Open Access
Add to BibSonomy
Abstract
Light-sheet microscopy has been widely used in high-speed fluorescence imaging with low phototoxicity, while the trade-off between the field-of-view and optical sectioning capability limits its application in large-scale imaging. Although Bessel beam light-sheet microscopy greatly enhances the light-sheet length with the self-healing ability, it suffers from the strong side-lobe effect. To solve these problems, we introduce the photobleaching imprinting technique in Bessel beam light-sheet microscopy. By extracting the non-linear photobleaching-induced fluorescence decay, we get rid of the large concentric side lobe structures of the Bessel beam to achieve uniform isotropic resolution across a large field-of-view for large-scale fluorescence imaging. Both numerical simulations and experimental results on various samples are demonstrated to show our enhanced resolution and contrast over traditional Bessel-beam light-sheet microscopy.
© 2020 Optical Society of America under the terms of the OSA Open Access Publishing Agreement
1. Introduction
Light-sheet fluorescence microscopy (LSFM) has been one of the main techniques for biological imaging in the past decade [1,2]. In LFSM, an oblique or orthogonal illumination plane generated by detection or illumination objective is used to reduce background fluorescence and phototoxicity [3,4]. By confining the excitation light to this light sheet, LSFM can achieve optical sectioning with wide-field detection. Although LSFM has proved its ability to image multi-cellular organisms at cellular resolution [5–7], a trade-off generally exists between the thickness of the light sheet and the imaging field of view (FOV) over which the light sheet remains thin [8,9]. In general, if an objective with a higher numerical aperture (NA) is used for excitation, the light sheet can be thinner but the FOV will be much smaller. Conversely, if we want to achieve larger FOV by using an objective with lower NA, the light sheet will be thicker. Moreover, when the sample is not very transparent, LSFM suffers from a high background level due to the excitation of scattered photons, which decreases image contrast and limits LSFM’s application [10–12]. Axial scanned [13] or tiling [14] light sheet microscopy could maintain a tight light sheet along a large imaging FOV, but require additional hardware to scan the excitation beam axially such as TAG lens or additional synchronizations with the detection camera.
The self-reconstructing Bessel beams are used in light-sheet microscopy for its increased depth of focus and good performance in scattering media [15,16]. We can easily create a Bessel beam by placing an annular aperture at the rear pupil plane of the excitation lens [17]. Adjusting the thickness of the annulus could help decouple the length of Bessel beams from their full width at half maxima, which can not be realized in Gaussian beams [18]. The self-healing or self-reconstruction characteristic of Bessel beams makes them able to regain their initial profile behind isolated obstacles and thus decrease artifacts in scattered medias [19]. However, the energy in the side lobe structure of the Bessel beam increases greatly with the decrease of the numerical aperture, which makes it have no superiority over the Gaussian beam in large-FOV conditions. To reduce the unexpected fluorescence excited by the side lobes, multi-photon excitation has been used and obtains a notable improvement in image contrast [20,21] by its nonlinear property in excitation efficiency. But the use of the femtosecond laser for multi-photon excitation is too expensive for many researchers.
In the meantime, the light-sheet fluorescence photobleaching imprinting microscopy (LSF-PIM) [22] has been proposed to obtain the nonlinear information by extracting high-order signals from photobleaching-induced nonlinear fluorescence decay. Instead of using the pulsed laser [23], the method uses time-series images taken by a conventional single-photon light-sheet microscope and can be easily implemented. While the LSF-PIM obtains a lower background level and better image contrast with the Gaussian light sheet, we expect that the non-linear property of the fluorescence decay works better with Bessel beams, especially in the low-NA conditions for large-scale wide-field light-sheet illumination due to the further suppression of the side lobes.
