Custom-access serial holography (CASH) is a new method for optical recording of neuronal activity in 3D at high speed in-vivo. Our implementation allows random access of 20 cells at 1 kHz up to 200 cells at 0.1 kHz in head-fixed behaving mice across a cortical space of 500 x 500 x 500 m3 size. Using fast acousto-optic spatial light modulation, every single laser pulse of a 40 kHz regenerative amplifier is individually patterned to serially access a selection of target cells with a square 5x5 spot excitation volume covering the cell body and, for the prevention of recording artefacts, the surrounding space in anticipation of the cell displacements during animal behavior. The recorded activity is corrected for neuropil signaling by weighted subtraction of a neuropil reference signal obtained by interleaved sampling of neuropil activity close to each cell. We performed 3D-CASH recordings of GCaMP6f expressing neurons in layer 2/3 and 5 of mouse primary visual cortex in response to moving contrast gratings and applied deconvolution to estimate spikes. Thanks to the fast recording permit by 3D-CASH, the cortical laminar structure is revealed in the temporal organization of the activity: pairwise correlation was higher between intralaminar vs. interlaminar neuron pairs; principal component analysis of the correlation matrix revealed a component assigning weights of opposite sign to neurons in different layers; closest follower spikes occurred with higher probability in a neuron of the same layer. 3D-CASH allows also following the response to the temporal periodicity of the stimulus, which features a phasic (R1) and a non-phasic component (R0). R1/R0 values are broadly distributed with weak bimodality resembling the transition between pure non-phasic response (complex receptive field) to phasic response (simple receptive field). Our data validate thus 3D-CASH as a method for assessing neuronal activity in 3D-distributed cortical circuits at high sampling rate.

This conference presentation, “Measurement of intraneuronal transport in vivo in zebrafish larvae brain by tracking nanocrystal-labelled endosomes with fast nonlinear microscopy” was recorded for the Biomedical Spectroscopy, Microscopy, and Imaging II conference at SPIE Photonics Europe 2022.

Understanding the causal relations between neuronal activity and behavior is one of the grand challenges of modern neuroscience. From a theoretical point of view, the requirements for reverse engineering the brain architecture are clear: we need to (1) measure neuronal activity from several individual cells, from which we can make hypothesis on their function; (2) manipulate (activate or inhibit) the same cells to test the accuracy of the hypothesis; (3) doing all this in animals during natural behavior. Two-photon (2P) micro-endoscopy could be the technological answer to these urgent needs and is today a thriving research field. However, 2P micro-endoscopy has so far focused on imaging neuronal activity, rather than manipulating it at will, which limits the ability to directly test possible causal links with behavior. To overcome this limitation, we have developed a new two-photon fiber bundle-based micro-endoscope (2P-FENDO) for the simultaneous functional imaging and optogenetic photostimulation of neurons in freely moving mice. By using computer generated holography, 2P-FENDO is capable of 2P optogenetic photostimulation of several neurons at once with cellular resolution. By optimising excitation and collection efficiencies, 2P-FENDO performs 2P functional imaging at one of the highest speeds so far demonstrated through an endoscope. Key novelty behind these results is the discovery that the fiber bundle, composed of ~15000 individual fiber cores, acts as a temporal multiplexing device, separating the laser pulses from each core in time of the right amount to avoid out of focus 2P fluorescence. This property results in a good axial resolution (< 15 µm) independently of the laser spot size. Proof-of-principle experiments were performed in head restrained and freely moving mice co-expressing jGcaMP7s and the opsin ChRmine in the visual or barrel cortex. On a field of view of 250 µm in diameter, we demonstrated functional imaging at a frame rate of up to 100 Hz and precise photostimulation of single and multiple cells (up to ~ 15, limited by the laser power, which could readily be increased with a more powerful photostimulation laser). With the capability to simultaneously image and control neuronal activity at single-cell resolution in freely moving animals, 2P-FENDO will enable to precisely define the functions of neurons in the brain and their interactions during naturalistic behaviours.

