Open Access
Add to BibSonomy
Abstract
Development of miniature two-photon microscopy (m2PM) has made it possible to observe fine structure and activity of neurons in the brain of freely moving animals. However, the imaging field-of-view of existing m2PM is still significantly smaller than that of miniature single-photon microscopy. Here we report that, through the design of low-magnification objective, large field-of-view scan lens and small tilt angle microscanner, a 2.5-g m2PM achieved a field-of-view of 1000 × 788 µm2, comparable to that of a typical single-photon miniscope. We demonstrated its capability by imaging neurons, dendrites and spines in the millimeter field-of-view, and simultaneous recording calcium activities, through a gradient-index lens, of approximately 400 neurons in the dorsal hippocampal CA1 in a freely moving mouse. Integrated with a detachable 1.2-g fast z-scanning module, it enables a 1000 × 788 × 500 µm3 volumetric neuronal imaging in the cerebral cortex. Thus, millimeter FOV m2PM provides a powerful tool for deciphering neuronal population dynamics in experimental paradigms allowing for animal’s free movement.
© 2023 Optica Publishing Group under the terms of the Optica Open Access Publishing Agreement
1. Introduction
With the development of miniaturized multiphoton microscopy owning deep penetration and inherent optical sectioning capabilities, it has become possible to obtain high-resolution images of the structures and activities of neurons in the brain of free-moving rodents [1–12]. For instance, using miniature two-photon microscopy (m2PM) capable of volumetric imaging at single-spine and single-neuron resolutions [3–6], researchers have mapped the functional network topography of the medial entorhinal cortex [7,8], deciphered microcircuit dynamics in the dorsomedial prefrontal cortex during social competition [9], and unraveled itch signal processing in the primary somatosensory cortex [10]. We recently demonstrated the capability of miniature three-photon microscopy by imaging calcium activities throughout the entire cortex and dorsal hippocampal CA1 up to 1.2 mm depth in mice [12]. However, existing miniaturized multiphoton microscopes have significantly smaller field of view (FOV) (up to 500 × 500 µm2) [8] than typical miniature single-photon microscopes (up to 1000 × 1000 µm2) [13,14], which means that fewer neurons can be recorded simultaneously. Recently, a miniature single-photon microscope with a larger FOV of a few millimeters has been developed [15]. In addition, existing miniaturized multiphoton microscopes used a fast z-scanning module (ZSM) to achieve a maximum z-scanning range of only 240 µm [8]. It is also desirable to extend the z-scanning range for the investigation of neuronal microcircuitry in three dimensions.
Here we report a millimeter FOV m2PM with the weight of 2.5 grams. An optical configuration for large FOV m2PM is devised with its key components including low-magnification objective with independent excitation numerical aperture (NA) and emission NA, large FOV scan lens and microscanner with a small tilt angle. The 2.5-g m2PM achieved an FOV of 1000 × 788 µm2, comparable with that of typical miniature single-photon microscopy and bench-top two-photon microscopy. It has optical resolutions of 1.47 µm laterally and 24.64 µm axially, and an imaging frame rate of 9 Hz at 600 × 512 pixels. Using our m2PM, subcellular structure of neurons in the millimeter FOV and the individual dendrites and spines in the superficial layer of the cortex were resolved. In addition, we recoded high signal-to-noise ratio (SNR) calcium activities of approximately 400 neurons in the dorsal hippocampal CA1 via a gradient-index (GRIN) lens relay in a free-moving mouse. Integrated with a detachable 1.2-g fast ZSM, it enables a 1000 × 788 × 500 µm3 volumetric neuronal imaging in the cerebral cortex, which is 6.6 times greater than that of existing miniature multi-photon microscopes.