In this paper, we propose the light-sheet fluorescence photobleaching imprinting microscopy with scanned Bessel beams (LSF-PIM-BB) and confirm its superior performance over conventional methods for large-FOV fluorescence imaging with uniform and thin optical sections. At first, we illustrate the principles of the method and demonstrate some intuitive simulation results. Further, we present experiments on fluorescein isothiocyanate (FITC) gel and fluorescence beads to show the higher axial resolution provided by the better sectioning ability. We also obtain images of a cleared mouse brain with greatly improved contrast and expect our method to be a simple and powerful tool for large-scale imaging in various biological applications.
2. Experimental setup
A light-sheet microscope with scanned Bessel beams is built for fluorescence imaging and the optical setup is shown in Fig. 1. In our setup, Bessel beams are generated by using an annular aperture [17], which is conjugated to the back pupil plane of the illumination objective (IO) through a 4f system. A galvanometer is used for scanning to form a light sheet perpendicular to the optical axis of the detection objective. We define the $x$ axis as the propagation direction of illumination beams and the $xy$ plane as the image plane of the detection objective. The $z$ axis is orthogonal to $xy$ plane, which is along the optical axis of the detection objective. The ratio of inner radius ($R_{min}$) and outer radius ($R_{max}$) of the annular mask is defined as $e$ (see the inset of Fig. 1). $R_{max}$ and $e$ determine the property of the Bessel beams and can be chosen according to the application. Generally, the $R_{max}$ represents the numerical aperture of the excitation beam. With the increase of $R_{max}$, the resolution of Bessel beam will increase and the expected persistence length will decrease. The parameter $e$ is a parameter for the practical implementation of Bessel beam. When the $e$ becomes smaller, the implemented Bessel beam will be similar to Gaussian beam and the resolution will be better with a shorter persistence length [8].
Fig. 1. Schematic of the system setup. The galvanometer (GM), annular mask (AM) and the back pupil plane of the illumination objective (IO) are conjugated. The thin annular illumination on the back pupil plane of IO forms a Bessel beam within the sample in the $x$ direction and the scanning of GM introduces laterally scanning along the FOV in the $y$ direction to generate a light sheet. The emitted fluorescence is collected by the detection objective (DO) and a tube lens (TL) onto the sCMOS camera.
Laser (Sapphire 488, Coherent) is used as the excitation source. The beam is collimated and expanded using a 1:10 relay lens system L1 ($f$=10mm) and L2 ($f$=100mm) before reaching the galvanometer (GM) (GVS211/M, Thorlabs). The light is transmitted onto the annular mask (AM) by a 1:1 relay lens system L3 ($f$=75mm) and L4 ($f$=75mm) and then onto back pupil plane of illumination objective (IO) (Plan N, 10x/0.25, Olympus) by another 1:1 relay lens system L5 ($f$=100mm) and L6 ($f$=100mm). GM, AM and the back pupil plane of IO are all at conjugate planes. Thus, a thin annular illumination is projected to the back pupil plane of the IO and produces a Bessel beam along $x$ direction within the sample. The galvanometer tilts in the y-direction, and the illumination beam is scanned laterally in the focal plane of the detection objective (DO) (UMPlanFL N, 10x/0.3W, Olympus) to illuminate the full field of view. The fluorescence is imaged via the detection objective and a tube lens(TL)(TTL180-A, Thorlabs) onto an sCMOS camera (ORCA-Flash4.0 V3, Hamamatsu). The sample is placed on a z-axis stage (M111.1, PI, not shown in Fig. 1) when doing 3D imaging.
3. Theoretical analysis and simulation
The principle of light-sheet fluorescence photobleaching imprinting microscopy(LSF-PIM) is described in [22], in which the light sheet is generated by a cylindrical lens. The method can also be extended to scanned light sheet microscopy with Bessel beams just as what we do.
