Compared with two-photon point-scanning microscopy, line-scanning temporal focusing microscopy breaks the limitation on imaging rate and maintains the axial resolution, which makes it promising for various biomedical studies. However, for deep tissue imaging, it suffers from reduced axial resolution and increased background noise due to sample induced wavefront distortion. Here, we propose a spatio-spectral focal modulation technique to enhance axial resolution and background rejection by simply subtracting an aberrated image, which is induced by a spatial light modulator, from an unaberrated image. The proposed technique could improve the axial resolution by a factor of 1.3 in our implementation, verified by both simulations and experiments. Besides, we show that compared with spatial modulation alone, spatio-spectral modulation induces less peak intensity loss caused by image subtraction. We further demonstrate the performance of our technique on the enhanced axial resolution and background rejection by deep imaging of cleared mouse brains and in vivo imaging of living mouse brains.
© 2018 Optical Society of America under the terms of the OSA Open Access Publishing Agreement
Two-photon laser scanning microscopy (TPLSM) becomes the golden standard in current biomedical studies, owing to its advantages in high resolution, intrinsic optical sectioning capability, deep penetration and biocompatibility [1–4]. However, the imaging speed of TPLSM is limited by the inertia of mechanical scanners, which hampers studies of most high-speed biological dynamics [5–7]. Recently, temporal focusing microscopy (TFM), a highly parallelized excitation technique, has been proposed to achieve optical sectioned excitation in biological tissues [8–13]. By spatially dispersing optical pulses with a grating, the pulse width is broadened temporally and the peak power is decreased everywhere except at the focal plane of the objective, thus ensuring that two-photon excitation is confined to this 2D plane. Different from conventional two-photon plane excitation scheme, TFM is able to maintain high axial resolution. Considering the potential improvement on imaging speed, TFM enables dynamic volumetric imaging which is inapplicable for TPLSM such as recording large-volume neural activities at tens of hertz . There are generally two modalities of TFM: planar illumination TFM in which samples are illuminated by a pulsed plane [15–17], and line-scanning TFM in which samples are illuminated by a mechanically swept pulsed line . Compared with planar illumination, line-scanning temporal focusing microscopy (LTFM) exhibits higher robustness to scattering and better axial confinement [19–21], benefiting from the fact that spatial focusing and temporal compression of excitation pulses are simultaneously achieved in LTFM. Due to the well balance between imaging speed and axial resolution, LTFM is ideal for laser processing  and large-scale imaging of biological dynamics .
However, suffering from scattering, the axial extent of the excitation volume of LTFM progressively broadens in the turbid medium as the penetration depth increases, which further decreases the axial resolution. Besides, for deep imaging, the scattering of emission photons degrades the image contrast and obscures the fine features of the specimen, due to the fact that LTFM uses array detectors to record the emission photons parallel .
To improve the axial resolution and reduce out-of-focus background, the focal modulation microscopy (FMM) was proposed [25–27]. In FMM, two successive acquisitions are required to get an improved image: one under an aberrated point-spread-function (PSF) and the other under the unaberrated PSF. A simple subtraction of the aberrated image from the unaberrated image could remove background noise, while preserving most of signals in focus. The aberrated image could be achieved by applying pre-designed wavefronts on spatial light modulators (SLMs)  or by introducing spectral and polarization modulation . However, the improved the axial resolution is usually achieved at the cost of lower peak intensity after subtraction, which may reduce the dynamic range of final images. Moreover, Durst et al. improved the axial confinement by combining sum-frequency generation with TFM, but the distorted temporal profile of the pulse and the power loss due to the partial obstruction of the excitation beam might be problematic .
In this paper, we propose a novel method that combines spatio-spectral FMM with LTFM to improve the axial resolution, maintain high peak intensity, and effectively reject background scattered emission photons. We first simulate that in our implementation, the proposed method which shows 1.3-fold axial resolution improvement with only 27% peak intensity dropped, then validate it experimentally via imaging with fluorescent beads. To further demonstrate the improvement of our spatio-spectral focal modulation technique, we conduct deep imaging of cleared Thy1-YFP mice brains and in vivo imaging of living Cx3Cr1-GFP mice brains.