2. Design principle
The illumination arm of m2PM consists of the microelectromechanical systems (MEMS) scanner, scan lens and finite objective, as shown in Fig. 1. Based on geometric optics, assuming that both object and image are in air, the expression of imaging FOV can be derived as
The longitudinal magnification can be expressed as [16]
The expression for excitation NA of the objective can be derived as
where D is the diameter of the incident laser beam and is limited by the mirror size of MEMS. As the lateral magnification m decreases, so does the excitation NA. For an average excitation power, $\left\langle P \right\rangle$, the number of fluorescent photons generated per fluorophore per unit time under two-photon excitation, F, can be expressed as [17,18]To take full advantage of the scanning angles of MEMS, the FOV of scan lens should be greater than 15°. In the configuration design of the scan lens, in order to correct the aberration introduced due to the increase of FOV, the Kellner eyepiece structure [20] was adopted and complicated. Finally, the combination of a single lens and two doublets realized the FOV aberration correction of the scan lens. The finite objective had a low magnification of 1.6 and adopted a double telecentric structure to ensure the constant size of FOV during z-scanning. The low-magnification objective corrected the large FOV and z-scanning aberrations by the combination of five single lenses and two doublets. The combination of the large FOV scan lens and low-magnification objective achieved an imaging FOV of 1.2 mm, as shown in Fig. 2(a). There is no vignetting of laser energy within the 1 mm FOV. Limited by the diameter of the miniature scan lens and objective, there is a 4% vignetting of laser energy at the field point of 1.1 mm and a 20% vignetting at the field point of 1.2 mm. The focal spot energy of the on-axis focal point within the Airy Disk (radius: 2.11 µm) is 84%, which is closed to the diffraction limit, and the focal spot energy of the ±15°focal point within the Airy Disk is 80% (Fig. 2(b)). Therefore, for two-photon excitation, the diffusion of focal spot energy due to aberration is negligible. In addition, the maximum optical distortion of the full FOV is less than 1%, which is also negligible (Fig. 2(c)).
Fig. 2. Optical design of the scan lens and objective. (a) Optical layout; (b) Fraction of enclosed energy of the focal spot at field points of 0°, ± 15°; (c) Percent of distortion at different FOV.
3. Results
The optical configuration and prototype of our millimeter FOV m2PM are shown in Fig. 3(a), (b). The headpiece weighed 2.5 grams, and was 16 × 9 × 25 mm3 in size. The illumination arm consists of a hollow fiber for 920-nm transmission, a collimator, a mirror, a MEMS with a small tilt angle, a large FOV scan lens, a dichroic mirror and a low-magnification objective. The fluorescence collection arm consists of the objective, the dichroic mirror, a condenser and a supple fiber bundle (SFB). The low-magnification objective has a lateral magnification of 1.6, a working distance of 1.08 mm, and an excitation NA of 0.25. By designing the fluorescence collection arm separately, the emission NA of the objective was increased to 0.45 to collect more fluorescence (Fig. 3(a)). Considering scanning speed and incident light diameter, the MEMS scanner (A3I12.2-1200AL-LCC20-TW, Mirrorcle) with a resonant frequency of 2.9 kHz is used, and provides a 1.2-mm diameter mirror with the scanning angles ${\theta _x}$= ± 13.4° and ${\theta _y}$= ± 10°. Limited by the size of the headpiece, the focal length of the scan lens was designed to be 3.7 mm. According to Formula (1) and as shown in Fig. 3(c), when the tilt angle of MEMS is 20°, the imaging FOV in the x directions can reach 1 mm. To overcome the space interference, a mirror-folded optical path was added between the collimator and MEMS. The MEMS scanning angles were fully utilized by designing a miniature scan lens with the FOV greater than ±15°. After the optimal design of scan lens and objective, the design resolution is close to the diffraction limit in the millimeter FOV (Fig. 3(d)). Fluorescence from the millimeter FOV is collected into a SFB with an optical diameter of 1.3 mm by a condenser after the dichroic mirror. Using 380-nm fluorescent beads placed under a coverslip, we determined on-axis resolutions of our m2PM to be 1.47 µm laterally and 24.64 µm axially in water, close to theoretical diffraction limit [21] (Fig. 3(e)). Using our m2PM, neurons and dendrites in the cerebral cortex were imaged at a 1000 × 788 µm2 FOV, and spines were resolved in the zoomed-in view (Fig. 3(f), (g)).