Using light sheet excitation, the measured light intensity within the FOV is an integration of the fluorescence emitted from all depths
where $C$ is a constant, $\mu _a$ is the absorption coefficient of the fluorophore, $F_b$ is the ballistic (unscattered) excitation fluence, $F_s$ is the scattered excitation fluence, $\mathrm {PSF}_z$ is the point-spread function (PSF) of the fluorescence imaging system at depth $z$.In most cases, photobleaching-induced decay in absorption coefficient $\mu _a$ can be seen as an exponential function [24,25] with time
where $k=B(F_b+F_s)$ is the photobleaching rate related to total excitation fluence, and $B$ is a constant.The time series image can be written as a polynomial of $t$ using Taylor expansion
In LSF-PIM, $I_n(x,y)$ is derived from the measured images to get a thinner equivalent optical section while keeping the FOV. Binomial expansion of the term $(F_b(z)+F_s(x,y,z))^{n+1}$ in $I_n(x,y)$ gives
For scanned bessel beams, the imaging FOV $d$ is related to the annular mask. We define $z_1$ as the point outside which $F_b(z_1)<0.05\cdot F_b(0)$ and $z_b$ as the first zero point of the generated Bessel beams. Assume that we use an annular with $e=0.9$, we have $z_1\approx 10\lambda /\mathrm {NA}$ and $z_b = 0.61\lambda /\mathrm {NA}$. Similarly, we assume that $F_b(z)\approx 0$ outside the optical section ($z<-z_1$ or $z>-z_1$) and $F_b(z)\gg F_s(x,y,z)$ within the optical section ($-z_1<z<z_1$). With an $e=0.9$ annular mask, we have $F_b(z)<0.53\cdot F_b(0)$ in $-z_1<z<-z_b$ and $z_b<z<z_1$. Binomial expansion of $(F_b(z)+F_s(x,y,z))^{n+1}$ gives
Fig. 2. Simulation results of effective excitation volumes. (a, b) Simulated excitation volumes with scanned Gaussian beams and Bessel beams (NA = 0.1 and $e$ = 0.9). (c, d) The corresponding first-order PIM component. (e, f) The corresponding fourth-order PIM component. Scale bar: 20$\mu$m. (g) Comparison of normalized equivalent illumination intensity along $z$-axis (dashed and solid lines in (a-f)). (h) Comparison of normalized equivalent illumination intensity along $x$-axis(dashed and solid lines in (a-f)).
For practical implementation, the PIM image is processed after the acquisition of time-lapse images. The process to obtain PIM component from time-lapse images could be describe as following steps:
- • Keep exciting the scanning Bessel-beam light sheet at one slice and take M time-lapsed images with the same exposure time $Im(1), Im(2),\ldots ,$ $Im(M)$;
- • For each point $(x,y)$ in a slice, we have a time lapse curve $Im(x,y,t)$ with intensity value $Im(x,y,1),Im(x,y,2),\ldots ,Im(x,y,M)$;
- • Use the exponential decay equation to fit the curve of $Im(x,y,t)$: $Im(x,y,t) = a(x,y)\times \exp (-b(x,y)t)$. In order to simplify the computation, we do logarithmic operation to the equation and get: $\mathrm {ln}(Im(x,y,t)) = \mathrm {ln}(a(x,y)) - b(x,y)t$, and then we can do a simple linear fit with the intensity value $Im(x,y,1),Im(x,y,2),$…$,Im(x,y,M)$ to get $\mathrm {ln}(a(x,y))$ and $b(x,y)$ for each point $(x,y)$;
- • For $n$th order PIM component, we could generate the image: ${Im_n}(x,y) = a(x,y) b(x,y)^n$
Simulation results of the effective excitation volume of LSF-PIM and LSF-PIM-BB are shown in Fig. 2. For the simulated excitation Gaussian beams, NA=0.1 and for the simulated excitation Bessel beams, NA=0.1 and $e=0.9$. Figures 2(a) and (b) shows the cross-sectional images of original scanned Gaussian beams and Bessel beams. The ring system of the Bessel beams is obvious and may lead to a serious decline in image quality. Figures 2(c) and (d) shows first-order PIM component (LSF-PIM($I_1$) and LSF-PIM-BB($I_1$)) and Figs. 2(e) and (f) shows fourth-order PIM component (LSF-PIM($I_4$) and LSF-PIM-BB($I_4$)) of both beams. The normalized intensity along z-direction is shown in Fig. 2(g) and the propagation along the x-direction is shown in Fig. 2(h). As expected, LSF-PIM-BB($I_4$) causes effective suppression of the ring system of Bessel beams. It’s clear that LSF-PIM-BB($I_4$) has almost the same thickness as LSF-PIM($I_4$) but its length is much longer.
