1. Introduction

High-speed three-dimensional (3D) observations using light microscopy are crucial for exploring the dynamic behavior and functions of biological specimens in the field of life science. For 3D observations, laser scanning fluorescence microscopes with a confocal pinhole to enable optical sectioning have been widely utilized [1]. However, this imaging technique is inherently time-consuming because it requires the stacking of multiple images by changing the observation plane to construct 3D images.

To increase the observation speed, various approaches have been developed [2]. Techniques utilizing, for example, multiple focal spots by a spinning disk [3–6], line scanning [7–10], light sheet illumination [11–17], and temporal focusing [18], can rapidly illuminate or excite a specific observation plane, effectively improving the imaging speed. Generally, these approaches still require the mechanical movement of the observation plane for 3D image construction, which eventually limits the temporal resolution of 3D imaging. Some recent light-sheet-microscopy techniques eliminate mechanical movement using an extended depth-of-field for detection [11] or axial movement using acousto–optic systems [12] to further increase the acquisition speed. However, because of the need for orthogonal configuration in the detection and illumination axes for light sheet illumination, this imaging technique imposes a limitation on the available numerical aperture (NA) of the detection objective lens.

Alternatively, the capturing of multiple axial planes simultaneously without changing the observation plane has been realized using diverse microscopy techniques exploiting specialized optical systems [19–22], computational approaches (e.g., light field imaging) [23–25], spatiotemporally multiplexing of an excitation beam [26–29], or inclined excitation and/or detection configuration [30–34]. Some of these techniques have achieved 3D imaging speeds of over 100 volumes per second (VPS) [9,12,33]. However, the complexity of the acquisition system or the required precise optical alignment of the system may restrict their imaging applications.

The application of a focal spot with an extended focal depth, formed by focusing an annular beam to produce a Bessel beam [35–40], is another promising approach that enables the acquisition of volumetric images of a specimen from a single raster scan of the focal spot. Although this method is readily introduced in the conventional framework of laser scanning microscopy and offers high spatial resolution in the lateral direction, it only records the projected images of a specimen along the axial direction. This drawback has recently been addressed by incorporating a wavefront-engineering approach for fluorescent signals [41,42]. By modulating the spatial-phase distribution of fluorescence signals based on the multiplexed computer-generated hologram (CGH) principle [42], the converted fluorescence wave exhibits lateral shifting behavior as it propagates [43]. Consequently, the depth information of the fluorescence signal can be retrieved as lateral information on the detection plane by imposing the lateral shifting behavior. This approach allows us to reconstruct 3D images with lateral and axial spatial resolutions comparable to those obtained using conventional laser scanning microscopy. However, due to the presence of widespread sidelobes around the center spot of the Bessel beam, this approach is restricted to microscopy techniques that employ nonlinear excitation processes, such as multiphoton excitation. Otherwise, the fluorescence waves excited by the sidelobes overlap on the detector plane, causing serious artifacts in the resulting images. This limits the applicability of this method to multiphoton excitation microscopy, where the impact of the sidelobes is effectively reduced.

In this study, we propose a method for wavefront-engineered light needle fluorescence microscopy that enables rapid 3D observation by one-photon excitation. We introduce an excitation beam, named a graded arc (GA) beam, which produces a Bessel beam-like focal pattern with reduced sidelobes in a specific lateral direction. The GA beam used in the proposed microscope technique can significantly alleviate the artifact caused by the sidelobes even in one-photon excitation fluorescence imaging. The proposed method enables the 3D observation of biological specimens from a single raster scan of the needle spot produced by the GA beam. In addition, using a commonly used deconvolution process, we obtained reconstructed 3D images with a spatial resolution comparable to that obtained using conventional confocal microscopy. Furthermore, the 3D tracking of multiple 200-nm particles suspended in water was achieved at a rate of 50.8 VPS by the continuous scanning of a needle spot under a conventional laser scanning microscopy framework. The proposed method will apply to not only fluorescence imaging but also other imaging techniques that detect linear response signals, such as reflected or scattered light in laser scanning microscopy.

2. Experimental setup

2.1 Formation of the graded arc beam

Figure 1(a) shows a schematic of the proposed method. As previously demonstrated [42], the wavefront of the emitted fluorescence signals is spatially modulated by a spatial light modulator (SLM) that displays a phase pattern designed based on a multiplexed CGH (see Appendix A). This wavefront modulation induces the linear shift of a point image along the lateral direction [H axis in Fig. 1(a)] in accordance with the depth position (z) of the point source at the focus. However, in one-photon excitation fluorescence imaging, the sidelobes of the excitation Bessel beam result in unfavorable fluorescence signals, which can be detected by the adjacent detection channels of a 1D array detector aligned along the H axis. Thus, the presence of sidelobes causes serious artifacts in the reconstructed images.

 

Fig. 1. Experimental setup and evaluation of the lateral shift behavior induced by wavefront modulation based on a multiplexed CGH. (a) Schematic of the proposed microscopy system. (b) Conceptual diagrams of the BA (left panel) and GA beams (right panel) at the pupil plane. (c)–(e) Example of laterally shifting point images on the detection plane simulated when a point source is located at (c) z = 5.5 µm, (d) z = −0.5 µm, and (e) z = −5.5 µm. Scale bar in (c)–(e) is 500 µm. (f) Intensity distribution of the Hz plane, reconstructed by exploiting the intensity profile along the white dashed lines by changing the z position of a point source. Horizontal and vertical scale bars are 500 and 5 µm, respectively.

Download Full Size | PDF

To eliminate these artifacts, it is unnecessary to suppress all the sidelobes of the Bessel beam. It will be sufficient to reduce them only along the H axis, as illustrated in the top right panel of Fig. 1(a). This is possible only in the present imaging technique, where the depth information is detected as spatially transposed signals along one direction (H axis) on the detection plane. The suppression of the sidelobes along a specific direction is achieved simply by focusing an arc-shaped beam formed by an annular amplitude mask with a defect angle [44]. Notably, such an arc-shaped mask has been realized as a binary amplitude mask [see the left pane in Fig. 1(b)]; thus, this beam is referred to as a binary arc (BA) beam in this study. However, the degree of sidelobe suppression and the resultant spot size of the BA beam largely depend on the defect angle, which is arbitrarily determined. In addition, the binary amplitude variation induces additional intensity ripples at the focus. Therefore, we propose a GA beam, which has a pair of opposed arc-shaped patterns similar to those of the BA beam. However, unlike the BA beam, the GA beam achieves a graded amplitude variation on the beam cross-section [see the right panel in Fig. 1(b)]. The azimuthal dependence of the intensity distribution of the GA beam [I_GA(ϕ)] is expressed as ${I_{\textrm{GA}}} \propto {\sin ^2}\phi $, where ϕ is the azimuthal angle of the beam cross-section. This intensity profile is advantageous for drastically reducing the sidelobes along the node direction, as described later.