Fig. 3. Design and performance of our millimeter FOV m2PM. (a) Optical configuration of the m2PM; (b) Picture of the m2PM; (c) X-direction imaging FOV changes with the tilt angle of MEMS; (d) Variation of Modulus of the Optical Transfer Function (OTF) with spatial frequency at field points of 0µm, 250µm, 500µm; (e) x-y (above) and y-z (below) images of a 380-nm fluorescent bead; (f) x-y images of neurons labeled with GCaMP6s in the cortex of an awake mouse at depths of 210 µm with a 1000 × 788 µm2 FOV (600 × 512 pixels); (g) Left, x-y images of dendrites labeled with YFP in the superficial layer of the cortex of an awake Thy1-YFPH transgenic mouse with a 1000 × 788 µm2 FOV (1400 × 1195 pixels), right, x-y images of dendrites and spines with a 550 × 433 µm2 FOV (1200 × 1024 pixels).
Using a GRIN lens (CLHS100GFT088, Gofoton) with a diameter of 1 mm for relay imaging, we achieved a large field of view observation of hippocampal neurons by our m2PM (Fig. 4). The imaging FOV is 1000 × 788 µm2 and imaging frame rate is 9 Hz at 600 × 512 pixels. Due to vignetting by the GRIN lens itself, the effective FOV is about 900 × 788 µm2. We recorded the calcium activities of about 400 neurons in the dorsal hippocampal CA1 for 18.3 minutes in a freely moving mouse. The mean SNR of raw data for these neurons calcium transients was 9.2, and 98% of the neurons calcium transients had a SNR greater than 3 (Fig. 4(e)).
Fig. 4. Dorsal hippocampal CA1 imaging using our m2PM via a GRIN lens relay in a freely moving mouse. (a) Optical layout of the m2PM combined with a GRIN lens; (b) a snapshot and track of a representative mouse in a box with 32 cm long and 25 cm wide, and the track was recorded in 18.3 minutes; (c) large FOV imaging of hippocampal neurons, left, single frame image, right, segmentation maps of 395 neurons recorded in 18.3 minutes; (d) Ca2+ traces of all 395 active neurons expressing GCaMP6s in (c). The optical power after the objective is 70 mW; (e) SNR of single-neuron calcium transients during the 18.3 minutes recording (n = 385 neurons, one mouse). Black center line, median; limits, 75% and 25%; whiskers, maximum and minimum.
Based on Formula (3), reducing the lateral magnification of the objective can increase the z-scanning range. By integrating with a 1.2-g detachable ZSM consisting of a fast electrically tunable lens (ETL, A-25H1-D0, Corning) and a pair of relay lenses, we showed that the change in ETL diopter from -32 to +42 translates into a focal plane shift from -500 µm to 0 µm (Fig. 5). The headpiece weighed 3.7 g after adding the ZSM (Fig. 5(b)). The ETL is conjugated with MEMS through the pair of relay lenses. Due to space constraints, the pair of relay lenses was designed in an asymmetric configuration that deviated from the 4F (conjugated) configuration. In the change of ETL diopter from -32 to +42, the designed resolution remains constant and is close to the diffraction limit (Fig. 5(d)). Using the m2PM with the ZSM, we achieved imaging of a 1000 × 788 × 500 µm3 volume in the cerebral cortex (Fig. 5(e)).