4. Sample preparation
4.1 Fluorescein isothiocyanate and fluorescent beads
Solid agarose samples containing fluorescein isothiocyanate (FITC) or 1$\mu$m diameter green fluorescent beads were used to demonstrate the characteristics of the method. For the former, FITC was first dissolved in water to form a 0.5mg/L solution and then uniformly mixed with melted 1% agarose. For the latter, the beads were first diluted to a 1:20 bead-water solution and then 1:10 mixed with 1% agarose. Both mixtures were transferred to a quartz cuvette and were cooled to room temperature to form gels.
4.2 Cleared mouse brain
Cleared mouse brain slices were imaged in our experiment and they were prepared following the uDISCO [26] method. was An adult mouse (CHAT-CHR2-EYFP) was deeply anesthetized, perfused with phosphate-buffered saline (PBS, pH 7.4) and fixed with 4% paraformaldehyde(PFA). The brain was removed from the mouse and was cut into 1mm coronal slices. The slices were sequentially dehydrated in a series of tert-butanol (30%, 50%, 70%, 80%, 90%, 96% and 100%, two hours each step) and incubated in BABB-D4(benzyl alcohol:benzyl benzoate:diphenyl ether = 4:8:3) until cleared at room temperature. The cleared slices were then stored in BABB-D4 at 4$^{\circ }$C. When imaging, a cleared slice was put in a quartz cuvette filled with BABB-D4 for refractive index matching.
5. Experimental results
5.1 Equivalent excitation volume of LSF-PIM-BB
To experimentally demonstrate the equivalent excitation volume of LSF-PIM-BB,we imaged a time-series images of Gaussian(NA = 0.04) and Bessel beam(NA = 0.04, $e=0.9$) in the FITC gel described in section 4.1. In this experiment, we set the GM to be static and obtained 100 time-lapse x-z plane images (50ms exposure time, 20 frames per second) of the fluorescent line. Images of the excitation line and its fourth-order PIM components were shown in Fig. 3: (a) LSFM(Gauss beams) (b) LSFM-BB (c)LSF-PIM($I_4$) (d)LSF-PIM-BB($I_4$). which could be roughly seen as the $x-z$ cross-section of the corresponding light sheet. To quantificationally analyze the effect of this method, normalized intensity distributions along $z$ axis at different $x$ were plotted in Figs. 3(e) and (f).
Fig. 3. Measured excitation of Gaussian beam and Bessel beam. (a, b) Fluorescence line in the FITC gel excited by Gaussian beam and Bessel beam. (c, d) Fourth-order PIM component of the Gaussian beam and Bessel beam. Scale bar: 100$\mu$m. (e, f) Normalized intensity along $z$ direction at different $x=300\mu$m, $600\mu$m and $900\mu$m (dashed and solid lines in (a-d)).
In the FITC gel, the unexpected fluorescence excited by the side lobes of the Bessel beam is severer than simulation due to scattering, which will decrease axial resolution and image contrast. By extracting the high-order PIM component of the Bessel beam (fourth-order in this experiment), we observed an efficient suppression in the ring system of the Bessel beam. The full width at half maximum (FWHM) of both beams was calculated from the intensity curves. With LSF-PIM-BB, we achieved a five-fold increase in the FWHM of the beam from $\sim 25\mu$m to $\sim 5\mu$m, which is close to the LSF-PIM, but the length of LFM-PIM-BB is much longer. However, as the non-linear effect can shorten the length of the beam, it’s necessary to get a trade-off between axial resolution and field of view.