2. Experimental setup and Methods
2.1 Optical design of LTFM combined with spatio-spectral FMM
The schematic of the proposed system is illustrated in Fig. 1. The design combines the LTFM setup with an SLM for focal modulation. The laser beam from the 80 MHz Chameleon Discovery (Coherent) with the pulse duration of ∼100 fs at the central wavelength of 920 nm is used for the two-photon excitation. The laser power is controlled by an electro-optical modulator. The temporal dispersion caused by transmissive optical components of the system is compensated by the laser built-in compensator. The laser beam is then expanded to 8 mm in diameter by a beam expander (BE05-10-B, Thorlabs). Expanded laser is afterwards polarization-rotated by a half-wave plate to ensure the grating diffraction efficiency, then scanned in the vertical direction by a one-dimensional galvanometer (GVS211, Thorlabs) and focused to a thin line by a cylindrical lens (f = 400 mm) at the surface of the diffraction grating (Edmund Optics, 830 lines/mm). A reflection mirror placed between the cylindrical lens and the grating directs the light at ∼50°incident angle to the grating to ensure that the central wavelength of 1st diffracted light is perpendicular to the grating surface. The spectrally spread pulse is collimated by a collimating lens (f = 500 mm) and modulated by a phase-only spatial light modulator (SLM, X10468-07, Hamamatsu) at the focal plane of the collimating lens. The diameter of formed two-dimensional beam in the SLM surface is about 12 mm along x-axis and 10 mm along y-axis. A 4f telescope (both with focal length f = 300 mm) delivers the modulated beam profile to the pupil plane of the objective (25 × , 1.05 NA, Olympus, XLPLN25XWMP2). A line-shaped laser line is formed in the focal plane of the objective, whose length is around 160 µm. For each scanning period, we capture the fluorescent image using an epi-fluorescence setup including a dichroic mirror (DMSP750B, Thorlabs), a bandpass filter (E510/80, Chroma), a 200 mm tube lens (TTL200-A, Thorlabs), and an sCMOS (Andor Zyla 5.5 plus). Three-dimensional imaging can be performed with axial movement of the sample stage and/or the piezo objective scanner (P725, PI).
The grating disperses the laser and introduces temporal chirp to the laser pulses. Pulse duration reaches its minimum only at the focal plane of the objective and broadens rapidly along axial direction, resulting in a compressed axial excitation. The cylindrical lens converges the laser beam into a thin line at the grating surface and the vertically scanning galvanometer mirror reflects the line-shaped laser beam, leading to a lateral scanning on the specimen.
Hardware synchronization is carried out using a multi-functional DAQ (NI Instrument, USB-6363), with three analog output channels controlling the EOM, galvanometer scanner, and piezo objective scanner, and two digital channels controlling the camera and the shutter. All the controls are conducted by the ScanImage software .
2.2 Focal modulation of LTFM
Different from conventional line scanning two-photon microscopy , in which excitation beams form a line before entering the objective, the back focal plane of the objective in LTFM is filled in two dimensions instead of one, as shown in Fig. 1(b). The two-dimensional phase profile could be divided into the spatial dimension (along the y-axis) and the spectral dimension (along the x-axis). Changing the phase function along the y-axis would shape the wavefront of the specific frequency ω, while changing the phase along x-axis would shape the dispersion of the pulse. The additional spectral dimension allows us to control the PSF more flexibly. Our goal is to subtract the distorted laser focus from the undistorted one to generate a laser focus with smaller axial FWHM, as shown in Fig. 1(c).
We conduct numerical simulations to study the focal modulation approach and illustrate the benefit of spatio-spectral modulation. The simulation parameters, such as effective NA in the experiment, the wavelength range of the output laser pulse, and the pulse width, are based on the experimental setup as described above. The simulation begins from the pupil plane of the objective, in which we shape the pulse beam profile with the theoretical description in Ref . The beam propagates to the objective via Fresnel diffraction , then forms the focus based on the Born and Wolf model [35, 36]. We investigate the two-photon excitation intensity to analyze the axial resolution as well as focal intensity under different modulations in the pupil plane.