Fig. 5. 3D volume imaging of 1000 × 788 × 500 µm3 in the cortex of an awake mouse by our m2PM integrating with a detachable ZSM. (a) Optical layout of our m2PM integrating with the detachable ZSM. (b) Structure and picture of our m2PM with the ZSM; (c) Corresponding relation between ETL diopter and focal plane position, obtained by optical path simulation; (d) Modulus of the OTF at different diopters of ETL. (e) 1000 × 788 × 500 µm3 volume imaging of neurons labeled with GCaMP6s in the cortex by our m2PM with the ZSM, and x-y images of neurons at representative depths of 90 µm, 200 µm, 310 µm and 430 µm with a 1000 × 788 µm2 FOV (600 × 512 pixels). The optical powers after the objective were 30 mW, 70 mW, 105 mW, and 145 mW respectively at the four representative depths.
4. Discussion and conclusion
We devised the optical configuration suitable for large FOV miniaturized multi-photon microscope. Its key elements include a low-magnification objective with independent excitation NA and emission NA, a large FOV scan lens, and a MEMS with a small tilt angle. The above results have demonstrated that our m2PM can achieve a FOV of 1000 × 788 µm2 while maintaining the lateral optical resolution less than 1.5 µm. Due to deep penetration and inherent optical sectioning capabilities, our millimeter FOV m2PM enables fine imaging of neuronal subcellular structures at spinal resolution, recording high SNR calcium activities of neurons in the dorsal hippocampal CA1 via a GRIN lens relay, which have significant advantages over miniature single-photon microscopes. Another significant advance is that by reducing the magnification of objective, the z-scanning range of the ZSM is expanded from 240 µm to 500 µm, enabling a 1000 × 788 × 500 µm3 volumetric neuronal imaging in the cerebral cortex, which is 6.6 times greater than the imaging volume of existing miniature multi-photon microscopes.
Our 2.5-gram millimeter FOV m2PM with its optical configuration maximizing the FOV is suitable for imaging fine structures and activities of neuron populations in the cerebral cortex and the dorsal hippocampus of freely behaving mice. We anticipate this new large FOV m2PM will greatly facilitate neuroscientists for investigation of neuronal population dynamics and microcircuitries in freely moving animals engaging in paradigms such as learning and memory, social interaction and fear conditioning. In addition, the optical design principles established here may find broader applications for future development of multi-photon miniscopes with large FOV and large z-scanning range.
5. Methods
A. The m2PM system
The system-level configuration and technology of our m2PM were similar to those of previously developed m2PM [3,4]. A femtosecond pulsing laser (ALCOR-920-1, Spark) was used to produce femtosecond laser pulses at a center wavelength of 920 nm. The maximum average power is about 1 W with a repetition rate of 80 MHz and pulse width of 120 fs. An acousto-optic modulator (MT110-B50A1.5-IR-HK, Opto-electronique) were used to control the optical power. The laser was coupled into a hollow-core fiber for 920-nm transmission by a lens. We designed the optical configuration of the headpiece using Zemax optical design software (OpticStudio 21.3.2, Zemax). The fiber terminal was pre-aligned with the collimator (#84128, Edmund Optics). The diameter of the collimated beam was 1.2 mm. After mirror and MEMS reflection, the laser beam passed through the large FOV scan lens, dichroic mirror, and finally was focused on the brain tissue by the low-magnification objective. All optical elements were coated with anti-reflection film. The objective and scan lens were checked by an eccentricity detector to ensure coaxiality. The precise structure of the headpiece ensured the assembly accuracy of each optical element. Finally, the alignment of the optical pupil was ensured by fine-tuning the position and angle of the MEMS.
The scattered fluorescence was collected by the objective and a condenser to the SFB, and then transferred by the SFB to the photomultiplier tube (PMT) module. The SFB has the same flexibility as those developed previously [3,4] with an effective optical diameter of 1.3 mm and an NA of 0.57. In the PMT module, fluorescence was firstly collimated by a collimating lens and passed through a low-pass filter (FF01-720/SP-25, Semrock) to remove the laser light. Then the green fluorescence entered the PMT (H10770PA-40, HAMAMATSU) through a bandpass filter (FF01-520/70-25, Semrock).