To generate a GA beam in our system, we exploited the polarization characteristics of SLMs implemented by liquid crystal devices. A linearly polarized, 488-nm continuous wave (CW) laser source (JUNO compact laser, Kyocera) was first converted into an azimuthally polarized beam using a segmented half-waveplate [45]. Thereafter, the converted beam was directed to SLM1 (SLM-100, Santec), which was placed at the position where the pupil plane of a water-immersion objective lens (CFI Apochromat LWD Lambda S 40XC WI, Nikon) with NA = 1.15 was projected using 4f relay systems. As in our previous work [42], the liquid crystal panel of SLM1 was divided into two regions displaying an axicon with a lens pattern (Region #1) and an annular mask (Region #2). The combined use of these patterns in a double-path configuration could efficiently transform an incident Gaussian beam into an annular-shaped beam. For the incidence of an azimuthally polarized beam, only the horizontal component, which corresponds to the direction parallel to the liquid crystal orientation of SLM1, is affected by this phase modulation. This polarization dependence enables the production of a GA-shaped intensity distribution. The residual π phase difference between the arcs, originating from the azimuthal polarization, was compensated for by adding a phase flip pattern in Region #2, enabling the generation of a GA beam with a homogeneous wavefront. We adjusted the inner and outer radii of the annular mask to generate a needle spot with the desired focal depth, measured by the full width at half-maximum (FWHM). In addition, the horizontal spot size of the central spot was designed to be approximately 235 nm, corresponding to that expected when a circularly polarized plane wave is focused with a wavelength of 488 nm under the same condition. For example, the inner and outer radii of the annular mask were set to 0.683a and 0.708a, respectively, where a is the radius of the pupil (2a = 11.5 mm) for a focal depth of 20 µm. See Appendix A for the combinations of the radii that produced a needle spot with other focal depths. The needle-shaped focal spot produced by the GA beam was raster-scanned on samples using two Galvano mirrors (83615KM40B, Cambridge Technology). The laser power was in the range of a few to 500 µW after the objective lens, depending on the experimental conditions.

2.2 Axially resolved, light needle scanning microscopy system

In the proposed system, the fluorescence signals emitted by the needle spot excitation were separated using a dichroic mirror (FF495-Di02-25 × 3, Semrock), and the wavefront was modulated by SLM2 (SLM-100, Santec). Through relay optics, SLM2 was located at the position corresponding to the pupil of the objective lens. Figures 1(c)–1(e) show the numerically simulated point images on the detector plane after applying the multiplexed CGH when a point source is located at different depth positions along the optical axis of the objective lens. In this simulation, we assumed that the fluorescence signals emitted from a point source with a wavelength of 515 nm were recorded using imaging optics with a magnification of 80. We considered 32-multiplexed CGH for detecting the depth range (d_range = 31 µm), which replicates a point image 32 times at laterally shifted positions along the H axis as the axial position (z) varies, as shown in Fig. 1(f). See Appendix A for a detailed description of the multiplexed CGH principle. By placing a 32-channel 1D array detector at the image plane, the fluorescence signals from different depth positions could be separately detected and their depth information could be retrieved. The actual design of the multiplexed CGH employed is described in Appendix A. The 1D array detector was composed of a custom-made optical fiber bundle with 32 multimode fibers (each having a core diameter of 60 µm) and aligned one-dimensionally with a pitch of 75 µm. The opposite end of the fiber bundle was coupled to a 32-channel linear array photomultiplier tube module (H11460-20, Hamamatsu). The output signals were recorded using a data-acquisition system in synchronization with Galvano mirrors to reconstruct 3D images.

3. Result

3.1 Evaluation of the sidelobe suppression

We examined the central spot size and the effect of sidelobe suppression for the GA and BA beams by calculating the intensity distribution of the focal spot, which corresponds to the point spread function (PSF) for excitation, as shown in Fig. 2(a). The degree of sidelobe suppression was evaluated from the ratio of the maximum intensity of the first sidelobe (I_s) to the central peak intensity (I_c), expressed as I_s/I_c. The maximum sidelobe intensity was evaluated along the suppressed direction (y axis), as shown in Fig. 2(b). A low intensity ratio indicates a large sidelobe suppression. The FWHM values of the central spot were measured for both the sidelobe-suppressed direction (y axis) and the nonsuppressed direction (x axis), which were normalized to 235 nm, the designed central spot size of a normal Bessel beam. For the BA beam, the spot size and the degree of the sidelobe suppression were evaluated by varying the defect angle from 0° to 90°, with a step of 10°. Figures 2(c) and 2(d) show the evaluated intensity ratio and the FWHM values, respectively, as a function of the defect angles of the BA beam. For the BA beam, sidelobe suppression is promoted at large defect angles (>∼50°) [black line in Fig. 2(c)]; however, the spot size along the y axis also increases significantly [black solid line in Fig. 2(d)]. In contrast, the GA beam can further suppress the sidelobes [red line in Fig. 2(c)] while maintaining a moderately small central spot in the y axis [red solid line in Fig. 2(d)]. Thus, the GA beam has a superior ability to form a needle spot with reduced sidelobes along one direction on the beam cross-section. Notably, the spot size of the GA beam along the x axis [dashed red line in Fig. 2(d)] slightly decreased compared with that for a normal Bessel beam. By contrast, the spot size along the y axis [red solid line in Fig. 2(d)] is 1.2 times larger than that of a normal Bessel beam, which potentially degrades resolution along the y axis.

 

Fig. 2. Evaluation of the excitation PSFs produced by the GA and BA beams. (a) Calculated intensity distributions at the foci of the BA (left) and GA beams (right). (b) One-dimensional intensity profile in the y axis of the focal plane. (c) Degree of sidelobe suppression, measured based on the ratio of the maximum sidelobe intensity to the central peak intensity. (d) The normalized FWHM values of the central spot as a function of the defect angle for the BA beam. (e), (f) Measured excitation PSF of the GA beam on the xy plane with z = 0 (e) and the zy plane (f). (g), (h) Measured 1D intensity profiles (filled circles) and the simulated results (solid lines) along the x axis (g) and the z axis (h). Scale bars in (a), (e), and (f) are 1 µm.

