This PDF file contains the front matter associated with SPIE Proceedings Volume 11248, including the Title Page, Copyright information, Table of Contents, Author and Conference Committee lists.

To visualize neuronal structure and function in the physiological context, optical microscopy that is non-invasive and capable of resolving sub-cellular structures has become the method of choice. Structured illumination microscopy (SIM) is a widefield fluorescence imaging technique that optically sections 3D samples, but its applications have been usually limited to in vitro samples. To apply SIM to in vivo imaging, we modified optical-sectioning SIM reconstruction algorithm and incorporated adaptive optics. We demonstrated fast, high-resolution in vivo imaging with optical sectioning for structural and functional interrogations of the brain in vivo.

An enabling technology for the monitoring of neural activity, multiphoton microscopy with Bessel focus scanning is a high-speed volumetric imaging method with subcellular lateral resolution. Similar to many other optical microscopy techniques, however, its axially extended Bessel-like focus experiences sample-induced optical aberrations, which lead to reduced resolution and image contrast at depth. In this study, we demonstrated an adaptive optical Bessel focus scanning multiphoton microscope with pupil-segmentation-based wavefront sensing and highly efficient sample-plane aberration correction. Applying it to mouse brain imaging in vivo, we observed up to threefold signal enhancement for functional spine imaging at depth.

In optical imaging, light propagation is affected by the medium inhomogeneities. Adaptive optics has been employed to compensate for sample-induced aberrations but the field-of-view is often limited to a single isoplanatic patch. Here, we propose a non-invasive approach based on the distortion matrix concept. This matrix basically connects any focusing point with the distorted part of its wave-front in reflection. Its time-reversal and entropy analysis allows to correct for high-order aberrations over multiple isoplanatic areas. We demonstrate a Strehl ratio enhancement up to 2500 and a diffraction-limited resolution until a depth of ten scattering mean free paths through biological tissues.

We demonstrate Lattice Light-Sheet microscopy combined with dual-lens Fresnel Incoherent Correlation Holography (LLS-FINCH) in the detection path. We will describe the system alignment steps, provide an experimental demonstration, and provide the system characteristics in comparison to those of conventional LLS (glass-optics based). We present the imaging performance of the microscope and the adaptive optics module with an example of post-data-acquisition image reconstruction which includes an adaptive optics correction process.

Light Sheet Microscopy has developed rapidly over the past decade and is the ideal approach for imaging model organisms such as zebrafish and other thick tissue specimens. Despite the superior optical sectioning capability, high imaging speed, and large field of view, the performance of light sheet microscopy still suffers from optical aberrations. We have implemented a scene-based Shack-Hartmann wavefront sensor for directly measuring the optical aberrations on the emission side of the light-sheet microscope. In this work, we show that our system is capable of AO correction using sensor based and sensorless based approaches. We demonstrate correction up-to one hundred microns deep in zebrafish and fruitfly embryos.

Adaptive optics normally concerns correction of phase aberrations; this has certainly been the case in microscopy. However, the performance of certain forms of microscope is also sensitive to polarization errors. We present the concept of vectorial adaptive optics (V-AO) to extend compensation techniques into the vectorial beam domain by including polarization. Full polarization and phase control are implemented based on a combined spatial light modulator (SLM) and deformable mirror (DM) system. We demonstrate that V-AO can manipulate/improve the point spread function (PSF) of microscopes and micro-endoscopes. We explore methods of control of such systems to enable feedback correction of polarization and/or phase aberrations. Widely-used high numerical aperture (NA) objective lenses and GRIN rod endoscope lenses based systems are used for testing. Comparisons of the performance using V-AO and without V-AO are demonstrated alongside a specific case of GRIN rod endoscope lenses based vectorial imaging. The enhancement of the normal/vectorial imaging of the samples validated that our technique may solve numerous residual optics error issues in various microscopic/endoscopic systems. This may pave the way for further application directions. It should be noted that the applications of V-AO are not confined to microscopes, may also find use such other areas of optics.