In order to illustrate how to decide when to stop an acquisition at one slice, we compared PIM images with different number of time-lapsed images. Generally speaking, sufficient number of images are required to estimate the parameters of the exponential decay curve for photo-bleaching and improve the noise robustness, such as illumination intensity fluctuations and shot noise. We conducted a comparison between the PIM components extracted from 10, 50 and 100 time-lapse images (50ms exposure time) of Bessel beam in the fluorescence dye. Figure 4 shows the PIM component extracted from different number of time-lapse images. More images with more photon numbers provide better estimations of the PIM components, while the PIM component extracted from 10 images is quite noisy and discontinuous. In this paper, we choose 100 as the number of time-lapse images based on the consideration of the continuous structures after reconstruction for fluorescence imaging.
Fig. 4. PIM component of Bessel beam with different number of time-lapsed images. 4th order PIM component using (a)10, (b)50 and (c)100 time-lapse images of Bessel beam in fluorescence dye. (d) Normalized intensity along $z$ direction at different $x=200\mu$m, $600\mu$m and $1000\mu$m. Scale bar: 100$\mu$m.
In addition, we further compared the different orders of PIM terms with the same time-lapsed images in Fig. 5. In general, if a higher order PIM term is used, the axial resolution will be better and the effective light-sheet length will be shorter. Furthermore, the side lobe effect cannot be totally eliminated if we choose a lower order PIM term. From the simulation, we found that the 4th PIM term can efficiently reduce the fluorescence excited by side lobes while keeping a relatively uniform beam. Figure 5 shows different order PIM terms extracted from the 100 time-lapse images of Bessel beams in fluorescence dye. In the 2nd PIM term, we can still see some side lobes. While in the 8th PIM term, the equivalent excitation intensity quickly decreased at the edge of the FOV. We choose the 4th PIM term here because it can maintain a large excitation area with significantly-reduced side lobes.
Fig. 5. Different order PIM terms with same time-lapsed images. (a)2nd, (b)4th and (c)8th PIM term extracted from 100 time-lapse images of Bessel beam in the fluorescence dye. (d) Normalized intensity along $z$ direction at different $x=200\mu$m, $600\mu$m and $1000\mu$m. Scale bar: 100$\mu$m.
5.2 Fluorescent beads imaging
To validate the effect of this method in weak scattering mediums, we imaged a solid agar sample containing 1$\mu$m diameter green fluorescent beads. A 3d image stack was collected with an NA = 0.1 and $e=0.8$ Bessel light sheet. We show the maximum projection along $z$ direction and $y$ direction in Figs. 6(a) and (b). For every single layer, 50 time-lapsed images were acquired (50ms exposure time, 20 images per second) and the fourth-order PIM image was extracted. Similarly, the maximum projection of the PIM stack is shown in Figs. 6(c) and (d). The corresponding areas in Figs. 6(b) and (d) are magnified in Fig. 6(f). We also show the magnified image of a single bead for comparison in Fig. 6(e).
Fig. 6. 3D image stack of fluorescent beads ($\bf {d=1\mu }$m) in solid agar with original Bessel beam and LSF-PIM-BB. Maximum intensity projection of the original image stack shown above and the maximum intensity projection of the fourth-order PIM component is shown below. (a, c) $x-y$ projection. Scale bar: 25$\mu$m. (b, d) $x-z$ projection. (e) Magnified images of the area marked by a dashed white rectangle in (b, d). Scale bar: 5$\mu$m. (f) Magnified images of the area marked by a solid white rectangle in (b, d). Scale bar: 5$\mu$m. (g) Statistics of the axial FWHM of 10 beads in the original image stack (cyan) and the fourth-order PIM component (red).
In conventional cases, the Bessel beams excite a lot of unwanted fluorescent beads outside their main lobe due to both the rings and scattering. The $x-z$ projection in Fig. 6(b) indicates that a bead can be excited by several light sheets nearby. The statistical axial full width at half maximum (FWHM) of the original image stack is 4.5$\pm$0.7$\mu$m, as shown in Fig. 6(g) (cyan). The axial resolution of the 4th PIM component is greatly improved owing to the thinner effective excitation volume. The axial FWHM is reduced to 2.2$\pm$0.2$\mu$m in Fig. 6(g) (red), which verifies the effectiveness of this method in weak scattering media. For a typical example, the two points marked by a white dashed rectangle could not be resolved in LSF-BB, but they can now be clearly separated in our LSF-PIM-BB.