Download Full Size | PDF

To verify this focusing property, we measured the excitation PSFs by imaging an isolated 200-nm yellow–green (YG) fluoresce bead fixed on a coverslip embedded in glycerin. Note that the measurement of the PSFs was conducted using a normal detection path with a single-point detector (R10467U-40, Hamamatsu), which was accomplished by inserting a mirror after the dichroic mirror to switch the optical path. Figures 2(e) and 2(f) show the intensity distribution on the xy and zy planes of a measured excitation PSF produced by a GA beam with a focal depth of 20 µm. The corresponding intensity profiles (red dots) along the y and z axes, as well as the calculated profiles (solid line), are plotted in Figs. 2(g) and 2(h), respectively. The measured central spot size along the y axis was 317 nm, and the degree of sidelobe suppression was estimated to be 0.063. These values agree well with the expected values (283 nm and 0.048). In addition, the measured focal depth was estimated to be 20.6 µm, which is almost identical to that of the designed needle spot. This system generated a visible needle spot with reduced sidelobes at the focus by focusing a GA beam, as expected.

3.2 PSF of the entire system

In principle, the emission PSF in our system is represented as a point image after the modulation of the wavefront of the fluorescence emission by the multiplexed CGH. Therefore, the PSF of the entire system depends on both the excitation and emission PSFs, and the signal detection by the 1D array detector is also considered. To evaluate the PSF of the proposed microscope system, we acquired 3D images of an isolated 200-nm YG bead by applying the 32-multiplexed CGH with d_range = 31 µm, as shown in Fig. 3. The image of an isolated YG bead located at the focus (z = 0) was successfully reconstructed through a single 2D scan of a needle spot with a focal depth of 31 µm. For a normal Bessel beam excitation [Fig. 3(a)], the bead images on the xy planes feature an artifact, indicated by an arrowhead in each panel, which moves along the y axis as the z position changes. This intensity variation is also confirmed by the diagonal intensity tail on the zy plane of Fig. 3(a), as indicated by the double-headed arrow. This artifact is caused by the sidelobes of the normal Bessel beam in one-photon excitation imaging. In contrast, as shown in Fig. 3(b), the use of the GA beam significantly diminished the appearance of this artifact with a reduction of up to 73% (see Appendix B). As a result, an improved bead image was obtained even in one-photon excitation imaging. Note that the residual sidelobes along the x axis are due to the sidelobes along the nonsuppressed direction inherently caused by the GA beam [x axis in Fig. 2(a)]. This leads to the reduction of the spatial resolution along the x axis (see Section 3.3).

 

Fig. 3. Evaluation of the PSFs of the entire system with the multiplexed CGH with d_range = 31 µm, measured by imaging an isolated 200-nm YG bead using (a) a normal Bessel beam and (b) a GA beam. The reconstructed intensity distributions of the bead image on the zy and zx planes are shown in the first and second rows, respectively. The xy planes at z = −2, −1, 0, 1, and 2 µm are shown in the third row of each figure. (c) Deconvolution with 20 iterations for (b). Each scale bar is 2 µm. (d)–(f) Intensity profiles of the measured PSFs along the x axis (d), y axis (e), and z axis (f). The red and green lines correspond to the profiles shown in (b) and (c), respectively. The blue lines correspond to the numerically simulated profiles of the PSFs using the GA beam.

Download Full Size | PDF

Figures 3(d)–3(f) show the intensity profiles of the measured PSF produced by the GA beam (red lines) along the x, y, and z axes, as well as those obtained by numerical simulations (blue lines). The experimental profiles agree with the simulated results. Nonetheless, the resultant PSF still contains the sidelobes along the x axis and the measurable intensity tail along the z axis, even when using a GA beam. The residual sidelobes along the x axis are due to the sidelobes along the nonsuppressed direction of the GA beam [x axis in Fig. 2(a)]. In principle, these sidelobes can be reduced by employing a 1D array detector with a fiber bundle having a small core diameter, with each fiber channel behaving as a confocal pinhole. In this setup, the diameter of each fiber channel corresponds to ∼1.4 Airy units (AU), calculated for an excitation wavelength of 488 nm. In contrast, the intensity tail along the z axis is attributed to the emission PSF converted by the multiplexed CGH. The applied phase pattern is produced by the superposition of different defocus and tilt wavefronts. Due to this multiplexing procedure, the emission PSF at the image plane typically involves weak but unignorable defocus components near the central spot, as shown in Figs. 1(c)–1(e). Such a peripheral intensity distribution is detected by neighboring detector channels and appears as the intensity tail along the z axis in the resultant PSF.

To further reduce these residual sidelobes and intensity tails, simple post-processing procedures, such as deconvolution, can be utilized. We applied the Richardson–Lucy (RL) algorithm [46,47] to deconvolve the measured PSF [Fig. 3(b)] using a numerically simulated PSF. Figure 3(c) shows the PSF obtained after deconvolution with 20 iterations. The corresponding intensity profiles along each axis are plotted as green lines in Figs. 3(d)–3(f). As shown in these figures, the residual sidelobes and intensity tails in the measured PSF were effectively reduced by applying a commonly employed deconvolution procedure with a limited number of iterations.

3.3 Spatial resolution

Figure 4(a) shows a typical example of a 3D image of 200-nm YG beads embedded in 4% agarose gel, acquired by scanning a needle spot with a focal depth of 31 µm with wavefront modulation (d_range = 31 µm) for fluorescence emission. The 3D-distributed YG beads were successfully visualized from a single scan of a needle spot. This condition produced a 3D image with a depth range of 31 µm, reconstructed from a 32-channel detector corresponding to an axial pitch of 1 µm [Fig. 4(b)]. To precisely evaluate the spatial resolution of the present setup, as performed in [42], we further constructed interpolated bead images by repeating image acquisitions and changing the observation plane within the depth range of 1 µm, with a pitch of 0.166 µm. Thereafter, the acquired frames were rearranged to form an interpolated image [Fig. 4(c)].

 

Fig. 4. Imaging of 200-nm YG beads embedded in agarose gel. (a) Three-dimensional view of the distribution of YG beads captured by a single scan of a needle spot produced by the GA beam. (b) Magnified view of a bead image in (a). (c) Interpolated image corresponding to the same bead in (b). Scale bars in (b) and (c) are 1 µm. (d)–(f) Bead sizes along the x axis (d), y axis (e), and z axis (f), as a function of the depth position of each bead.