We have developed a 1.2-g detachable ZSM for fast multiplane imaging. To reduce the impact of the headpiece weight and slight shift in center of gravity after the addition of ZSM, we tied a string to the SFB. The string was supported by two pulleys, with a weight attached to the other end of the string, which counteracted the weight of the headpiece, thus reducing the effect on the mice's free behavior.
B. Animals
C57BL/6 mice (8-16 postnatal weeks, 25-30 grams of weight) and Thy1-YFPH transgenic mice (8–16 postnatal weeks, 25-30 grams of weight) were used in these experiments. All procedures, including animal breeding and experimental manipulation, were approved by the Peking University Animal Use and Care Committee and complied with the Association for Assessment and Accreditation of Laboratory Animal Care standards. Animals were housed with a 12-h light/dark cycle at 22 °C with 50% humidity and had free access to food and water.
C. Virus injection
Mice were anesthetized with 1.5% isoflurane in the air at a flow rate of 0.4 L/min and mounted on a stereotactic frame, AAV2/9.hSyn.GCaMp6s.WPRE.hGH.pA was injected into M1 or CA1 using the stereotactic coordinates (AP: + 0.7 mm, ML: -1.7 mm, DV: -0.4 mm for M1; AP:-2.2 mm, ML:-1.8 mm, DV:-1.5 mm for CA1). The titer of the viruses was adjusted to 2 × 1012 vg/mL, a volume of 200 nl was slowly injected per site at a speed of 20 nl/min. The injection needle was kept in each site for another 5 min after injection.
D. Surgical procedure
One week after virus injection, mice were anesthetized with isoflurane, cranial window was prepared on M1, or GRIN lens was implanted above CA1. For cranial window preparation, a headpiece baseplate was attached to the skull with cyanoacrylate, then reinforced with dental cement mixed with carbon powder, a round cranial window (4.3 mm diameter) was centered over M1, a piece of the skull (4.3 mm diameter) was carefully removed with the dura remained intact, a piece of glass coverslip (4.3 mm diameter, CG00C2, No. 0 BK-7, 85–115 µm, Thorlabs) was placed on the cranial window and sealed with cyanoacrylate and dental cement. For GRIN lens implantation, a round cranial window (1.2 mm diameter) was centered over injection site for CA1, bone debris and dura were carefully cleared, the cortex and corpus callosum above CA1 were asperated, 1-mm-diameter GRIN lens was implanted above CA1, sealed with cyanoacrylate and dental cement, a headpiece baseplate was attached to the skull.
E. Placement of the headpiece
The protocol for mounting, dismounting, and remounting the headpiece consisted of four steps, as described previously [4]. Step 1, the baseplate with a coverslip is fixed over a cranial window with dental cement. Step 2, positioning the headpiece. The headpiece is first screw-fastened to its holder and then placed onto the baseplate using a triaxial motorized stage. The triaxial motorized stage moves the headpiece precisely to focus on the depth we want to locate. Once a region of interest is found, the holder is fixed to the baseplate with dental cement. Step 3, dismounting the headpiece by unscrewing and unplugging from the holder. Step 4, to remount, a drop of water is placed on the coverslip of the baseplate; then, the headpiece is plugged in and screw-fastened to the holder. In most experiments, we could find the same focal plane used previously since the holder was fixed firmly.
F. Data processing and analysis
The open-source software Image J was used for averaging, cropping, registration, pseudocolor-coding, and three-dimensional projection of raw images. All cross-section profiles, statistical graphs, traces of calcium transients, and other coordinate graphs were drawn with Origin 9 (OriginLab) and GraphPad Prism 8 (GraphPad Software). All three-dimensional reconstruction figures were made using Imaris (Bitplane). Calcium imaging data were analyzed offline with custom-written programs in MATLAB (MathWorks). We used an annular ring subtraction method to present calcium signals as relative fluorescence changes (ΔF/F).