Brain responsiveness and its activation complexity are linked to the level of consciousness (Tononi et al., 2004). However, how these features change across brain states is still not clear. The combination of Transcranial Magnetic Stimulation and hd-EEG recordings represents the standard method to address this issue in humans. A preclinical analogous in lab animals would provide novel mechanistic insights on the brain-state-dependent complexity of the brain. A powerful technique to study mesoscale cortical connectivity in mice exploits wide-field fluorescence microscopy. This approach provides simultaneous information of neuronal ensemble activity from distributed cortical areas, while optogenetic has been demonstrated to be a powerful tool to activate cortical neuronal clusters. However, all-optical systems that combine these techniques critically suffer from crosstalk between imaging and photostimulation (Emiliani et al. 2015). Brain responsiveness and its activation complexity are linked to the level of consciousness (Tononi et al., 2004). However, how these features change across brain states is still not clear. The combination of Transcranial Magnetic Stimulation and hd-EEG recordings represents the standard method to address this issue in humans. A preclinical analogous in lab animals would provide novel mechanistic insights on the brain-state-dependent complexity of the brain. A powerful technique to study mesoscale cortical connectivity in mice exploits wide-field fluorescence microscopy. This approach provides simultaneous information of neuronal ensemble activity from distributed cortical areas, while optogenetic has been demonstrated to be a powerful tool to activate cortical neuronal clusters. However, all-optical systems that combine these techniques critically suffer for crosstalk between imaging and photostimulation (Emiliani et al. 2015). Here we established an all-optical method combining wide-field fluorescence imaging of the red-shifted calcium indicator jRCaMP1b and transcranial optogenetic stimulation of Channelrhodopsin-2 (ChR2). To achieve a cortex-wide expression of the calcium indicator, an adeno-associated virus (AAV.PHP.eb) carrying jRCaMP1b under the control of the synapsin promoter was injected in the retro-orbital sinus of anesthetized mice. This led to a uniform expression of the functional indicator in the whole cortex, giving the possibility to visualize the neuronal activity propagation in all the cortical areas. Due to the high opsins expression required for effective optogenetic stimulation, AAV9-ChR2 was locally injected in the somatosensory cortex (S1). Results show that in awake mice, optogenetic stimulations at increasing laser power evoke a distributed cortical response in several areas in the two cortical hemispheres, whereas, during anesthesia, stimulation led to a localized reponse limited in space and time. These results suggest that response complexity decrease with the levels of consciousness, as observed in pathological patients affected by disorders of consciousness (Massimini et al., 2009).

To understand the brain computation paradigms and the causal interactions in complex neuronal networks, we need methods and technologies to record and perturb neuronal distributions over large fields of view. In this application, two-photon (2P) imaging has become a cornerstone microscopy technique, widely used for deep optical access in biological samples and selective light targeting with submicrometric resolution. In parallel to structural and functional imaging, 2P optogenetics has represented a game-changer, allowing targeted stimulation of specific neural circuits. However, the long commutation times and refresh rates of traditional scanning methods substantially hinder near-simultaneous multi-site 3D stimulation. Acousto-optic deflectors (AODs), owing to their fastest scanning and refresh rates, can fulfil the temporal requirements for concurrent activation of sparsely distributed neurons. Nevertheless, their applicability to 2P optogenetics in large volumes has been limited so far by the massive efficiency drop along the optical axis during their use in axial scanning. To counteract this drawback, a compensation software module is frequently employed to flatten the power distribution throughout the volume. However, the power threshold is reduced to the minimum intensity value addressable, lowering the peak intensity released in the centre of the axial scan. Here, we propose a unique approach for overcoming this drawback which provided lifted axial power distribution while maintaining a uniform lateral illumination range. We tested this method by the 2P photoactivation of optogenetic actuators in 3D in zebrafish larvae, showing how the probability of evoking an electrophysiological response and the relative neuronal activity amplitude improved by carefully optimizing the light targeting time on different axial planes. In conclusion, fast and uniform axial light addressing with AODs enables unprecedented 3D 2P optostimulation, formerly not feasible. Furthermore, this approach can be adopted as an upgrade for existing microscopes designed for volumetric imaging, providing 3D multi-site imaging and random-access illumination.