Download Full Size | PDF

By measuring the FWHM values of each constructed bead image, we evaluated the spatial resolution of the system. The blue circles in Figs. 4(d)–4(f) plot the estimated bead sizes, corresponding to the spatial resolution along the x, y, and z axes, as a function of the axial position of each bead. The median sizes in the x, y, and z axes were 0.562 µm (0.525–0.624 µm), 0.386 µm (0.363–0.433 µm), and 1.66 µm (1.37–2.10 µm), respectively, where the numbers in parentheses represent the interquartile range. This result implies that the proposed technique provides a constant spatial resolution without noticeable dependence on the axial position. The slight deviation appearing near the edge of the depth range, particularly for the axial resolution, is partly due to the chromatic dispersion of SLMs for fluorescence signals with a certain bandwidth. The fluorescence signals emitted at a point distance from the focus are detected by the detection channels located near the edge of the array detector, and a substantial lateral shift is required for the fluorescence signals modulated by the multiplexed CGH. In this situation, the chromatic dispersion of fluorescence signals blurs the emission PSFs on the detection plane, deteriorating the spatial resolution along the z axis (see Appendix C).

We compared the spatial resolution of the proposed system with that of a conventional confocal microscope. The spatial resolution measured at different axial positions for confocal imaging using a Gaussian excitation beam under a confocal pinhole of ∼1.6 AU is plotted as gray circles in Figs. 4(d)–4(f). The corresponding median sizes and interquartiles in the x, y, and z axes were 0.284 µm (0.274–0.297 µm), 0.276 µm (0.263–0.290 µm), and 0.707 µm (0.687–0.726 µm), respectively. Compared with confocal microscopy, this imaging technique provides spatial resolutions of the same order of magnitude in all directions. The relatively large differences for the x and z axes are due to the residual sidelobes caused by the GA beam and the intensity tails along the z axis of the PSF, respectively, as mentioned in Section 3.2. To fully exploit the potential spatial resolution, we eliminated these intensity distributions by applying the deconvolution process as mentioned above. The red circles in Figs. 4(d)–4(f) correspond to the resultant spatial resolutions after deconvolution with 20 iterations. The median values and the interquartile estimated in the x, y, and z axes were improved to 0.192 µm (0.179–0.224 µm), 0.237 µm (0.217–0.257 µm), and 1.12 µm (0.969–1.39 µm), respectively. After the deconvolution process, the spatial resolution along the x axis was notably enhanced compared with that along the y axis. This is attributed to the elliptical focal spot having a narrow width along the x axis, produced by focusing the GA beam.

3.4 Observation of biological specimens

To demonstrate the applicability of the present method to biological imaging, we examined a fixed brain slice of a Thy1-YFP-H mouse, expressing enhanced yellow fluorescence proteins (EYFP). The Thy1-YFP-H mice were anesthetized with isoflurane and perfused with phosphate-buffered saline (PBS), followed by 4% paraformaldehyde (PFA, in PBS). The brain was surgically removed and treated with 4% PFA at 4°C overnight, after which the 4% PFA was replaced with PBS. The brain was cut into 200-µm-thick coronal slices and treated with Tissue-Clearing Reagent CUBIC-L (T3740, Tokyo Chemical Industry) at room temperature for a few days, as previously reported [48]. After clearing, they were mounted with CUBIC-L. All animal studies were performed in line with the Animal Research approved by the Institutional Animal Care and Use Committee of the National Institute of Natural Sciences and according to the guidelines of the National Institute for Physiological Sciences (Approval No. 23A062).

The sample was imaged using the GA beam, producing a needle spot with a focal depth of 31 µm under wavefront modulation (d_range = 31 µm). Figure 5(a) shows a reconstructed 3D image projected along the z axis with color-coded depth. Figure 5(b) is a deconvolved image of Fig. 5(a) using RL method with 10 iterations. As an example, the sectional image of Fig. 5(b) at z = 0.5 µm is also shown in Fig. 5(c). The needle spot was raster-scanned in an area of 150 × 150 µm² by 512 × 512 px², with a dwell time of 20 µs. The observation-depth range was 31 µm by 32 px. For comparison, the same region was captured using a conventional confocal microscope with a Gaussian beam, which required 32 scans by changing the observation plane to construct the 3D image [Fig. 5(d)]. The result shown in Fig. 5(a) clearly demonstrates that the present system could visualize the 3D structures of the pyramidal neurons and dendrites in the brain slice. Furthermore, the deconvolution process resulted in an imaging quality comparable to that obtained by conventional confocal microscopy [Fig. 5(d)].

 

Fig. 5. Acquisition of pyramidal neurons in a fixed Thy1-YFP-H mouse brain. (a), (b) Reconstructed 3D image projected along the z axis with color-coded depth (a) and deconvolved volume image using the RL method with 10 iterations (b). (c) Sectional image at z = 0.5 µm, extracted from the deconvoluted volume image. The observation area was 150 × 150 × 31 µm³, acquired by a single scan of a needle spot with a focal depth of 31 µm under the detection condition of d_range = 31 µm. (d) Corresponding image obtained by a conventional confocal microscope using a Gaussian beam, acquired for the same volume with 32-image stacking. (e) Magnified image, corresponding to the region (30 × 30 µm²) indicated by the red rectangle in (b), acquired and reconstructed again using a needle spot with a focal depth of 10 µm and the detection condition of d_range = 10 µm. (f) The deconvolved volume image of (e) with 10 iterations. (g) Sectional image at z = 1.0 µm in (f). (h) The image obtained by conventional confocal imaging for the same region as (e). Scale bars in (a)–(d) are 30 µm and those in (e)–(h) are 5 µm.

Download Full Size | PDF

Figure 5(e) shows a magnified image (30 × 30 µm² by 512 × 512 px²) of the region corresponding to the red rectangle in Fig. 5(b), which was reconstructed by scanning an excitation needle spot with a focal depth of 10 µm and wavefront modulation with d_range = 10 µm. Figure 5(f) is a deconvolved 3D image of Fig. 5(e) with 10 iterations. In this acquisition, we employed a 16-multiplexed CGH, which was designed for 16 detection channels to increase the signal intensity. In principle, an N-multiplexed CGH will reduce the signal intensity on each channel by a factor of N [42]. Thus, the use of the 16-multiplexed CGH instead of the 32-multiplexed one allows us to visualize finer structures with an improved signal-to-noise ratio, although the number of available slices along the axial direction is decreased. Indeed, fine structures, such as spine structures [Fig. 5(f)], were clearly observed by our system as conventional confocal imaging [Fig. 5(h)]. Importantly, the switching of the acquisition area and the depth range was accomplished by changing the phase patterns displayed on SLM1 and SLM2, without any mechanical adjustments of the optical system.