Firstly, we add a π spatial phase function into the pupil plane of the objective [Fig. 2(b), the added phase fills 70% of the dispersed beam and locates at the center of the beam], i.e. the conjugated plane of the SLM modulation plane. The simulation result shows that the induced aberrations have much less effect on axial wings of the PSF, but the peak intensity drops by a factor of 2.3 [shown in blue curve of Fig. 2(d)]. In practical imaging, one prefers the aberrated PSF of low peak intensity, as high peak intensity of aberrated PSF would reduce the peak intensity of the subtracted PSF [shown in blue curve of Fig. 2(e)], which further reduces the dynamic range of the captured image. Instead, if we further add a π spectral phase function [Fig. 2(c), the added phase fills 60% of the dispersed beam and locates at the center of the beam], the peak intensity of the aberrated PSF would drop by a factor of 3.6, but the intensity away from axial focus center can be retained, as shown in the green curve of Fig. 2(d). In Fig. 2(e), we could see that after subtraction the spatio-spectral modulation leads to a PSF of higher peak intensity but almost the same FWHM compared with the one with spatial modulation only, and the effect of background rejection is apparent. As the simulation results show, the axial FWHM decreases from 3.2 µm to 2.4 µm by using our spatio-spectral FMM, i.e. the axial resolution increase by 1.3 folds in our current implementation, while the peak intensity drops 27% compared with 43% by simply using spatial focal modulation. Besides, Fig. 2(f) shows that our method would not compromise the lateral resolution.
3. Experimental Results
All procedures involving mice were approved by the Animal Care and Use Committees of Tsinghua University.
3.1 Experimental validation of spatio-spectral focal modulation
We experimentally demonstrate the fidelity of our technique in enhancing axial resolution by imaging fluorescent microspheres. We monitored 500 nm-diameter green-labeled fluorescent microspheres (T14792, Thermo fisher) and acquired a z-stack by driving the piezo objective scanner. All the solid lines in Fig. 3 represent the simulation results while the makers represent the corresponding experimental results. As Fig. 3 shows, compared to the normal PSF (red), the aberrated PSF (blue) has lower peak intensity but remains the wing intensity. For the subtracted PSF (green), we sacrifice a little peak intensity but achieve improved axial resolution by about 1.3 folds, which agrees well with previous simulation results under our current implementation.
3.2 Deep imaging of the cleared Thy1-YFP mouse brain
To further investigate the effectiveness of our technique in biological tissues, we imaged the cleared brains of the Thy1-YFP (H line) mice (JAX No. 003782) by uDISCO (ultimate 3D imaging of solvent-cleared organs) . The mice were deeply anesthetized, and perfused with phosphate buffered saline (pH 7.4) followed by 4% paraformaldehyde for fixation. The brains were removed from the mouse bodies and were cut into 1 mm coronal slices after fixation overnight. The slices were sequentially dehydrated in a series of tert-butanol (30%, 50%, 70%, 80%, 90%, 96% and 100%, two hours each step) at room temperature. The dehydrated slices were incubated in BABB-D4 (BABB: benzyl alcohol/benzyl benzonate = 1/2, BABB-D4: BABB/diphenyl ether = 4/1) for more than one hour at room temperature until they became optically transparent.
We took the image at ~300 μm depth under the surface of the slices. We show the maximum intensity projections (MIPs) of a 20 µm thick x-y image stack without and with the spatio-spectral FMM in Figs. 4(a) and 4(b), respectively. It can be seen that the image achieved with our technique has much clear background compared with the original one. In Fig. 4(c), we show MIPs along y-axis of a 10 µm thick x-z image stack [marked by the dashed red box in (a)], and the corresponding intensity fluctuation along indicated lines. It shows that with our spatio-spectral focal modulation technique, the axon of the YFP-expressing neuron shows increased sharpness in z-axis owning to the improved axial resolution. We also quantitatively compare the background reduction in Fig. 4(d), in which the signals along the dashed lines in Figs. 4(a) and 4(b) are plotted. As we could see, the structure is better resolved and the background is much lower with our spatio-spectral focal modulation. This is because the out-of-focus signal is not affected by the SLM induced aberration  and therefore it is eliminated by image subtraction in our proposed method.
3.3 In vivo imaging of the living Cx3Cr1-GFP mouse brain
We also demonstrate the performance of the proposed technique in in vivo imaging of living Cx3Cr1-GFP mouse (JAX No. 005582) brains. After craniotomy , we allowed 2-3 weeks for the mice to fully recover from optical window implanting. In Fig. 5(a), we show the MIP along the z-axis of a 20-µm-thick image stack (50-70 μm under the dura) acquired with our spatio-spectral focal modulation technique. To further validate the axial resolution improvement enabled by our method, we investigate the signal intensity from images acquired with and without our focal modulation. In Fig. 5(b), we show MIPs along y-axis of a 10 µm thick x-z image stack [marked by the dashed yellow box in (a)], and the corresponding intensity fluctuation along indicated lines. It shows that with our spatio-spectral focal modulation technique, the process of the GFP-expressing microglia shows higher contrast in z-axis, benefiting from the enhanced axial resolution and background rejection. In addition, we investigate two adjacent processes along z-axis via MIP along x-axis of a 10 µm thick y-z image stack in Fig. 5(c). It is apparent that such adjacent structures could be clearly resolved with our spatio-spectral focal modulation technique. However, they are hard to be distinguished in the image acquired with conventional LTFM. We also investigate the background reduction by our technique by quantitatively comparing the microstructures of microglia [the signals along the dashed lines in Fig. 5(a)]. The improvement is obvious, as shown in Figs. 5(d) and 5(e).