Fluorescence microscopy has succeeded in attaining super-resolution localization of single emitters in cellular biology. However, 3D localization deep inside tissue is still challenging. A few years ago, we developed SELFI: self-interference 3D super-resolution microscopy, a framework for 3D single-molecule localization within multicellular specimens and tissues. Here, we extend the capability of SELFI to the near-infrared (NIR) region where carbon nanotubes (CNTs) are strong emitters. The aim of this work is to develop NIR SELFI for single-particle tracking applications of CNTs in live brain tissues or NIR quantum dots. SELFI uses a diffraction grating placed on the optical path of the sample image, generating an interference pattern within diffraction limited images of point emitters. A single image obtained with NIR SELFI contains two independent variables: the intensity distribution to extract the intensity centroid to determine the lateral localization, and the wavefront curvature (provided by the interfringes) to get the axial super-localization. SELFI was first developed to localize red emitting dyes and quantum dots. The performance of the system is examined by means of the standard deviation and root mean square error of the localizations. The experiments performed show that the 3D-precision and accuracy achieved with NIR SELFI are both below 100 nm for emission around 1000 nm and high photon budget. Therefore, we can now achieve 3D localization in the NIR, permitting 3D single-particle tracking of CNTs at video rate in complex environments.

Neurodegenerative diseases such as Alzheimer’s disease present abnormalities in intraneuronal transport, suggesting the relevance of measuring this key biological process. In 2017, a sensitive method to measure changes in intraneuronal endosomal transport has been reported in 2D cultures of neurons using fluorescent nanodiamonds (fNDs) [1]. The high brightness, photostability and absence of cytotoxicity allow fNDs to be tracked with 50 nm spatial and 50 ms time resolutions. This nanoparticle tracking based-approach applies also to multiphoton imaging, opening the possibility of transport measurement in vivo. We use nanocrystals possessing a large nonlinear second order optical response. First results indicate that the intraneuronal transport measurement can be inferred from nonlinear microscopy data, opening applications to thicker samples owing to the low background of multiphoton imaging. In order to get a high spatio-temporal resolution (around 10 nanometers at 1 ms), we are developing a two-photon microscope, based on a digital holography method [2]. A Digital Micromirror Device (DMD) is used as a spatial light modulator, allowing a fast 3D motion of the excitation volume. We aim at reaching a time resolution below the millisecond and super-localization regime in the tens nanometer range using orbital tracking. References : [1] S. Haziza, et al. Nat. Nanotechnol. 12 (2017), 322. [2] Geng, Q., Gu, C., Cheng, J. & Chen, S. Digital micromirror device-based two-photon microscopy for three-dimensional and random-access imaging. Optica 4, 674 (2017)

Spinal cord injuries (SCI) affect between 2.5 and 4 million patients worldwide, with no current curative treatment. To understand the mechanisms underlying the absence of spontaneous regeneration following injury, we are combining the non-linear multiphoton microscopy (MPM) technique with force measurements via atomic force microscopy (AFM), in a mouse model, to monitor the glial scar, a scar that inhibits the axonal regeneration by forming a physical and chemical barrier composed mainly of astrocytes and microglia. We recorded 2-photon excited fluorescence (2PEF) and second harmonic generation (SHG) signals of excised mice SC injured tissues in MPM at 72h, 1week and 6 weeks post-lesion, and further performed polarization dependent measurements of the SHG signal to assess the preferential orientation of the collagen bundles. Our MPM images revealed a strong SHG signal at 1 week post injury, due to the formation of fibrillary collagen fibers (collagen type I) by the injury site. The SHG signal was increased at 6 weeks after injury, and associated with (1) a higher fiber density (2) a shorter fiber length and less fibers oriented in the same direction. AFM based force spectroscopy measurements, performed at the same post-lesion time-points to map the elastic properties of the spared grey and white matters and injured (lesion) parts of the tissue, suggested an increase of the lesion area stiffness over time. These results together indicate the presence of a fibrotic process seven days after injury, that is further increased at later time points. We similarly started to investigate the effect of a treatment (pharmacological transient depletion of microglia/macrophage proliferation) in mice that underwent SCI. Our preliminary results suggested an increase in fibers length in treated tissues, as well as a reduction of the collagen extension around the injury site.