3.5 High-speed three-dimensional tracking

To demonstrate the capability of rapid 3D imaging, we recorded the 3D motion of 200-nm YG beads suspended in water, as shown in Fig. 6(a). The reconstructed volume was 15 × 7.5 µm² by 64 × 32 px², with a depth range of 20 µm, attained by 32-multiplexed CGH with d_range = 20 µm using a needle spot with a focal depth of 20 µm. By conducting continuous raster scans of the needle spot with a dwell time of 4 µs, we achieved continuous and real-time 3D imaging at a rate of 50.8 VPS. This imaging speed enabled the 3D capture of the fast Brownian motion of 200-nm particles, as also shown in Visualization 1. Moreover, the 3D trajectories of the particles with a voxel size of 0.24 × 0.24 × 0.65 µm³ were successfully analyzed from the recorded data using TrackMate [49]. The analyzed results are displayed in Fig. 6(b) and Visualization 2. In this analysis, only the trajectories of a particle that was tracked for more than 10 frames are displayed, and 31 trajectories were extracted. Notably, in this result, despite that no deconvolution process was performed on the obtained images, individual beads were clearly distinguished. This imaging performance is advantageous for recording the dynamic motion of small particles, which can be potentially applied to broad fields, including biopharmaceutical research [50,51].

 

Fig. 6. Tracking of the 3D motion of YG beads suspended in water with an acquisition rate of 50.8 VPS. (a) Three-dimensional view of the reconstructed image at t = 2.16 s. The observation volume is 15 × 7.5 × 20 µm³, with the detection condition being d_range = 20 µm. (b) Three-dimensional trajectory of the Brownian motion of each bead, recorded over 6 s. Only the beads wandering over 10 frames within the observation volume are depicted.

Download Full Size | PDF

Additionally, we evaluated the diffusion coefficient from the 3D trajectory of 11 particles that were tracked for more than 50 frames. Following a similar procedure reported in [52], we analyzed the mean square displacement (MSD) of the selected particles in the x, y, and z axes for each lag time (τ) and estimated the diffusion coefficients from the slope of the MSD (τ) by a linear fitting using the first 10 time-lag points. The averaged diffusion coefficients along x, y, and z axes were 1.94 (±1.00) µm²/s, 1.64 (±0.46) µm²/s, and 2.31 (±1.09) µm²/s, respectively, where the number in the parentheses represents the standard deviations. From this analysis, the 3D diffusion coefficient was estimated to be 1.96 (±0.91) µm²/s, which was almost identical to a theoretical value of 1.86 µm²/s. The theoretical value was calculated using the Stokes-Einstein equation with the parameters of a temperature of 300 K, a viscosity of 0.89 mPa·s, and a hydrodynamic diameter of 265 nm (obtained by the dynamic light scattering measurement for the particle). The slight difference in the estimated diffusion coefficients along the x, y, and z axes can be attributed to the non-identical spatial resolution and the pixel pitch along three axes, associated with the localization accuracy [42].

4. Discussion and conclusion

The use of a GA beam enables 3D imaging in one-photon excitation fluorescence microscopy using a needle spot without serious artifacts caused by the sidelobes of the Bessel beam. The present approach will largely extend the range of choices for the excitation wavelength and types of fluorescence probes. This is a distinctive feature compared to the previous technique [41,42] which could be applied only to multiphoton excitation microscopy using a near-infrared femtosecond laser source. The applicability of the present approach to one-photon excitation microscopy suggests its possible application, beyond fluorescence imaging, to imaging modalities that detect linear response signals, including scattered light. In addition, the use of light with wavelength in the visible region for excitation or illumination will result in better spatial resolution compared with the results of two-photon imaging. In this regard, as mentioned in Section 3.3, the residual sidelobes on the x axis and the intensity tails along the z axis in the resultant PSF need to be addressed to fully exploit the potential spatial resolution. A promising approach demonstrated in this study is the use of deconvolution for the acquisition of raw images. Another solution could be the use of a Bessel beam with suppressed sidelobes in all directions [53–55], although a more precise wavefront control is required for the excitation beam. It is worth mentioning that a GA beam has potential applicability to other imaging techniques beyond light needle microscopy. Light sheet microscopy under one-photon excitation is one of the microscope techniques that needs such a needle spot with an extended focal depth and reduced sidelobes [44]. The use of a GA beam may offer an optimum spot to form thin-sheet excitation also in light sheet microscopy.

The effectiveness of the proposed method was confirmed by its ability to visualize fine structures, such as dendritic spines in a mouse brain sample. The shape and distribution analyses of dendritic spines are particularly crucial for understanding brain function [56,57]. Thus, this method will be highly valuable in neuroscientific research for such analyses. It might also be possible to capture nervous activities using the present method combined with a calcium indicator, e.g., G-CaMP [58]. This combined approach would lead to functional imaging in three dimensions with a high spatio-temporal resolution.

One constraint of the proposed method is the relatively low signal intensity due to the wavefront modulation achieved using the multiplexed CGH. As mentioned before, the signal intensity obtained at each detector, in principle, decreases by a factor of N when using an N-multiplexed CGH. Therefore, observing fluorescence samples with a low quantum efficiency is challenging under large multiplexing conditions. Owing to the trade-off relationship between the signal intensity and multiplicity, the detectable signal intensity can be improved by reducing the multiplicity while decreasing the number of slices. A major advantage of the present method is its controllable multiplicity (N) and observation depth (d_range). These parameters can be readily adjusted by changing the phase pattern displayed on the SLMs. Owing to this flexibility, one can choose an optimized imaging condition depending on the observation objects without mechanical adjustments.

In summary, we proposed a beam, named a GA beam, which can produce a needle spot with suppressed sidelobes along one direction while maintaining a small central spot with an extended focal depth. By applying this beam to wavefront-engineered light needle microscopy, multiple observation planes were simultaneously captured, even for one-photon excitation imaging, without serious artifacts. The present microscopy technique adopting a visible-light GA beam exhibited high spatial resolutions in all directions, enabling the 3D imaging of biological samples with an imaging quality comparable to that obtained using conventional confocal microscopy. Furthermore, this method was applied to capture the 3D motion of 200-nm beads suspended in water, enabling fast 3D tracking of individual particles. The present technique provides superior controllability for switching the observation-depth range and the number of observation planes, which can be applied to various bioimaging scenarios that require rapid acquisition of 3D images.