We have achieved enhanced axial resolution and background rejection in line-scanning temporal focusing microscopy by focal modulation. Limited by the low pulse energy available from our laser source, we only demonstrate the proof-of-concept biological experiments in this paper. To achieve images with higher penetration depth and higher signal-to-noise ratio (SNR), we need to choose a small field of view, which limits the actual excitation NA in the system . A femtosecond laser with low repetition rate but high pulse energy could improve the SNR as well as fully use the NA of the adopted objective .
On choosing the fill factors of π phase for spatial and spectral phase modulation, it is a comprehensive consideration between axial resolution and final peak intensity. Besides, it depends on other parameters, such as the wavelength range of the laser, effective NA of the system, intensity distribution of the beam, etc.
For cases where sample motion occurs or fast dynamical imaging (including morphorlogy change or functional imaging) is required, the images with and without introducing aberration should be acquired successively and quickly. However, the practical imaging speed may be limited by the refreshing rate of the phase modulation component. In our demonstration, a SLM is employed whose refreshing rate might not be high enough. Note that the used phase modulation is not complicated in our proposed method, thus a high-frame-rate but lower-pixel-count phase modulator such as the deformable mirror could support high-speed imaging, which eliminates the effects of sample motion and distortion.
In summary, we have demonstrated a spatio-spectral focal modulation technique that could improve the imaging axial resolution as well as preserve the peak intensity after subtraction algorithm, by fully using the properties of line-scanning temporal focusing microscopy. The proposed technique could also eliminate the background noise effectively in deep tissue imaging. Through both simulations and experiments, we investigated the performance of the spatio-spectral focal modulation and validated the performance improvement in deep imaging of cleared mouse brains and in vivo imaging of living mouse brains.
National Natural Science Foundation of China (NSFC) (No. 61771287, No. 61327902).
YZ thanks Yingjun Tang for helps in sample preparation. LK thanks the support from Tsinghua University and the “Thousand Talents Plan” Youth Program.
References and links
4. A. Song, A. S. Charles, S. A. Koay, J. L. Gauthier, S. Y. Thiberge, J. W. Pillow, and D. W. Tank, “Volumetric two-photon imaging of neurons using stereoscopy (vTwINS),” Nat. Methods 14(4), 420–426 (2017). [CrossRef] [PubMed]
6. W. Yang, J. E. Miller, L. Carrillo-Reid, E. Pnevmatikakis, L. Paninski, R. Yuste, and D. S. Peterka, “Simultaneous Multi-plane Imaging of Neural Circuits,” Neuron 89(2), 269–284 (2016). [CrossRef] [PubMed]
7. L. Kong, J. Tang, J. P. Little, Y. Yu, T. Lämmermann, C. P. Lin, R. N. Germain, and M. Cui, “Continuous volumetric imaging via an optical phase-locked ultrasound lens,” Nat. Methods 12(8), 759–762 (2015). [CrossRef] [PubMed]
10. Y. Meng, W. Lin, C. Li, and S. C. Chen, “Fast two-snapshot structured illumination for temporal focusing microscopy with enhanced axial resolution,” Opt. Express 25(19), 23109–23121 (2017). [CrossRef] [PubMed]
11. O. Hernandez, E. Papagiakoumou, D. Tanese, K. Fidelin, C. Wyart, and V. Emiliani, “Three-dimensional spatiotemporal focusing of holographic patterns,” Nat. Commun. 7, 11928 (2016). [CrossRef] [PubMed]