We present our recent study combined multi-armed Bandits algorithm in reinforcement learning with spontaneous Raman microscope for the acceleration of the measurements by designing and generating optimal illumination pattern “on the fly” during the measurements while keeping the accuracy of diagnosis. We present our simulation and experimental studies using Raman images in the diagnosis of follicular thyroid carcinoma and non-alcoholic fatty liver disease, and show that this protocol can accelerate more than a few tens times in speedy and accurate diagnoses faster than line-scanning Raman microscope that requires the full detailed scanning over all pixels. The on-the-fly Raman image microscopy designs to accelerate measurements by combining one of reinforcement machine learning techniques, bandit algorithm utilized in the Monte Carlo tree search in alpha-GO, and a programmable illumination system. Given a descriptor based on Raman signals to quantify the likelihood of the predefined quantity to be evaluated, e.g., the degree of cancers, the on-the-fly Raman image microscopy evaluates the upper and lower confidence bounds in addition to the sample average of that quantity based on finite point/line illuminations, and then the bandit algorithm feedbacks the desired illumination pattern to accelerate the detection of the anomaly, during the measurement to the microscope. Most conventional bandit algorithms assume that infinite number of measurements or samples provides us with 100% accuracy. However, in Raman measurements we should develop both a Raman descriptor to quantify the degree of anomaly, and a new algorithm to take into account the finite accuracy lower than 100%. This microscope can also be applied to other problems, besides detection of cancer cells, such as anomaly or defects of materials. The algorithm itself is also beneficial and transferrable to the other microscopes such as infrared image microscope.

High-speed Raman spectroscopy has enabled label-free characterization of molecules in cells and materials in a space- and time-resolved manner. Among these, time-domain Raman spectroscopy (TDRS) techniques, such as Fourier-transform coherent anti-Stokes Raman scattering (FT-CARS) and impulsive stimulated Raman scattering (ISRS) spectroscopies, have unique capabilities such as high spectral acquisition rates, broadband spectral sensitivity in the fingerprint region, and nonresonant-background-free spectral acquisition. With a few exceptions, most TDRS studies have focused only on the fingerprint region (200 – 1800 cm-1) because the ultrashort pulses typically used for ultrabroadband (200 – 3200 cm-1) spectral acquisitions are difficult to generate and handle. For example, detecting Raman peaks above 3000 cm-1 necessitates a pulse duration of < 10 fs, which demands an expensive laser source and careful dispersion control. Furthermore, with sub-10-fs pulses, Raman detection sensitivity in the fingerprint is compromised because the spectral power density is diluted in the spectrally broad ultrashort pulse. The present research demonstrates FT-CARS spectroscopy covering both the fingerprint and CH-stretching regions by employing synchronized mode-locked Ti:Sapphire and Yb-doped fiber lasers as the light source. With this method, we show that ultra-broadband FT-CARS spectra can be obtained without using sub-10-fs pulses, which significantly mitigates experimental complexity. More importantly, ultra-broadband Raman detection can be achieved in this scheme without compromising the sensitivity in the fingerprint region, unlike previous ultrashort-pulse approaches. The present method will significantly broaden the application range of TDRS for biomedical and material science research.

Discoid Lupus Erythematosus is chronic autoimmune disease that disproportionately affects parts of the population. Presently, the biochemical events involved in the formation of the disease and elements of the pathophysiology are poorly understood, demonstrating a need for improved analysis. We present the results of our multimodal imaging combining Raman spectroscopy and mass spectrometry and their chemometric analysis and models. We distinguish physiological features and how they differ between healthy and DLE. We show that by fusing the data we are able to build a classification model that can differentiate the two with higher accuracy than either technique alone. The findings from this study can serve as a basis for improved biomedical diagnostics and better informed potential treatment options.

Stimulated Raman Scattering (SRS) microscopy is a Coherent Raman technique that emerged in recent years as a powerful tool for biomedical imaging due to its high specificity, high speed, and label-free capability. Despite its advantages, SRS microscopy can be affected by nonlinear competing phenomena, namely two-photon absorption (TPA), cross-phase modulation (XPM), and thermal lensing (TL), which generate a background signal that reduces the achievable specificity and sensitivity. These competing processes are quasi-instantaneous and spatially non-uniform in heterogeneous samples and require customized setups to be canceled in SRS acquisitions. A robust approach for background-free SRS measurements is the frequency-modulation (FM) SRS, which is based on the broad spectral dependence of the parasitic effects (typically tens of nanometers) compared to the narrower band of the SRS effect (~ 1 nm). Performing a differential measurement at two different wavenumbers, respectively on- and off- Raman resonance, it is possible to selectively detect the SRS process. Different solutions for background cancellation via FM have been reported in the literature, but they present various drawbacks, such as the limited applicability over certain ranges of the vibrational spectrum or the necessity to modify the optical setup when performing measurements at different Raman shifts. We propose an FM-SRS configuration realized for the first time with an acousto-optic tunable filter, able to perform measurements from the fingerprint to the CH-stretch region of the spectrum without any modification of the optical setup. We determined its efficiency in canceling the background signal due to different types of competing effects on various samples: polymer beads, human hair, and human cells. These results underline the importance of an effective cancellation of background signals of diverse nature when collecting SRS images. Our FM-SRS setup demonstrated critical advantages compared to other FM configurations.