Appendix

A. Design of the multiplexed CGHs

The fluorescence wavefront is modulated by applying a phase pattern designed using the principle of multiplexed CGH [43]. The phase pattern applied to the emitted fluorescence is designed by the coherent superposition of CGH elements that achieve both defocus correction and wavefront tilt. Readers are referred to [42] for details. Briefly, the designed phase pattern displayed on SLM2 is expressed as follows:

$$\begin{array}{{c}} {\; \; \; \; \; \; \; \; \; \; \; \; \; W({\xi,\; \eta})= \arg\left\{ {\mathop \sum \limits_{i = 1}^N \textrm{exp}[{ - i{\psi_{\textrm{defocus}}}({\xi ,\eta ;{z_i}})+ i{\psi_{\textrm{tilt}}}({\xi ,\eta ;{z_i}})+ 2\pi i{c_i}}]} \right\},\; \; \; \; \; \; \; \; \; \; \; \; \; (1)} \end{array}$$

where (ξ, η) is the coordinate on the SLM2 corresponding to the pupil plane of an objective lens, N is the multiplicity of the CGH element, ${\psi _{\textrm{defocus}}}({\xi ,\; \eta ;{z_i}})$ is the defocus wavefront when a point source is located at z_i in the objective space, ${\psi _{\textrm{tilt}}}({\xi ,\; \eta ;{z_i}})$ is the intended wavefront tilt that shifts the image position of point source z_i laterally on the detector plane, and c_i (= 0–1) is the initial phase value of each CGH element. The purpose of the multiplexed CGH is to image the fluorescence emission from a point source located at z = z_i on the input face of each optical fiber in an N-channeled fiber bundle in accordance with the linear relationship, H = αz, where α is a coefficient. The value of α is determined by the magnification of the imaging system and the pitch of the 1D fiber bundle.

The phase patterns of the multiplexed CGH were prepared according to the imaging conditions, including the observation-depth range (d_range), multiplicity (N), and emission wavelength. In the design of the phase patterns, the initial phase value (c_i) was optimized to avoid unfavorable interference fringes in the resultant emission PSF. To determine the value of c_i, we employed an optimization procedure based on a genetic algorithm. The optimized values of c_i employed in this study are summarized in Table 1.

Table 1. Initial phase values for the multiplexed CGHs employed in this study

View Table

When applying the specific multiplexed CGH to fluorescence signals, we employed a needle spot for excitation with a focal depth compatible with the observation-depth range, d_range. The combinations of the inner and outer radii of an annular mask, displayed on SLM1, for a GA beam to produce a needle spot with a focal depth of 31, 20, 10 µm at the focus are (0.687a for inner, 0.704a for outer), (0.683a, 0.708a), and (0.670a, 0.719a), respectively, where a is the pupil radius of the water-immersion objective lens with NA = 1.15 employed in this study.

B. Artifact reduction in light needle microscopy by a GA beam

We have evaluated the artifact reduction by a GA beam compared to a normal Bessel beam. Figures 7(a) and 7(b) are the maximum intensity projection along the z axis of the measured PSFs [Figs. 3(a) and 3(b)] when using the Bessel beam and the GA beam, respectively. The intensity within 30 px along the x axis from the center, indicated by the yellow rectangle in Figs. 7(a) and (b), were integrated and evaluated along the y axis as shown in Fig. 7(c). The artifacts appearing as sidelobes were clearly reduced by the use of the GA beam. The maximumpeak intenisty of the sidelobes, indicated by the red arrowhead in Fig. 7(c), was reduced by 73%.

 

Fig. 7. Evaluation of the artifact reduction by a GA beam. Maximum intensity projection of the measured PSF produced by the Bessel beam (a) and the GA beam (b). Scale bars in (a) and (b) are 1 µm. (c) Normalized intensity profiles along the y axis were obtained by integrating the intensities within 30 px along the x axis from the center [indicated by the yellow squares in (a) and (b)].

Download Full Size | PDF

C. Effect of chromatic dispersion on the SLMs

We evaluated the effect of the chromatic dispersion of SLMs. To examine this effect, we imaged an isolated 200-nm YG bead without scanning by applying wavefront tilt on SLM2, causing a lateral shift of the bead image along the H axis. The resultant intensity distributions at the detection plane were captured by a charged-coupled device camera, which was placed instead of the 1D array detector. Figures 8(a)–8(e) show the intensity distributions obtained at each tilt angle, and Fig. 8(f) plots the 1D intensity profiles along the H axis. Figures 8(g)–8(l) show the result of a similar evaluation for the case of the insertion of a bandpass filter with a central wavelength of 513 nm and a bandwidth of 17 nm (FF01-513-17-25, Semrock). Without the bandpass filter, the intensity distributions in the H axis were observed to disperse significantly for large tilt angles, indicating the influence of chromatic dispersion by the SLMs. In the present setup, a 32-channel fiber bundle with a pitch of 75 µm, employed at the entrance of the 1D array detector, was used, requiring tilt angles of −0.22° to 0.22°. Thus, it is expected that the influence of chromatic dispersion may deteriorate the axial resolution near the edge of the detection depth range when a high multiplicity is employed.

 

Fig. 8. Evaluation of the effect of chromatic dispersion of SLMs. (a)–(e) Intensity distributions of point images on the detection plane, measured by a 2D camera, when only a wavefront tilt with angles of (a) 0.2°, (b) 0.1°, (c) 0°, (d) −0.1°, and (e) −0.2° are applied for fluorescence. (f) One-dimensional intensity profiles of the measured fluorescent spot along the H axis. (g)–(l) Results obtained in the same manner as (a)–(f); however, a bandpass filter (central wavelength, 513 nm; bandwidth, 17 nm) was inserted in the detection path. Scale bars in (a)–(e) and (g)–(k) are 500 µm.

Download Full Size | PDF

Funding

Fusion Oriented Research for Disruptive Science and Technology (JPMJFR222F); Japan Society for the Promotion of Science (22H01979, 22H04926, 22K06960); Japan Agency for Medical Research and Development (JP19dm0207078).

Acknowledgments

We thank Citizen Watch Co., Ltd. for the provision of the liquid crystal devices.