Then we imaged a $x-y$ plane of densely distributed 1$\mu$m diameter green fluorescent beads, in order to prove axial resolution and image contrast enhancement in a scattering medium. The excitation Bessel beams have an NA=0.06 and $e=0.8$ and the Gaussian beams have the same NA. 100 time-lapse x-y plane images (50ms exposure time, 20 frames per second) of the densely distributed beads were captured and the fourth-order PIM image was extracted. Figure 7 shows the (a) LSFM (b) LSFM-BB (c) LSF-PIM($I_4$) and (d) LSF-PIM-BB ($I_4$) images of the beads. Figures 7(e) and (f) shows the maximum projection intensity along the y-axis of the derived PIM components. The fluorescent beads are almost uniformly distributed in the agarose so that the y-axis total intensity curve can show the improvement of the effective field of view with high contrast.
Fig. 7. Image of fluorescent beads ($\bf {d=1\mu }$m) in solid agar with Gaussian and Bessel beams. Image with (a) conventional scanned Gaussian beams and (b) conventional scanned Bessel beams are shown. Scale bar: 100$\mu$m. Fourth-order PIM component of (c) Gaussian beams and (d) Bessel beams are extracted. The max projection along $y$ direction of (c) and (d) is plotted in (e) and (f).
Generally speaking, the Bessel beams excite a lot of unwanted fluorescent beads outside their main lobe (Fig. 7(b)) than Gaussian beams (Fig. 7(a)). The background fluorescent signal of Bessel beams is much stronger than that of Gaussian beams. After extracting their fourth-order PIM component, both beams show a great increase in image contrast and axial resolution. As we can expect, the light sheet formed by Gaussian beams (Fig. 7(c)) suffers a lot from the decrease in FOV while the light sheet formed by Bessel beams (Fig. 7(d)) performs well over a range of 1mm.
5.3 Cleared mouse brain imaging
The main application of light-sheet fluorescence microscopy is to observe large cleared biological samples. We tested our method by imaging a slice of cleared mouse brain(the preparation of the sample is described in section 4.2). The original image and the fourth-order PIM image extracted from 100 images(50ms exposure time, 20 frames per second) are shown in Figs. 8(a) and (b). For better comparison, we magnify a part containing two fluorescent points in Figs. 8(c) and (d) and plot the intensity distribution in Fig. 8(e). As usual, the conventional LSFM-BB image suffers from the severe background and low image contrast. By using its PIM component, background excited by the side lobes are almost eliminated. As is illustrated in Fig. 8(e), the features of the brain can be seen more clearly with the increased image contrast. In addition, we provide an example of the photobleaching and fitting curve for cleared mouse brain images. We picked a point in the area marked by dashed white box in Fig. 8, which is further magnified in Figs. 8(f) and (g). Figure 8(h) is the original time-lapse data of that point and its fitting curve. Despite the noise fluctuation in the original time-lapse data, we can still obtain a good fitting curve with 100 time points.
Fig. 8. Images of cleared mouse brain. (a) Light-sheet fluorescence image taken with scanned Bessel beams. (b) Fourth-order PIM component extracted from time-lapse images. Scale bar: 100$\mu$m. (c,d) Magnification of the part marked by a white box in the original image and the PIM image. Scale bar: 5$\mu$m. (e) Normalized intensity along dashed and solid lines in (c, d). (f,g) Magnification of the part marked by a dashed white box in the original image and the PIM image. Scale bar: 5$\mu$m. (h)Original time-lapse data of the point marked by an arrow and its fitting curve.