We present our recent developments in utilizing a compact and portable light source for high-speed multicolor stimulated Raman scattering imaging (SRS) in biomedical and medical environments. The source combines a rapid and wide tunability for accessing Raman bands between 700 and 3300 cm-1 with high stability in terms of power (deviation < 0.3 %) and wavelength (deviation < 0.5 pm) over more than 100 h. We highlight applications in metabolic cell imaging and the identification of pharmaceuticals in complex environments such as skin by harvesting contrast from several Raman bands.

Total Internal Reflection Fluorescence Microscopy (TIRFM) exploits an evanescent field induced at the boundary between high and low refractive index media to selectively excite the sample inside a very thin region (from 100 to 300 nm depending on the illumination angle) above the coverslip surface. The minimum exposure of the sample to light above the excitation slice reduces significantly the out-of-focus fluorescence and phototoxicity which are major issues in live-cell imaging. It has become an indispensable tool in biology, in particular to study the molecular traffic at the cell plasma membranes. However, in many applications, the lateral resolution of TIRF, which is diffraction limited to about 300 nm, is not sufficient. In addition, the optical sectioning of the evanescent illumination of TIRF is seldom perfect. Propagative waves stemming from imperfections in the optical train of the instrument and/or light scattering by the sample itself are able to excite the fluorescence in the volume of the sample. When the latter is densely marked, these leaks result in out-of-focus fluorescence which deteriorates the signal to noise ratio. To improve simultaneously the lateral resolution and the image contrast, and to address the difficulties related to the control of the illumination patterns, we propose to adapt the recently developed Random Illumination Microscope (RIM) to the TIR configuration. We show that this approach yields a two-fold resolution gain and ameliorates the image contrast without compromising the ease of use of standard TIRFM. We apply TIRF-RIM to calibrated targets and to fixed and live biological samples with a sub-100nm resolution.

This conference presentation, “Axially-swept adaptive optics light sheet fluorescence microscopy for high resolution neuroimaging in the drosophila brain” was recorded for the Biomedical Spectroscopy, Microscopy, and Imaging II conference at SPIE Photonics Europe 2022.

Endoscopes with 3D imaging have been known for some time. Additional information about the depth allows further statements during the examination and better visualisation. However, the measuring heads of such endoscopes are bulky, since they must accommodate optics that enable axial scanning. Hence, the field of application remains very limited. We present an endoscope without optics on the distal end and a significantly smaller measuring head in the sub-millimetre range. This enables endoscope technology to be used in new areas of surgery, such as in brain or cochlea. Conventional endoscopes are too large for these regions. A static phase correction has been demonstrated to be sufficient to maintain phase information. Hence, programmable optics like spatial light modulators are no longer needed. Therefore, we applied 3D printed phase masks using 2-photon polymerisation. This allows a robust and cost-efficient system to be realised. In addition to the process of printing phase correction DOEs, we also present a new setup which allows the sample in front of the endoscope head to be imaged through the fibre bundle directly to a camera sensor. No raster scan is required like in past approaches. Hence, an image can be generated in a single shot without further computational reconstruction.

The retina is an epithelium composed of different cell layers with unique optical properties and detects light by photoreceptor neurons for visual function. The quest for suitable measurement methods to detect the health status of retinal tissues is ongoing. We study the capability of the optical transmission matrix, which fully describes the transition of a light field propagating through a scattering sample. Despite its rich information content, the transmission matrix is commonly just used for light delivery through scattering media. Digital holography is employed to measure the transmitted light. We demonstrate that singular value decomposition of the transmission matrix allows to discriminate phantom tissues with varying scattering coefficient. We apply these findings to retinal organoid tissues. Application of an inducer of retinal damage in animals, caused cell death and structural changes in human retinal organoids, which resulted in distinct changes in the transmission matrix. Our data indicate that the analysis of the transmis-sion matrix can distinguish pathologic changes of the retina towards the development of imag-ing-based biomarkers. Laser microscopy of retinal organoid samples from human induced plu-ripotent stem cells is a disruptive technology that promises paradigm shifts for biomedicine.