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. T. Wilson and A. R. Carlini, “Size of the detector in confocal imaging systems,” Opt. Lett. 12(4), 227–229 (1987). [CrossRef]  

2. J. Mertz, “Strategies for volumetric imaging with a fluorescence microscope,” Optica 6(10), 1261–1268 (2019). [CrossRef]  

3. A. lchihara, T. Tanaami, K. lsozaki, et al., “High-speed confocal fluorescence microscopy using a Nipkow scanner with microlenses for 3-D Imaging of single fluorescent molecule in real time,” Bioimages 4(2), 57–62 (1996). [CrossRef]  

4. K. Otomo, T. Hibi, T. Murata, et al., “Multi-point scanning two-photon excitation microscopy by utilizing a high-peak-power 1042-nm Laser,” Anal. Sci. 31(4), 307–313 (2015). [CrossRef]  

5. J. Bewersdorf, R. Pick, and S. W. Hell, “Multifocal multiphoton microscopy,” Opt. Lett. 23(9), 655–657 (1998). [CrossRef]  

6. K. Seiriki, A. Kasai, T. Nakazawa, et al., “Whole-brain block-face serial microscopy tomography at subcellular resolution using FAST,” Nat. Protoc. 14(5), 1509–1529 (2019). [CrossRef]  

7. J.-M. Tsang, H. J. Gritton, S. L. Das, et al., “Fast, multiplane line-scan confocal microscopy using axially distributed slits,” Biomed. Opt. Express 12(3), 1339–1350 (2021). [CrossRef]  

8. C. Martin, T. Li, E. Hegarty, et al., “Line excitation array detection fluorescence microscopy at 0.8 million frames per second,” Nat. Commun. 9(1), 4499 (2018). [CrossRef]  

9. H. Mikami, J. Harmon, H. Kobayashi, et al., “Ultrafast confocal fluorescence microscopy beyond the fluorescence lifetime limit,” Optica 5(2), 117–126 (2018). [CrossRef]  

10. J.-L. Wu, Y.-Q. Xu, J.-J. Xu, et al., “Ultrafast laser-scanning time-stretch imaging at visible wavelengths,” Light: Sci. Appl. 6(1), e16196 (2016). [CrossRef]  

11. O. E. Olarte, J. Andilla, D. Artigas, et al., “Decoupled illumination detection in light sheet microscopy for fast volumetric imaging,” Optica 2(8), 702–705 (2015). [CrossRef]  

12. M. Duocastella, G. Sancataldo, P. Saggau, et al., “Fast inertia-free volumetric light-sheet microscope,” ACS Photonics 4(7), 1797–1804 (2017). [CrossRef]  

13. E. H. K. Stelzer, “Light-sheet fluorescence microscopy for quantitative biology,” Nat. Methods 12(1), 23–26 (2015). [CrossRef]  

14. J. Huisken, J. Swoger, F. Del Bene, et al., “Optical sectioning deep inside live embryos by selective plane illumination microscopy,” Science 305(5686), 1007–1009 (2004). [CrossRef]  

15. Z. Jiao, Z. Zhou, Z. Chen, et al., “Simultaneous multi-plane imaging light-sheet fluorescence microscopy for simultaneously acquiring neuronal activity at varying depths,” Optica 10(2), 239–247 (2023). [CrossRef]  

16. J. M. Girkin and M. T. Carvalho, “The light-sheet microscopy revolution,” J. Opt. 20(5), 053002 (2018). [CrossRef]  

17. S. Takanezawa, T. Saitou, and T. Imamura, “Wide field light-sheet microscopy with lens-axicon controlled two-photon Bessel beam illumination,” Nat. Commun. 12(1), 2979 (2021). [CrossRef]  

18. E. Papagiakoumou, E. Ronzitti, and V. Emiliani, “Scanless two-photon excitation with temporal focusing,” Nat. Methods 17(6), 571–581 (2020). [CrossRef]  

19. P. M. Blanchard and A. H. Greenaway, “Simultaneous multiplane imaging with a distorted diffraction grating,” Appl. Opt. 38(32), 6692–6699 (1999). [CrossRef]  

20. C. Maurer, S. Khan, S. Fassl, et al., “Depth of field multiplexing in microscopy,” Opt. Express 18(3), 3023–3034 (2010). [CrossRef]  

21. S. Abrahamsson, J. Chen, B. Hajj, et al., “Fast multicolor 3D imaging using aberration-corrected multifocus microscopy,” Nat. Methods 10(1), 60–63 (2013). [CrossRef]  

22. A. Badon, S. Bensussen, H. J. Gritton, et al., “Video-rate large-scale imaging with Multi-Z confocal microscopy,” Optica 6(4), 389–395 (2019). [CrossRef]  

23. H. Li, C. Guo, D. Kim-Holzapfel, et al., “Fast, volumetric live-cell imaging using high-resolution light-field microscopy,” Biomed. Opt. Express 10(1), 29–49 (2019). [CrossRef]  

24. Y.-G. Yoon, Z. Wang, N. Pak, et al., “Sparse decomposition light-field microscopy for high speed imaging of neuronal activity,” Optica 7(10), 1457–1468 (2020). [CrossRef]  

25. Z. Zhang, L. Bai, L. Cong, et al., “Imaging volumetric dynamics at high speed in mouse and zebrafish brain with confocal light field microscopy,” Nat. Biotechnol. 39(1), 74–83 (2021). [CrossRef]  

26. S. Weisenburger, F. Tejera, J. Demas, et al., “Volumetric Ca²⁺ imaging in the mouse brain using hybrid multiplexed sculpted light microscopy,” Cell 177(4), 1050–1066.e14 (2019). [CrossRef]  

27. J. Demas, J. Manley, F. Tejera, et al., “High-speed, cortex-wide volumetric recording of neuroactivity at cellular resolution using light beads microscopy,” Nat. Methods 18(9), 1103–1111 (2021). [CrossRef]  

28. A. Cheng, J. T. Gonçalves, P. Golshani, et al., “Simultaneous two-photon calcium imaging at different depths with spatiotemporal multiplexing,” Nat. Methods 8(2), 139–142 (2011). [CrossRef]  

29. X. Yang, H. Zhang, Z. Liu, et al., “Time-stretch chromatic confocal microscopy for multi-depth imaging,” Laser Photonics Rev. 17(12), 2300387 (2023). [CrossRef]  

30. J. Kim, T. Li, Y. Wang, et al., “Vectorial point spread function and optical transfer function in oblique plane imaging,” Opt. Express 22(9), 11140–11151 (2014). [CrossRef]  

