High-contrast multifocus microscopy with a single camera and z-splitter prism

Abstract

Optical microscopy has been an indispensable tool for studying complex biological systems, but is often hampered by problems of speed and complexity when performing 3D volumetric imaging. Here, we present a multifocus imaging strategy based on the use of a simple z-splitter prism that can be assembled from off-the-shelf components. Our technique enables a widefield image stack to be distributed onto a single camera and recorded simultaneously. We exploit the volumetric nature of our image acquisition by further introducing a novel extended-volume 3D deconvolution strategy to suppress far-out-of-focus fluorescence background to significantly improve the contrast of our recorded images, conferring to our system a capacity for quasi-optical sectioning. By swapping in different z-splitter configurations, we can prioritize high speed or large 3D field-of-view imaging depending on the application of interest. Moreover, our system can be readily applied to a variety of imaging modalities in addition to fluorescence, such as phase-contrast and darkfield imaging. Because of its simplicity, versatility, and performance, we believe our system will be a useful tool for general biological or biomedical imaging applications.

Fig. 1.

(a) Schematic of our multifocus imaging system. (b)–(d) Construction of 3-, 6-, and 9-plane z-splitter prisms. (e)–(g) Fitted axial intensity profile of a 1 μm fluorescent bead measured with 3-, 6-, and 9-plane prisms. (h)–(j) Theoretical and measured axial focal positions associated with 3-, 6-, and 9-plane prisms. (k) Simultaneous 9-depth transmission imaging of a USAF resolution target using a 9-plane prism. Δz = 22.1 μm.

Fig. 2.

(a) Schematic of EV-3D deconvolution algorithm. (b) Color-coded EDOF image of the raw 9-plane image stack. (c) Color-coded EDOF image of the deconvolved stack using RL-3D algorithm. (d) EV-3D estimated background far-out-of-focus fluorescence in (b) generated from volume extension V_EV − V_I. (e) Color-coded EDOF image of the deconvolved stack using EV-3D algorithm. Scale bars are 100 μm. (f) SBR distribution as a function of axial extension w. Box plots: box, 25th to 75th percentiles; whiskers, 1.5× interquartile range from the 25th and 75th percentiles; middle horizontal line, median; “×” symbol, average.

Fig. 3.

High-speed multifocus in vivo imaging of GCaMP7-labeled mouse brain over a large FOV. (a) (max–min) projections of three different focal planes over an 8 min recording at 50 Hz. The total 3D FOV is 1.2 × 1.2 × 0.22 mm³. From left to right, 200, 345, and 186 neurons are identified in each plane. Scale bar, 100 μm; Δz = 110 μm. (b) Calcium traces of subset of neurons during 60–120 s recording period. (c)–(e) Comparison of SNR for neurons 267–271 during 120–180 s recording period. (f) Distribution of ΔF/F conditioned on ΔF/F > 3σ_n in cases of no deconvolution, RL-3D deconvolution, and EV-3D deconvolution. (g) SNR distribution conditioned on ΔF/F > 3σ_n in cases of no deconvolution, RL-3D deconvolution, and EV-3D deconvolution. Box plots the same as Fig. 2(f). (h) Pairwise cross-correlation as a function of inter-neuron distance in cases of no deconvolution, RL-3D deconvolution, and EV-3D deconvolution. Solid line, average cross-correlation; shaded area, ±1 std.

Fig. 4.

Video-rate multifocus tracking of freely moving C. elegans. (a) Color-coded EDOF image of one frame of the C. elegans video before deconvolution. (b) Color-coded EDOF image of the same frame after EV-3D deconvolution. (c), (d) Individual 9-plane images over the green rectangular regions in (a), (b), respectively. (e) One frame from the video used for C. elegans tracking. (f) All 27 traces of tracked head ganglia positions from a 104 s video. (g) Position of the head ganglia of a single worm [indicated by red arrow in (e)] from 12 s to 86 s. Line color represents depth. (h)–(j) Extracted skeletons of a single worm at times t = 30.0 s, 60.0 s, and 66.7 s, respectively. Red dots represent tracked head ganglia positions. Positions in (h)–(j) correspond to the rightmost, leftmost, and middle red dotted positions in (g). All scale bars are 50 μm. For visualization, the gamma values of all images are set to 0.5.

Fig. 5.

Multifocus phase-contrast imaging of living rotifers. (a) Large FOV (1.1 × 1.1 × 0.7 mm³) imaging of multiple rotifers in their natural state. Δz = 88.6 μm. Scale bar, 200 μm. (b) Multifocus image of a single rotifer at high resolution. Δz = 5.5 μm. Scale bar, 50 μm. (c), (d) High speed imaging (100 Hz) of beating cilia of a rotifer. (c) and (d) show two frames at times t = 0.6 s and t = 24.7 s, respectively. Δz = 5.5 μm. Scale bar, 50 μm. (e) Expanded view of a single rotifer within the blue square in (a). (f)–(h) Expanded view of stomach, mastax, and corona regions [yellows squares at depths +2 Δz, −2 Δz, and −4 Δz in (b)] of a rotifer at different depths. (i), (j) Expanded view of cilia from the red square regions in (c), (d), respectively. The focal position of the cilia moved from +3 Δz at t = 0.6 s to −2 Δz at t = 27.4 s. a.u., arbitrary unit.

Fig. 6.

Multifocus darkfield imaging of entire organisms. (a)–(c) All-in-focus images of radiolaria, volvox, and dictydium. Scale bar, 100 μm. (d), (e) 9-plane images of a living Daphnia magna at times t = 6 s and t = 58 s, respectively. Δz = 88.6 μm. Scale bar, 200 μm. (f)–(h) Expanded views of boxes in (d) (e), where distinct features can be seen across depths. a.u., arbitrary unit.