Optical imaging techniques that provide free space, label free imaging are powerful tools in obtaining structural and biochemical information in biological samples. To date, most of the optical imaging technologies create images with a specific contrast and require multimodality integration to add additional contrast. In this study, we demonstrate spectroscopic Thermo-elastic Optical Coherence Tomography (TE-OCT) as a potential tool in tissue identification. TE-OCT creates images based on two different forms of contrast: optical reflectance and thermo-elastic deformation. TE-OCT uses short laser pulses to induce thermo-elastic tissue deformation and measures the resulting surface displacement using phase-sensitive OCT. In this work we characterized the relation between thermo-elastic displacement and optical absorption, excitation, fluence and illumination area. The experimental results were validated with a 2-dimensional analytical model. Using spectroscopic TE-OCT, the thermo-elastic spectra of elastic phantoms and tissue components in coronary arteries were extracted. Specific tissue components, particularly lipid, an important biomarker for identifying atherosclerotic lesions, can be identified in the TE-OCT spectral response. As a label-free, free-space, dual-contrast, all-optical imaging technique, spectroscopic TE-OCT holds promise for biomedical research and clinical pathology diagnosis.

Extracellular matrix (ECM) has important functions in cell proliferation, differentiation, and migration, which influence the development and progression of cancer. ECM in tumor microenvironment experiences changes in composition and structure that can appear early in tumor development and could serve as a biomarker for cancer diagnostics. In addition, some changes in ECM may correlate with the rate of tumor progression or its tendency to form metastases and would allow to predict future tumor development [1]. Collagen is an important structural protein found in ECM. It has a non-centrosymmetric structure, and, thus, can be easily visualized using second harmonic generation (SHG) microscopy. SHG microscopy employs certain polarimetric techniques to gain detailed information about the organization of collagen in various tissues [2]. In this work, polarimetric SHG microscopy is used to acquire collagen images from normal and cancerous regions of human colon and pancreas histological samples. Texture analysis is performed on SHG intensity and polarization images to characterize the distribution of ultrastructure parameters in the tissue. Significant differences are observed in collagen ultrastructure between normal and tumor areas. Further, collagen structures of colon and pancreas tumor microenvironments are compared to investigate relative differences in ECM organization between the tissues. Finally, a machine learning classifier is used to group the acquired images in tumor and normal groups. The results show potential for development of novel cancer diagnostic technique using polarimetric second harmonic generation microscopy and texture analysis. [1] Winkler, J. et al., “Concepts of extracellular matrix remodelling in tumour progression and metastasis”, Nat Commun 11, 5120 (2020). [2] Golaraei, A. et al., “Polarimetric second-harmonic generation microscopy of the hierarchical structure of collagen in stage I-III non-small cell lung carcinoma,” Biomed. Opt. Express 11, 1851-1863 (2020).

Development and metastasis of cancer are known to change the structure of extracellular matrix (ECM), which affect the tumor's further growth and spread. A substantial part of ECM is comprised of collagen, which is a noncentrosymmetric structure. As a result, it generates second harmonic signals, dependent on the polarization of incoming light. This property of collagen led to the applications of polarization-resolved second-harmonic generation (P-SHG) microscopy in investigating collagen ultrastructure changes in different cancers. In this work, multiphoton absorption fluorescence (MPF), third-harmonic generation (THG) and polarimetric second-harmonic generation (P-SHG) measurements were performed on various types and staging of human melanoma histological sections. Reduced polarimetry techniques, employing linear and circular polarization states, were used to obtain polarimetric SHG parameters of collagen in both normal and cancerous tissues. These parameters provide important information about the structural properties of collagen. The parameter distributions were analyzed using a grey-level co-occurrence matrix (GLCM), which allows to obtain statistical parameters, such as correlation, contrast, entropy, angular second moment and inverse difference moment. Statistical tests were performed on polarimetric and texture analysis data in order to determine whether parameter distribution differences in normal and cancerous tissues are statistically significant. Furthermore, a machine learning classifier algorithm was trained to distinguish normal tissues from cancerous using aforementioned polarimetric and texture parameters as predictors. Firstly, separate training and testing datasets were formed from each sample and classification was carried out for each of them individually and afterwards, a common training dataset was used for all samples. The results suggest that normal and cancerous skin tissues can be distinguished from each other with the help of multimodal nonlinear polarimetric microscopy. Also, depending on the type and stage of melanoma, the differences in some polarimetric and texture parameters are more pronounced, suggesting its possible application in melanoma diagnostics and differentiation.