31. M. B. Bouchard, V. Voleti, C. S. Mendes, et al., “Swept confocally-aligned planar excitation (SCAPE) microscopy for high-speed volumetric imaging of behaving organisms,” Nat. Photonics 9(2), 113–119 (2015). [CrossRef]  

32. T. Li, S. Ota, J. Kim, et al., “Axial plane optical microscopy,” Sci. Rep. 4(1), 7253 (2014). [CrossRef]  

33. V. Voleti, K. B. Patel, W. Li, et al., “Real-time volumetric microscopy of in vivo dynamics and large-scale samples with SCAPE 2.0,” Nat. Methods 16(10), 1054–1062 (2019). [CrossRef]  

34. S. Yordanov, K. Neuhaus, R. Hartmann, et al., “Single-objective high-resolution confocal light sheet fluorescence microscopy for standard biological sample geometries,” Biomed. Opt. Express 12(6), 3372–3391 (2021). [CrossRef]  

35. E. J. Botcherby, R. Juškaitis, and T. Wilson, “Scanning two photon fluorescence microscopy with extended depth of field,” Opt. Commun. 268(2), 253–260 (2006). [CrossRef]  

36. G. Thériault, Y. De Koninck, and N. McCarthy, “Extended depth of field microscopy for rapid volumetric two-photon imaging,” Opt. Express 21(8), 10095–10104 (2013). [CrossRef]  

37. G. Thériault, M. Cottet, A. Castonguay, et al., “Extended two-photon microscopy in live samples with Bessel beams: steadier focus, faster volume scans, and simpler stereoscopic imaging,” Front. Cell. Neurosci. 8, 139 (2014). [CrossRef]  

38. S. Ipponjima, T. Hibi, Y. Kozawa, et al., “Improvement of lateral resolution and extension of depth of field in two-photon microscopy by a higher-order radially polarized beam,” Microscopy 63(1), 23–32 (2014). [CrossRef]  

39. R. Lu, W. Sun, Y. Liang, et al., “Video-rate volumetric functional imaging of the brain at synaptic resolution,” Nat. Neurosci. 20(4), 620–628 (2017). [CrossRef]  

40. C.-P. Chang, K. Otomo, Y. Kozawa, et al., “Single-scan volumetric imaging throughout thick tissue specimens by one-touch installable light-needle creating device,” Sci. Rep. 12(1), 10468 (2022). [CrossRef]  

41. Y. Kozawa and S. Sato, “Light needle microscopy with spatially transposed detection for axially resolved volumetric imaging,” Sci. Rep. 9(1), 11687 (2019). [CrossRef]  

42. Y. Kozawa, T. Nakamura, Y. Uesugi, et al., “Wavefront engineered light needle microscopy for axially resolved rapid volumetric imaging,” Biomed. Opt. Express 13(3), 1702–1717 (2022). [CrossRef]  

43. T. Nakamura, S. Igarashi, Y. Kozawa, et al., “Non-diffracting linear-shift point-spread function by focus-multiplexed computer-generated hologram,” Opt. Lett. 43(24), 5949–5952 (2018). [CrossRef]  

44. F. O. Fahrbach, V. Gurchenkov, K. Alessandri, et al., “Self-reconstructing sectioned Bessel beams offer submicron optical sectioning for large fields of view in light-sheet microscopy,” Opt. Express 21(9), 11425–11440 (2013). [CrossRef]  

45. Y. Kozawa, T. Hibi, A. Sato, et al., “Lateral resolution enhancement of laser scanning microscopy by a higher-order radially polarized mode beam,” Opt. Express 19(17), 15947–15954 (2011). [CrossRef]  

46. L. B. Lucy, “An iterative technique for the rectification of observed distributions,” Astron. J. 79(6), 745–754 (1974). [CrossRef]  

47. W. H. Richardson, “Bayesian-based iterative method of image restoration*,” J. Opt. Soc. Am. 62(1), 55–59 (1972). [CrossRef]  

48. E. A. Susaki, K. Tainaka, D. Perrin, et al., “Whole-Brain Imaging with Single-Cell Resolution Using Chemical Cocktails and Computational Analysis,” Cell 157(3), 726–739 (2014). [CrossRef]  

49. J.-Y. Tinevez, N. Perry, J. Schindelin, et al., “TrackMate: An open and extensible platform for single-particle tracking,” Methods 115, 80–90 (2017). [CrossRef]  

50. H. Fukuda, H. Kuramochi, Y. Shibuta, et al., “Analysis of Brownian motion trajectories of non-spherical nanoparticles using deep learning,” APL Mach. Learn. 1(4), 046104 (2023). [CrossRef]  

51. J. Gross, S. Sayle, A. R. Karow, et al., “Nanoparticle tracking analysis of particle size and concentration detection in suspensions of polymer and protein samples: Influence of experimental and data evaluation parameters,” Eur. J. Pharm. Biopharm. 104, 30–41 (2016). [CrossRef]  

52. B. Louis, R. Camacho, R. Bresolí-Obach, et al., “Fast-tracking of single emitters in large volumes with nanometer precision,” Opt. Express 28(19), 28656–28671 (2020). [CrossRef]  

53. J. G. George, K. Dholakia, and S. Bhattacharya, “Generation of Bessel-like beams with reduced sidelobes for enhanced light-sheet microscopy,” Opt. Continuum 2(7), 1649–1660 (2023). [CrossRef]  

54. S. Mori, “Side lobe suppression of a Bessel beam for high aspect ratio laser processing,” Precis. Eng. 39, 79–85 (2015). [CrossRef]  

55. G. Di Domenico, G. Ruocco, C. Colosi, et al., “Cancellation of Bessel beam side lobes for high-contrast light sheet microscopy,” Sci. Rep. 8(1), 17178 (2018). [CrossRef]  

56. J. C. Fiala, J. Spacek, and K. M. Harris, “Dendritic spine pathology: Cause or consequence of neurological disorders?” Brain Res. Rev. 39(1), 29–54 (2002). [CrossRef]  

57. M. Segal, “Dendritic spines and long-term plasticity,” Nat. Rev. Neurosci. 6(4), 277–284 (2005). [CrossRef]  

58. J. Nakai, M. Ohkura, and K. Imoto, “A high signal-to-noise Ca²⁺ probe composed of a single green fluorescent protein,” Nat. Biotechnol. 19(2), 137–141 (2001). [CrossRef]