6. Conclusions
We have demonstrated that the axial resolution of Bessel light-sheet fluorescence microscopy can be greatly improved by extracting the photobleaching-induced fluorescence decay. We validated the LSF-PIM-BB method by imaging fluorescent beads and showed its application in observing cleared mouse brain. In the experiments, fluorescence excited by the side lobes of the Bessel beams as well as scattering in the sample can be greatly suppressed due to the non-linear effect of photobleaching, decreasing the thickness of Bessel light sheet. In addition, our method can be implemented on any Bessel-Beam LSM without any hardware modifications. With more and more requirements for large-scale imaging of biological samples such as embryos and cleared brains, we anticipate our LSF-PIM-BB can attract great interests with its superior performance in both FOV and resolution.
Funding
Beijing Municipal Science and Technology Commission (Z181100003118014); Tsinghua University Initiative Scientific Research Program; China Postdoctoral Science Foundation (2019M660644).
Acknowledgments
We thank Minmin Luo and Qiru Feng from the National Institute of Biological Sciences (NIBS) for providing the transgenic mouse brain.
Disclosures
The authors declare no conflicts of interest.
References
1. F. Strobl, A. Schmitz, and E. H. K. Stelzer, “Improving your four-dimensional image: traveling through a decade of light-sheet-based fluorescence microscopy research,” Nat. Protoc. 12(6), 1103–1109 (2017). [CrossRef]
2. A. H. Voie, D. Burns, and F. Spelman, “Orthogonal-plane fluorescence optical sectioning: three-dimensional imaging of macroscopic biological specimens,” J. Microsc. 170(3), 229–236 (1993). [CrossRef]
3. R. Galland, G. Grenci, A. Aravind, V. Viasnoff, V. Studer, and J.-B. Sibarita, “3d high-and super-resolution imaging using single-objective spim,” Nat. Methods 12(7), 641–644 (2015). [CrossRef]
4. B. Yang, X. Chen, Y. Wang, S. Feng, V. Pessino, N. Stuurman, N. H. Cho, K. W. Cheng, S. J. Lord, L. Xu, D. Xie, R. D. Mullins, M. D. Leonetti, and B. Huang, “Epi-illumination spim for volumetric imaging with high spatial-temporal resolution,” Nat. Methods 16(6), 501–504 (2019). [CrossRef]
5. R. Tomer, K. Khairy, F. Amat, and P. J. Keller, “Quantitative high-speed imaging of entire developing embryos with simultaneous multiview light-sheet microscopy,” Nat. Methods 9(7), 755–763 (2012). [CrossRef]
6. P. J. Keller and M. B. Ahrens, “Visualizing whole-brain activity and development at the single-cell level using light-sheet microscopy,” Neuron 85(3), 462–483 (2015). [CrossRef]
7. J. B. Treweek, K. Y. Chan, N. C. Flytzanis, B. Yang, B. E. Deverman, A. Greenbaum, A. Lignell, C. Xiao, L. Cai, M. S. Ladinsky, P. J. Bjorkman, C. C. Fowlkes, and V. Gradinaru, “Whole-body tissue stabilization and selective extractions via tissue-hydrogel hybrids for high-resolution intact circuit mapping and phenotyping,” Nat. Protoc. 10(11), 1860–1896 (2015). [CrossRef]
8. T. A. Planchon, L. Gao, D. E. Milkie, M. W. Davidson, J. A. Galbraith, C. G. Galbraith, and E. Betzig, “Rapid three-dimensional isotropic imaging of living cells using bessel beam plane illumination,” Nat. Methods 8(5), 417–423 (2011). [CrossRef]
9. T. Vettenburg, H. I. Dalgarno, J. Nylk, C. Coll-Lladó, D. E. Ferrier, T. Čižmár, F. J. Gunn-Moore, and K. Dholakia, “Light-sheet microscopy using an airy beam,” Nat. Methods 11(5), 541–544 (2014). [CrossRef]