When imaging a sample, inhomogeneities in refractive index cause blur in the image and decrease resolution. Adaptive optics (AO) is a technique that can correct for the resulting aberrations. The most common implementation of AO uses a single deformable mirror that is conjugate to the pupil. A single pupil-conjugate corrective device provides correction over a limited field of view owing to field-dependent aberrations. To overcome this limitation, an additional specimen-conjugate deformable mirror can be used. However, adding a second reflective correction device significantly increases system complexity. We have developed a closed-loop multiconjugate AO system for field-dependent aberration correction in a confocal fluorescence microscope. A 140-actuator deformable mirror is used in the pupil plane with a custom 37-element transmissive deformable phase plate inserted in a sample-conjugate plane. Both devices are calibrated and controlled in closed-loop using a Shack-Hartmann sensor in combination with an integral control law. The sensor consists of an EMCCD and lenslet array with a 500 μm pitch and a 47 mm focal length. Results from a Drosophila ovary and HeLa cells are presented.

I will discuss our efforts in developing computational microscopy techniques that can provide improved robustness and scalability to multiple scattering problems. First, I will discuss a new model for quantifying the effects of anisotropic scattering on image quality degradation. In particular, I will illustrate how to quantitatively relate the macroscopic scattering properties to the microscopic parameters used in the model. Next, I will discuss a deep learning approach to invert the effect of scattering. Particular emphasis will be placed on the scalability of this approach and how the model can be generalizable to different objects/media by extracting statistically invariant information.

Multimode fiber (MMF) endoscopy offers high spatial resolution in an ultra-small form factor, yet several technical challenges remain to be addressed to realize practical MMF imaging. Here, we demonstrate a strategy for confocal reflectance imaging through MMF to improve contrast owing to optical sectioning and enable volumetric imaging. Instead of physically focusing the light into the sample, it uses a series of distinct illumination patterns obtained by varying the proximal coupling of the illumination. This alleviates the need for active wave-control and translates into a speed advantage critical for practical MMF imaging.

Structured illumination microscopy has been used with great success on single cell samples. However, it cannot be applied to multicellular thick samples because aberrations caused by the sample not only reduce the intrinsic resolution and contrast, they cause the SIM reconstruction to fail. Although the structured illumination pattern can still be projected onto the focal plane inside the sample when imaging deep into thick biological samples, the resulting raw fluorescent images are corrupted resulting in poor reconstructed SIM images with poor signal to noise ratio, degraded spatial resolution and artifacts that can call into question the reliability of the image. To image in thicker samples with diffraction limited resolution, adaptive optics can be used to correct the optical aberrations due to the sample. Here, we combine three-dimensional structured illumination microscopy (3DSIM) and adaptive optics (AO), demonstrating full three-dimensional aberration-free super-resolution imaging deep into thick multicellular samples. We applied a frequency-based metric function in image-based sensorless AO method for aberration corrections. In some cases, we also applied a customized dot-array illumination pattern to optimize the image spectrum. Through the imaging of various samples, we show that the image-based sensorless AO method performs a satisfying and robust correction of different aberrations with minimal photobleaching. The final three-dimensional image achieves a resolution of ~120 nm laterally and ~500 nm axially with optical sectioning, which is a two-fold resolution enhancement without any nonlinear deconvolution methods. AO-3DSIM provides a reliable solution for three-dimensional super-resolution imaging in vivo with improved fidelity.

Sensorless adaptive optics (AO) has been widely used in optical microscopy to improve imaging quality in scattering tissue without additional wavefront sensing devices. The traditional image metric-based sensorless AO method requires multiple frames to assess aberrated wavefront, which is time consuming and even inaccurate when the aberration becomes large due to distortion mode crosstalk. Here we propose a neural network based wavefront sensing method which can accurately predict wavefront distortions across different aberration scales in a single-shot. Compared to the traditional method, the neural network approach reduces the prediction time by over one thousand folds. We validate the superior performances of neural network-based approach in both accuracy and speed through numerical simulations.