G. SNR calculation of neuronal activity
We first estimated events for every ΔF/F trace by MLspike [22]. We kept the events larger than one standard deviation over the mean of ΔF/F trace. SNR was calculated as the signal ratio over the noise's standard deviation. The signal was determined as the mean amplitude of events. The noise statistics were extracted from 15-second ΔF/F episodes, which had to lie at least 10 seconds before and 40 seconds after any event [7].
Funding
STI2030-Major Projects (2021ZD0202205, 2022ZD0212100); National Natural Science Foundation of China (31327901, 31830036, 32293211, 61975002, 8200907151); CAMS Innovation Fund for Medical Sciences (2019-I2M-5-054); National Postdoctoral Program for Innovative Talents (BX20190011).
Acknowledgments
We thank Dr. D. Li from Peking University, Dr. C. Guo from Beijing Institute of Collaborative Innovation, Dr. L. Zhang from PKU-Nanjing Institute of Translational Medicine, J. Tian, Y. Hu, Y. Guo, J. Tong, Y. Zhao, J. Wen, Y. Zhang from Beijing Transcend Vivoscope Biotech Co.
Disclosures
The authors declare no conflicts of interest.
Data availability
Data underlying the results presented in this paper are not publicly available at this time but may be obtained from the authors upon reasonable request.
References
1. F. Helmchen, M. S. Fee, D. W. Tank, and W. Denk, “A miniature head-mounted two-photon microscope high-resolution brain imaging in freely moving animals,” Neuron 31(6), 903–912 (2001). [CrossRef]
2. S. Chen, Z. Wang, D. Zhang, A. Wang, L. Chen, H. Cheng, and R. Wu, “Miniature fluorescence microscopy for imaging brain activity in freely-behaving animals,” Neurosci. Bull. 36(10), 1182–1190 (2020). [CrossRef]
3. W. Zong, R. Wu, M. Li, Y. Hu, Y. Li, J. Li, H. Rong, H. Wu, Y. Xu, Y. Lu, H. Jia, M. Fan, Z. Zhou, Y. Zhang, A. Wang, L. Chen, and H. Cheng, “Fast high-resolution miniature two-photon microscopy for brain imaging in freely behaving mice,” Nat. Methods 14(7), 713–719 (2017). [CrossRef]
4. W. Zong, R. Wu, S. Chen, J. Wu, H. Wang, Z. Zhao, G. Chen, R. Tu, D. Wu, Y. Hu, Y. Xu, Y. Wang, Z. Duan, H. Wu, Y. Zhang, J. Zhang, A. Wang, L. Chen, and H. Cheng, “Miniature two-photon microscopy for enlarged field-of-view, multi-plane and long-term brain imaging,” Nat. Methods 18(1), 46–49 (2021). [CrossRef]
5. C. J. Engelbrecht, R. S. Johnston, E. J. Seibel, and F. Helmchen, “Ultra-compact fiber-optic two-photon microscope for functional fluorescence imaging in vivo,” Opt. Express 16(8), 5556–5564 (2008). [CrossRef]
6. W. Piyawattanametha, E. D. Cocker, L. D. Burns, R. P. Barretto, J. C. Jung, H. Ra, O. Solgaard, and M. J. Schnitzer, “In vivo brain imaging using a portable 2.9 g two-photon microscope based on a microelectromechanical systems scanning mirror,” Opt. Lett. 34(15), 2309–2311 (2009). [CrossRef]
7. H. A. Obenhaus, W. Zong, R. I. Jacobsen, T. Rose, F. Donato, L. Chen, H. Cheng, T. Bonhoeffer, M. B. Moser, and E. I. Moser, “Functional network topography of the medial entorhinal cortex,” Proc. Natl. Acad. Sci. U. S. A. 119(7), e2121655119 (2022). [CrossRef]