12. C. Y. Chang, Y. Y. Hu, C. Y. Lin, C. H. Lin, H. Y. Chang, S. F. Tsai, T. W. Lin, and S. J. Chen, “Fast volumetric imaging with patterned illumination via digital micro-mirror device-based temporal focusing multiphoton microscopy,” Biomed. Opt. Express 7(5), 1727–1736 (2016). [CrossRef] [PubMed]
14. R. Prevedel, A. J. Verhoef, A. J. Pernía-Andrade, S. Weisenburger, B. S. Huang, T. Nöbauer, A. Fernández, J. E. Delcour, P. Golshani, A. Baltuska, and A. Vaziri, “Fast volumetric calcium imaging across multiple cortical layers using sculpted light,” Nat. Methods 13(12), 1021–1028 (2016). [CrossRef] [PubMed]
15. T. Schrödel, R. Prevedel, K. Aumayr, M. Zimmer, and A. Vaziri, “Brain-wide 3D imaging of neuronal activity in Caenorhabditis elegans with sculpted light,” Nat. Methods 10(10), 1013–1020 (2013). [CrossRef] [PubMed]
16. J. N. Yih, Y. Y. Hu, Y. D. Sie, L. C. Cheng, C. H. Lien, and S. J. Chen, “Temporal focusing-based multiphoton excitation microscopy via digital micromirror device,” Opt. Lett. 39(11), 3134–3137 (2014). [CrossRef] [PubMed]
17. E. Papagiakoumou, F. Anselmi, A. Bègue, V. de Sars, J. Glückstad, E. Y. Isacoff, and V. Emiliani, “Scanless two-photon excitation of channelrhodopsin-2,” Nat. Methods 7(10), 848–854 (2010). [CrossRef] [PubMed]
18. Z. Li, J. Hou, J. Suo, C. Qiao, L. Kong, and Q. Dai, “Contrast and resolution enhanced optical sectioning in scattering tissue using line-scanning two-photon structured illumination microscopy,” Opt. Express 25(25), 32010–32020 (2017). [CrossRef] [PubMed]
20. J. K. Park, C. J. Rowlands, and P. T. C. So, “Enhanced Axial Resolution of Wide-Field Two-Photon Excitation Microscopy by Line Scanning Using a Digital Micromirror Device,” Micromachines (Basel) 8(3), 85 (2017). [CrossRef] [PubMed]
21. G. Sela, H. Dana, and S. Shoham, “Ultra-deep penetration of temporally-focused two-photon excitation,” 8588, 858824 (2013).
22. B. Sun, P. S. Salter, C. Roider, A. Jesacher, J. Strauss, J. Heberle, M. Schmidt, and M. J. Booth, “Four-dimensional light shaping: manipulating ultrafast spatiotemporal foci in space and time,” Light Sci. Appl. 7(1), 17117 (2018). [CrossRef]
23. H. Dana, A. Marom, S. Paluch, R. Dvorkin, I. Brosh, and S. Shoham, “Hybrid multiphoton volumetric functional imaging of large-scale bioengineered neuronal networks,” Nat. Commun. 5, 3997 (2014). [CrossRef] [PubMed]
27. S. Piazza, P. Bianchini, C. Sheppard, A. Diaspro, and M. Duocastella, “Enhanced volumetric imaging in 2-photon microscopy via acoustic lens beam shaping,” J. Biophotonics 11(2), e201700050 (2018). [CrossRef] [PubMed]
32. G. J. Brakenhoff, J. Squier, T. Norris, A. C. Bliton, M. H. Wade, and B. Athey, “Real-time two-photon confocal microscopy using a femtosecond, amplified Ti:sapphire system,” J. Microsc. 181(3), 253–259 (1996). [CrossRef] [PubMed]
34. J. W. Goodman, Introduction to Fourier optics (Roberts and Company Publishers, 2005).
36. E. Wolf and M. Born, Principles of optics (Pergamon Press, 1980).
37. 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] [PubMed]
39. L. C. Cheng, C. Y. Chang, C. Y. Lin, K. C. Cho, W. C. Yen, N. S. Chang, C. Xu, C. Y. Dong, and S. J. Chen, “Spatiotemporal focusing-based widefield multiphoton microscopy for fast optical sectioning,” Opt. Express 20(8), 8939–8948 (2012). [CrossRef] [PubMed]
40. O. D. Therrien, B. Aubé, S. Pagès, P. D. Koninck, and D. Côté, “Wide-field multiphoton imaging of cellular dynamics in thick tissue by temporal focusing and patterned illumination,” Biomed. Opt. Express 2(3), 696–704 (2011). [CrossRef] [PubMed]