10. P. J. Keller, A. D. Schmidt, A. Santella, K. Khairy, Z. Bao, J. Wittbrodt, and E. H. Stelzer, “Fast, high-contrast imaging of animal development with scanned light sheet–based structured-illumination microscopy,” Nat. Methods 7(8), 637–642 (2010). [CrossRef]
11. J. Mertz and J. Kim, “Scanning light-sheet microscopy in the whole mouse brain with hilo background rejection,” J. Biomed. Opt. 15(1), 016027 (2010). [CrossRef]
12. L. Silvestri, A. Bria, L. Sacconi, G. Iannello, and F. Pavone, “Confocal light sheet microscopy: micron-scale neuroanatomy of the entire mouse brain,” Opt. Express 20(18), 20582–20598 (2012). [CrossRef]
13. K. M. Dean and R. Fiolka, “Uniform and scalable light-sheets generated by extended focusing,” Opt. Express 22(21), 26141–26152 (2014). [CrossRef]
14. D. Wang, Y. Jin, R. Feng, Y. Chen, and L. Gao, “Tiling light sheet selective plane illumination microscopy using discontinuous light sheets,” Opt. Express 27(23), 34472–34483 (2019). [CrossRef]
15. F. O. Fahrbach, P. Simon, and A. Rohrbach, “Microscopy with self-reconstructing beams,” Nat. Photonics 4(11), 780–785 (2010). [CrossRef]
16. F. O. Fahrbach and A. Rohrbach, “Propagation stability of self-reconstructing bessel beams enables contrast-enhanced imaging in thick media,” Nat. Commun. 3(1), 632 (2012). [CrossRef]
17. Y. Lin, W. Seka, J. Eberly, H. Huang, and D. Brown, “Experimental investigation of bessel beam characteristics,” Appl. Opt. 31(15), 2708–2713 (1992). [CrossRef]
18. J. Durnin, “Exact solutions for nondiffracting beams. i. the scalar theory,” J. Opt. Soc. Am. A 4(4), 651–654 (1987). [CrossRef]
19. F. O. Fahrbach and A. Rohrbach, “A line scanned light-sheet microscope with phase shaped self-reconstructing beams,” Opt. Express 18(23), 24229–24244 (2010). [CrossRef]
20. O. E. Olarte, J. Licea-Rodriguez, J. A. Palero, E. J. Gualda, D. Artigas, J. Mayer, J. Swoger, J. Sharpe, I. Rocha-Mendoza, R. Rangel-Rojo, and P. Loza-Alvarez, “Image formation by linear and nonlinear digital scanned light-sheet fluorescence microscopy with gaussian and bessel beam profiles,” Biomed. Opt. Express 3(7), 1492–1505 (2012). [CrossRef]
21. F. O. Fahrbach, V. Gurchenkov, K. Alessandri, P. Nassoy, and A. Rohrbach, “Light-sheet microscopy in thick media using scanned bessel beams and two-photon fluorescence excitation,” Opt. Express 21(11), 13824–13839 (2013). [CrossRef]
22. L. Gao, L. Zhu, C. Li, and L. V. Wang, “Nonlinear light-sheet fluorescence microscopy by photobleaching imprinting,” J. R. Soc., Interface 11(93), 20130851 (2014). [CrossRef]
23. F. Helmchen and W. Denk, “Deep tissue two-photon microscopy,” Nat. Methods 2(12), 932–940 (2005). [CrossRef]
24. L. Song, E. Hennink, I. T. Young, and H. J. Tanke, “Photobleaching kinetics of fluorescein in quantitative fluorescence microscopy,” Biophys. J. 68(6), 2588–2600 (1995). [CrossRef]
25. E. J. Peterman, S. Brasselet, and W. Moerner, “The fluorescence dynamics of single molecules of green fluorescent protein,” J. Phys. Chem. A 103(49), 10553–10560 (1999). [CrossRef]
26. C. Pan, R. Cai, F. P. Quacquarelli, A. Ghasemigharagoz, A. Lourbopoulos, P. Matryba, N. Plesnila, M. Dichgans, F. Hellal, and A. Ertürk, “Shrinkage-mediated imaging of entire organs and organisms using udisco,” Nat. Methods 13(10), 859–867 (2016). [CrossRef]