Standard imaging systems provide a spatial resolution that is ultimately dictated by the numerical aperture. In biological tissues, the resolution degraded by scattering which limits the imaging at a depth. Here, we exploit the properties of speckle patterns embedded into a strongly scattering matrix to illuminate the sample at high spatial frequency content. Combining adaptive optics with a custom deconvolution algorithm, we obtain a resolution improvement of a factor < 2.5. Our Scattering Assisted Imaging (SAI, M. Leonetti et al., Sci. Rep. 9:4591 (2019)) provides an effective solution to increase the resolution when long working distance optics are needed.

We demonstrate both theoretically and experimentally how to achieve wave states that are optimal for transferring momentum, torque, etc. on a target of arbitrary shape embedded in an arbitrary environment.

Scattering media scramble light paths, create seemingly random speckle patterns and hinder even our simple visualization of objects. Here, we demonstrate stochastic optical scattering localization imaging (SOSLI) to achieve super-resolution non-invasively through not only static, but also dynamic scattering media (up to 80% decorrelation). A camera captures multiple speckle patterns created by stochastic emitters in the object. Then our computational approach can retrieve a super-resolution image of hidden objects, surpassing the diffraction limit by factor of five, while posing no fundamental-limit in achieving higher spatial-resolution. Our demonstration paves the way to non-invasively visualize various biological samples with unprecedented levels of detail.

This talk reviews ongoing work towards computational imaging of the living eye, primarily for the purpose of quantifying corneal transparency, for which to date no objective clinical tool exists.

Non-invasive retinal imaging has greatly facilitated the research of eye disease and neurodegenerative disorders in the central nervous system (CNS). Two-photon microscopy is a powerful tool for in vivo imaging of mouse retina because it provides intrinsic optical sectioning capability and the infrared laser is less likely to excite the photoreceptors. However, the dilated mouse eye has large optical aberrations, which must be corrected to achieve high-resolution or even diffraction-limited imaging. Here, we developed an adaptive optics (AO) two-photon microscope for in vivo imaging of retinal neurons through the eyeball of living mouse. We used the two-photon excited fluorescence signal of retina as the guide star to measure and correct the aberration of mouse eye. After AO correction, the fluorescence signal was increased by at least fivefold and the fine structures such as axons of retinal ganglion cells (RGC) were clearly resolved. To take advantage of the non-invasive high-resolution imaging, we demonstrated functional calcium imaging of RGC responding to the light stimulations.

Heterogeneities of biological tissues can strongly affect light propagation at large depths by distorting the initial wavefront. Inspired by previous works in acoustics, we have developed a matrix approach to Full-Field Optical Coherence Tomography (FF-OCT) to push back the fundamental limit of aberrations and multiple scattering. An analysis of the correlations of the matrix allows to correct for aberrations and forward multiple scattering over multiple isoplanatic areas (contrary to classic adaptive optics). Here, we report on the application of this approach to the imaging of the monkey cornea and the quantitative measurement of the corneal transparency.

Blood flow dynamics in the human retina can be assessed by laser Doppler perfusion imaging by digital holography with near-infrared light, but motion and aberrations of the eye can prevent accurate coherent image formation with standard free-space wave propagation methods. We present a fully digital aberration correction scheme in post-processing by tuning a multi-parameter phase mask in the reciprocal plane of the retina, through iterative optimization of the quality of power Doppler images. The parameters are the amplitudes of low-order Zernike polynomials controlling shift/tilt, defocus and astigmatism. Numerical correction for motion and ametropia reveals smaller vessels than standard reconstruction.