8. W. Zong, H. A. Obenhaus, E. R. Skytøen, H. Eneqvist, N. L. de Jong, R. Vale, M. R. Jorge, M. B. Moser, and E. I. Moser, “Large-scale two-photon calcium imaging in freely moving mice,” Cell 185(7), 1240–1256.e30 (2022). [CrossRef]
9. C. Zhang, H. Zhu, Z. Ni, Q. Xin, T. Zhou, R. Wu, G. Gao, Z. Gao, H. Ma, H. Li, M. He, J. Zhang, H. Cheng, and H. Hu, “Dynamics of a disinhibitory prefrontal microcircuit in controlling social competition,” Neuron 110(3), 516–531.e6 (2022). [CrossRef]
10. X. Chen, Y. Liu, N. Xu, and Y. Sun, “Itch perception is reflected by neuronal ignition in the primary somatosensory cortex,” Natl. Sci. Rev. 9(6), nwab218 (2022). [CrossRef]
11. A. Klioutchnikov, D. J. Wallace, M. H. Frosz, R. Zeltner, J. Sawinski, V. Pawlak, K. M. Voit, P. S. J. Russell, and J. N. D. Kerr, “Three-photon head-mounted microscope for imaging deep cortical layers in freely moving rats,” Nat. Methods 17(5), 509–513 (2020). [CrossRef]
12. C. Zhao, S. Chen, L. Zhang, D. Zhang, R. Wu, Y. Hu, F. Zeng, Y. Li, D. Wu, F. Yu, Y. Zhang, J. Zhang, L. Chen, A. Wang, and H. Cheng, “Miniature three-photon microscopy maximized for scattered fluorescence collection,” Nat. Methods 20(4), 617–622 (2023). [CrossRef]
13. J. A. Lecoq, R. Boehringer, and B. F. Grewe, “Deep brain imaging on the move,” Nat. Methods 20(4), 495–496 (2023). [CrossRef]
14. K. K. Ghosh, L. D. Burns, E. D. Cocker, A. Nimmerjahn, Y. Ziv, A. E. Gamal, and M. J. Schnitzer, “Miniaturized integration of a fluorescence microscope,” Nat. Methods 8(10), 871–878 (2011). [CrossRef]
15. J. R. Scherrer, G. F. Lynch, J. J. Zhang, and M. S. Fee, “An optical design enabling lightweight and large field-of-view head-mounted microscopes,” Nat. Methods 20(4), 546–549 (2023). [CrossRef]
16. J. W. Smith, Modern optical engineering (fourth edition) (SPIE, 2008), pp. 42–45.
17. W. Denk, J. H. Strickler, and W. W. Webb, “Two-photon laser scanning fluorescence microscopy,” Science 248(4951), 73–76 (1990). [CrossRef]
18. J. P. Zinter and M. J. Levene, “Maximizing fluorescence collection efficiency in multiphoton microscopy,” Opt. Express 19(16), 15348–15362 (2011). [CrossRef]
19. C. Xu and W. W. Webb, “Multiphoton excitation of molecular fluorophores and nonlinear laser microscopy,” in Topics in Fluorescence Spectroscopy (volume 5), J. Lakowicz, ed. (Plenum, 1997), pp. 471–479.
20. M. Laikin, Lens Design(4th edn) (CRC, 2007), pp. 121–122.
21. W. R. Zipfel, R. M. Williams, and W. W. Webb, “Nonlinear magic: multiphoton microscopy in the biosciences,” Nat. Biotechnol. 21(11), 1369–1377 (2003). [CrossRef]
22. T. Deneux, A. Kaszas, G. Szalay, G. Katona, T. Lakner, A. Grinvald, B. Rózsa, and I. Vanzetta, “Accurate spike estimation from noisy calcium signals for ultrafast three-dimensional imaging of large neuronal populations in vivo,” Nat. Commun. 7(1), 12190 (2016). [CrossRef]
