Fig 1.
A) Key components of the microscope head. Shaded colors illustrate beam deflection angles. B) Cartoon of the beam conditioning components. First the beam is split by 50%, the intensity adjusted with an electro-optic modulator (EOM) and the beam width with a telescope before entering the microscope head. C) 3D drawing of the full microscope, highlighting beam conditioning subsystem (blue) and the microscope head (yellow). For scale, the spacing between two holes on the optical table is 1 inch or 2.5 cm. D) Photograph of the microscope head with key elements labeled. This section is visible in panel A.
Fig 2.
Optical design of the excitation pathway.
A) Zemax design with the scanning mirror, tube and scan lens groups. Colors indicate light rays produced by different mirror angles of the scanning mirror. Notice how the rays converge in the image plane. B) CAD design built around the optical design in A, see also Fig 1C. C) Cuts through the scan and tube lens assemblies, which are fabricated from off-the-shelf lenses housed in SM2 lens tubes. The specific lenses are stated in the text. D) Simulation of the optical system’s two-photon point spread function (PSF-2p) for 0 deg deflection angle, and two example radial sections. E) PSF-2p in the xz plane across the full range of deflection angles. Shown below is Petzval field curvature (blue), and radial (purple) and axial (yellow) full-width-at-half-maximum of the PSF-2p. Error bars are resolution limits of the Huygens PSF estimates in sequential mode.
Fig 3.
Optical design of collection system.
A) Zemax design of the collection optics. Colors indicate different emission angles. These rays were computed for the green emission light path. Small letters denote filters: a is a FF665-Di02–40x55; b: a near-infrared blocking filter (FF01–680/SP-50) (not shown in the Zemax simulation); c: FF562-Di03–40x52 and d: FF01–525/45-32 (BrightLine; AVR Optics). Small numbers denote lenses. 1 is a LA1384-A; 2: LB1607-A and 3: ACL25416U-A (all Thorlabs). B) CAD design built around the Zemax model in A, with several covers removed to expose the interior. The light paths from the laser and fluorescent emission are indicated in transparent colors. Notice how the dichroic mirror splits fluorescence and excitation light, and directs the former to the photomultiplier tubes (PMTs). The mirrors/filters and lenses are numbered/lettered as in A). Two more filters were added relative to the simulation. b blocks residual IR excitation light λ > 680 nm (FF01–680/SP-50), and e is a FF01–600/52-32 filter for the red channel. C) Shown is the angle collection graph of the collection optics, which shows the fraction of collected light as function of the angle it emerges from the objective back aperture. This was simulated in Zemax for the design in A. The collection optics were optimized to collect essentially all light over a ≈ ±8 deg emission angle.
Fig 4.
Custom electronic control circuit and box.
A) Control circuit to set and read the gain of the two H16201P-40 photomultiplier tube modules used for red and green channel. The two modules are connected with two 4-wire connectors, and their current gain is displayed on a small digital panel meter. B) Wiring of the panel meter for single-ended configuration. C) The circuit in a prototype-box, and mounted to the side of a rack. The two gains for green (top) and red (bottom) channel are visible. A gain of around ≈ 0.8 is typical for Calcium imaging to maximize the faint signal. A gain of ≈ 0.6 is appropriate for very bright samples.
Fig 5.
A) Measurement of the local magnification with a 100 μm calibration target in a thin Fluorescein film below a coverslip. Shown in green is a 700 × 700 μm2 square as reference. B) Histogram of the magnifications in A, measured across all squares. C) Two-photon intensity measurements from a uniform bath of Fluorescein. Shown are pixel values (maximum is 32768). D) Horizontal and vertical profiles through the data in C. The green box is indicated by the shaded green area. Range is ±1 standard deviation. E) Histogram of the intensities from within the green box in C. F) Volumetric measurement of a 0.2 μm bead sample in 1% Agarose. G) Average of N = 38 beads. H) Horizontal and vertical profiles through data in G and a Gaussian fit reveal a radial resolution of 760 ± 30 nm. I) Estimate of the axial resolution shows 5.4 ± 0.9 μm.
Fig 6.
Example imaging data from simple samples.
A) Autofluorescence of a part of a dandilion flower (Taraxacum officinale) in vivo. Notice the pollen grains embedded in the plant material. Optical sectioning allows to focus below the plant surface to the grains. B) Zoom and volumetric image of a small region (white box) of the sample in A. Optical sectioning shows the surface of the plant and the slice where the pollen grains come into focus. Notice the complex 3D structure of the pollen grains, and their embedding below a layer of red-fluorescent plant material. C) Histological sample showing a stain for glial fibrillary acidic protein in rat astrocytes. D) Image of a cotton napkin (Kimwipe) with depth coded as color. Zooming into a small region (white box) shows how strands of cotton are woven together, and how the pattern changes with depth. Notice how image quality deteriorates in this strongly scattering sample.
Fig 7.
Example calcium imaging in transgenic Zebrafish (Danio rerio).
A) 3D rendering of a 300 μm deep volumetric image of a five day old Zebrafish larva with Fiji [36]. B) Dorsal slice through the same animal and two zoom views down to cellular scale. C) Color-coded ≈ 10 min of time-averaged volumetric imaging data of three 20 μm-spaced planes using the piezoelectric collar attached to the objective. The layers are colored in red/green/blue. No motion correction was applied. Notice the blurry averages, caused by motion of the fish. D) Running Suite2p on the data in C produces numerous regions of interest (ROIs) with complex Calcium dynamics. Left: Suite2p applies motion correction, resulting in a much crisper average picture. Right: Time series of a few ROIs over 10 min.
Fig 8.
Example calcium imaging in adult Drosophila melanogaster.
A) Imaging from the brain of a head-fixed adult fly [38]. A dissection is performed to remove the cuticle from the back of the fly’s head and provide optical access to the brain, and the imaged brain volume is outlined. The transgenic flies express pan-neuronal GCaMP6s and myr-tdTomato, driven by the nsyb promoter. B) Average projections over ten 5 μm-spaced planes of both imaging channels, acquired at a temporal resolution of 1 volume/s. XY boundaries of exemplar ROIs are indicated. C) The channels are motion corrected and spatially clustered ROIs are extracted. Examples of spontaneous calcium activity extracted from ROIs in the brain volume.
Fig 9.
Examples for teaching principles of two-photon microscopy.
A) Transmitted laser power as function of the control voltage applied to the Pockels crystal for two different crystal orientations. The red orientation is far from the optimum and produces a small dynamic range for intensity control. The blue orientation is close to the optimal. The thin purple line is a best fit (see text). B) Fluorescence in Fluorescein as a function of applied dispersion compensation. A fit allows to estimate the pulse width. Note also the symmetric shoulders not captured by the fit, suggesting a non-Gaussian pulse shape.