The lensless endoscope represents the ultimate limit in miniaturization of imaging tools: an image can be transmitted through a (multi-mode or multi-core) fiber by numerical or physical inversion of the fiber's pre- measured transmission matrix. However, the transmission matrix changes completely with only minute conformational changes of the fiber, which has so far limited lensless endoscopes to fibers that must be kept static. In this work we report a lensless endoscope which is exempt from the requirement of static fiber by designing and employing a custom-designed conformationally invariant fiber. We give experimental and theoretical validations and determine the parameter space over which the invariance is maintained.

We present a fast holographic endoscope with needle footprint. Robustness versus bending is achieved by a single sided self-calibration of the fiber bundle. Incorporating an adaptive lens and 2-axis galvo scanner allows 3D scans with one Mega voxel per second. While the 3D imaging is demonstrated at 3µm fluorescent beads, we also demonstrate the applicability in 2-axis cell rotators by precisely rotating two foci with an angle uncertainty below 1°. For the future, we expect this approach will lead us to create tools for deeper tissue imaging and cell manipulation.

Miniaturisation of endoscopes can be achieved using lensless endoscope probes, which enhances in vivo deep- tissue imaging technology. The necessity of a detailed understanding of light propagation through optical fibres is paramount, since beam focusing and scanning at tissue require beam shaping at the proximal end of the fibre. For stable light delivery and collection, the sensitivity of various fibre profiles against fibre deformations needs to be reviewed. We present a numerical simulation tool investigating optical field propagation through multimode and multicore optical fibres, emphasizing fibre-bending deformations. The simulation tool enables user to choose optimum fibre with best possible realistic parameters for any application.

One limitation of microendoscopy is the device footprint that should be minimal for many applications. Here we present a minimally-invasive endoscope based on a multimode fiber that combines photoacoustic and fluorescence sensing. With the use of a fast spatial-light modulator, it is possible to rapidly learn the transmission matrix during a prior calibration step. A focused spot can then be produced to raster-scan a sample at the distal tip of the fiber. Our setup provides both photoacoustic and fluorescence microscopic images of test samples in vitro (fluorescent beads and red blood cells) through a single multimode fiber.

Specimen induced aberrations can have detrimental effects in all types of high-resolution microscope. In this study, we present a sensorless technique that uses a deformable mirror (DM) to correct aberrations of both the system and sample. Using a laser-free confocal microscope, with patterned disk illumination and detection. The system is based on a commercial confocal module (Clarity, Aurox Ltd., UK) that uses Light Emitting Diode (LED) illumination to obtain optically sectioned 3D images. The results obtained show that the setup was able to correct aberrations of biological samples used in the study. These systems will help researchers working on various biological systems to obtain improved quality images when focussing deep into thick specimens.

We have developed a microscopic adaptive optics (AO) system that corrects wavefront phase errors induced by complex structures of biological samples. The technique of correlation-based Shack-Hartmann (SH) sensing used in the AO system enables wavefront measurement using complex structures in a target as the reference. However, sub-images in the SH sensor become deformed dependently on the positions of sub-apertures as the NA of the microscopic objective is higher. This often deteriorates the accuracy of wavefront sensing. To mitigate the undesirable effect, we here propose a differential wavefront sensing technique with a mathematical formula, which is expected to measure wavefront at a better precision. Because differences in image shapes are less significant between nearby SAs, correlations between adjacent SAs are measured in the proposed method. We confirmed that the AO system worked as designed by experiments.

In optics in general, a sharp aberration-free image is normally the desired goal, and the whole field of adaptive optics has developed with the aim of producing blur-free images. Likewise, in ophthalmic optics we normally aim for a sharp image on the retina. But even with an emmetropic, or well-corrected eye, chromatic and high order aberrations affect the image. We describe two different areas where it is important to take these effects into account and why creating blur correctly via rendering can be advantageous. Firstly we show how rendering chromatic aberration correctly can drive accommodation in the eye and secondly report on matching defocus-l generated using rendering with conventional optical defocus.