GaAsBi has attracted research for near-infrared (NIR) optoelectronics because bismuth incorporation causes a far greater band gap reduction per unit strain than indium incorporation. The bismuth atoms induce the formation of many localised states near the valence band maximum, which can take part in radiative transitions and result in a large broadening of the luminescence spectrum. The large linewidths observed in GaAsBi are typically seen as a disadvantage of the material and researchers have focussed on reducing the density of localised states. Superluminescent light emitting diodes with peak emission centred around 1050 nm are useful for ophthalmology applications such as OCT since these wavelengths are less strongly absorbed by ocular media. In this case, a large LED spectral linewidth leads to an improved axial resolution in OCT, enabling better imaging and subsequent analysis by doctors. Commercial LED based OCT light sources operating at 1050nm rely on emission from both ground and excited states in InGaAs quantum wells and have a linewidth around 70nm. State-of-the-art OCT light sources based on multiple layers of InAs self-assembled quantum dots have achieved linewidths of 160nm. Existing unoptimised GaAsBi single quantum well structures grown in our group by molecular beam epitaxy with a peak wavelength of 1050nm have a spectral linewidth of around 67 nm, which nearly matches the commercial LEDs used for OCT. This is despite our devices only containing emission from the ground state in the quantum wells. With careful control of the bismuth content and well thickness in future devices, the linewidth of GaAsBi based devices could match or exceed the state-of-the-art for NIR broadband light sources. In this work we study the applicability of GaAsBi quantum well LEDs as a competitor to InGaAs quantum well and InAs quantum dot LEDs for broadband NIR light sources. We show simulations of LED structures to find the optimum LED design parameters that will give the broadest linewidth centred on 1050nm while retaining an approximately Gaussian emission shape. The growth challenges associated with growing the structures are also discussed.

Multiphoton microscopy is a very important and powerful technique in the application of in-vivo drosophila brain imaging. However, the photon budget, which is compromising between imaging speed, spatial resolution and accumulating time, remains a huge challenge in biological compatibility. Rapid dual-resonant volumetric multiphoton microscopy combines a tunable acoustic gradient (TAG) lens and a resonant mirror, which can achieve up to 8 kHz frame rate and hundreds of hertz volumes per second temporal resolution. Adaptively sampling each laser pulse by an embedded field programmable gate array (FPGA) with up to 80 MHz pixel rate enables efficient but restricted signal accumulation. This study has developed a generative neural network to restore images from degradation of low signal-to-noise ratio (SNR), missing pixels and pattern residual. A series of training data by adding noise and Lissajous scanning path into the known fluorescent bead images were used to pretrain the model for the rapid dual resonant volumetric multiphoton microscopy system. Experimental results verify that the axial distortion and the image resolution of noisy fluorescent bead images can be effectively restored with the deep-restoration neural network. The mushroom body (MB) of drosophila brain which contains thousands of Kenyon cells in a 200 × 200 × 100 µm3 volume is utilized to demonstrate the strategy. The deep-restoration rapid dual resonant volumetric multiphoton microscopy image not only maintains 256 × 256 × 128 voxels and ~30 volumes per second, but also significantly improves image quality which is compatible to the ground truth. However, in-vivo imaging is time-varying. Base on in-vitro image datasets, in-vivo images via transfer learning is utilized to enables fast and improved image quality efficiently. Despite of one pulse per pixel, in-vivo drosophila brain imaging with deep-restoration not only keeps the advantage of temporal resolution, but also obtains well image quality.

Optical microscopy allows to perform structural and functional imaging within large volume of tissues with subcellular resolution. Non-linear microscopy allows the interrogation of neuronal activity in mammalian brains but remains limited because of scattering and optical aberrations. To overcome these issues, Adaptive Optics (AO) strategies have been implemented to retrieve the microscope imaging quality while addressing important imaging depths. A first AO strategy implemented in non-linear microscopy relies on a sensorless configuration, but is a time-consuming iterative process hardly compatible with photobleaching issues. A second approach is based on direct wavefront sensing using Shack-Hartmann wavefront sensors and has proved its efficiency on in vivo experiments. However, this method fails at large depths because of the strong scattering of the emitted fluorescence. A method for direct wavefront sensing more resilient to scattering of the fluorescence emission would therefore facilitate the use of AO in optical microscopy. This work proposes an alternative method of direct wavefront measurement, which relies on the cross-correlation of images of an extended source obtained through a microlens array. This extended-source Shack-Hartmann wavefront sensor (ESSH) requires to be coupled to an optical sectioning method. Its efficiency has been proven when coupled to Light Sheet Fluorescence Microscopy in the adult drosophila brain in weekly scattering conditions. Here, we show that it allows quantitative aberration measurements through highly scattering fixed brain slices, up to four times the scattering length of the tissue. We demonstrate that it is more resilient to scattering compared to the current centroid-based approach. Taking advantage of its geometry, this new wavefront sensor also provides scattering coefficient measurements of biological tissues. Finally, we present its implementation on a two-photon microscope within a closed–loop configuration for in depth neuroimaging in mouse brain and compare its performances in scattering media to the classical centroid approach.

