surface to account for the heterogeneity produced when scanning a finite-size light beam. We tested a GPU implementation of W-SAFT with numerical simulations and showcase its performance on experimental data acquired from targets embedded in tissue mimicking phantoms.
The acoustically-mismatched skull bone poses significant challenges for the application of ultrasonic and optical techniques in neuroimaging, still typically requiring invasive approaches using craniotomy or skull thinning. Optoacoustic imaging partially circumvents the acoustic distortions due to the skull because the induced wave is transmitted only once as opposed to the round trip in pulse-echo ultrasonography. To this end, the mouse brain has been successfully imaged transcranially by optoacoustic scanning microscopy. Yet, the skull may adversely affect the lateral and axial resolution of transcranial brain images. In order to accurately characterize the complex behavior of the optoacoustic signal as it traverses through the skull, one needs to consider the ultrawideband nature of the optoacoustic signals.
Here the insertion loss of murine skull has been measured by means of a hybrid optoacoustic-ultrasound scanning microscope having a spherically focused PVDF transducer and pulsed laser excitation at 532 nm of a 20 μm diameter absorbing microsphere acting as an optoacoustic point source. Accurate modeling of the acoustic transmission through the skull is further performed using a Fourier-domain expansion of a solid-plate model, based on the simultaneously acquired pulse-echo ultrasound image providing precise information about the skull's position and its orientation relative to the optoacoustic source. Good qualitative agreement has been found between the a solid-plate model and experimental measurements.
The presented strategy might pave the way for modeling skull effects and deriving efficient correction schemes to account for acoustic distortions introduced by an adult murine skull, thus improving the spatial resolution, effective penetration depth and overall image quality of transcranial optoacoustic brain microscopy.
Characterizing disease progression and identifying possible therapeutic interventions in stroke is greatly aided by the use of longitudinal function imaging studies. In this study, we investigate the applicability of real-time multispectral optoacoustic tomography (MSOT) as a tool for non-invasive monitoring of the progression of stroke in the whole brain. The middle cerebral artery occlusion (MCAO) method was used to induce stroke. Mice were imaged under isoflurane anesthesia preoperatively and at several time points during and after the 60-minute occlusion. The animals were sacrificed after 24 hours and their excised brains frozen at -80°C for sectioning. The cryosection were stained using H&E staining to identify the ischemic lesion. Major vessels are readily identifiable in the whole mouse head in the in vivo optoacoustic scans. During ischemia, a reduction in cerebral blood volume is detectable in the cortex. Post ischemia, spectral unmixing of the optoacoustic signals shows an asymmetry of the deoxygenated hemoglobin in the hemisphere affected by MCAO. This hypoxic area was mainly located around the boundary of the ischemic lesion and was therefore identified as the ischemic penumbra. Non-invasive functional MSOT imaging is able to visualize the hypoxic penumbra in brains affected by stroke. Stopping the spread of the infarct area and revitalizing the penumbra is central in stroke research, this new imaging technique may therefore prove to be a valuable tool in the monitoring and developing new treatments.
Acoustic resolution optoacoustic microscopy is a powerful modality allowing imaging morphology and function at depths up to a few centimeters in biological tissues. This optoacoustic configuration is based on a spherically-focused ultrasonic transducer raster scanned on an accessible side of the sample to be imaged. Volumetric images can then be formed by stacking up the recorded time-resolved signals at the measured locations. However, the focusing capacity of a spherically-focused transducer depends on its aperture and the acoustic spectrum of the collected signals, which may lead to image artifacts if a simplistic reconstruction approach is employed. In this work, we make use of a model-based reconstruction procedure developed in three dimensions in order to account for the shape of spherically focused transducers in acoustic resolution optoacoustic microscopy set-ups. By discretizing the transducer shape to a set of sub-sensors, the resulting model incorporates the frequency-dependent transducer sensitivity for acquisition of broadband optoacoustic signals. Inversion of the full model incorporating the effects of the transducer shape is then performed iteratively. The obtained results indicate good performance of the method for absorbers of different size emitting optoacoustic waves with different frequency spectra.