Trough intravital functional optical coherence tomography (OCT) in mouse models, this study investigates physiological processes within the fallopian tube in vivo. The transport of oocytes and embryos through the oviduct (fallopian tube) is a fundamental reproductive processes of clinical importance. However, because mammalian fertilization and embryogenesis take place deep within the female body, these processes are hidden from direct observation. Therefore, much of what we know about the innerworkings of the female reproductive tract is extrapolated from in vitro and ex vivo experimental settings and does not necessarily represent the native state, limiting success in management of reproductive disorders. This study presents first in vivo volumetric dynamic imaging of oocytes and embryos as they are transported through the mouse oviduct. By implementation of new functional OCT methods, we established methods for tracking oviductal ciliary function and individual sperm movements. Supported by dynamic volumetric visualizations, the study reveals a variety of intriguing never-before-seen dynamic behaviors and suggest new regulatory mechanisms driving reproductive processes.

Polarimetric second harmonic generation (SHG) microscopy techniques are powerful tools to reveal sub-molecular information from biological specimens. Among biological samples collagen with a noncentrosymmetric structure and efficient SHG conversion has been the focus of many studies. Since collagen remodeling takes place due to cancer progression, it is important to develop tools to detect and understand the ultrastructural changes in collagen assembly using polarimetric nonlinear microscopy. Several polarimetric techniques have been developed to probe susceptibility ratios, in-plane orientation, and out of the image plane orientation of collagen. Polarization-In Polarization-Out (PIPO) and SHG circular dichroism (SHG-CD) techniques have been used to calculate the out of the image plane orientation and chirality of collagen. In this work, we study the correlation between SHG-CD and the chiral susceptibility ratio (C) in order to reveal the collagen chirality, and the collagen fiber tilt out of image plane. A numerical modeling is used to understand the relation between aforementioned parameters and the chirality and out of the image plane orientation of collagen. The results of numerical modeling show similar behaviors for SHG-CD and the chiral susceptibility ratio (C) calculated from PIPO measurements. The results obtained from rat tail tendon collagen confirms that the sign of both SHG-CD and C ratio changes by flipping the sample as it is predicted by the numerical modeling. The results also show that both SHG-CD and C ratio may become miscalculated when antiparallel chiral fibers are present in the focal volume of the microscope. The results of this study confirm that polarimetric SHG microscopy techniques are able to reveal 3D structure of biological samples and therefore they are beneficial to the diagnosis of collagen related diseases.

This conference presentation, “3D molecular phenotyping of the human brain Broca’s area using light-sheet fluorescence microscopy” was prepared for the Biomedical Spectroscopy, Microscopy, and Imaging II conference at SPIE Photonics Europe 2022.

This conference presentation, “Monitoring dynamic collagen reorganization in human dermis during uniaxial stretching with second/third harmonic generation microscopy” is part of the Biomedical Spectroscopy, Microscopy, and Imaging II conference at SPIE Photonics Europe 2022.

This conference presentation, “Label-free stimulated Raman scattering imaging utilized for correlating silicone content in breast tissue with capsular contracture in an intra-patient study” was prepared for the Biomedical Spectroscopy, Microscopy, and Imaging II conference at SPIE Photonics Europe 2022.

This conference presentation, “Stimulated Raman scattering simulation for imaging optimization” was prepared for the Biomedical Spectroscopy, Microscopy, and Imaging II conference at SPIE Photonics Europe 2022.

This conference presentation, “Instant histopathological assessment of fresh lung biopsies taken for diagnosing interstitial lung diseases using third and second harmonic generation microscopy” was recorded for the Biomedical Spectroscopy, Microscopy, and Imaging II conference at SPIE Photonics Europe 2022.