Understanding how neural circuits process information requires rapid measurements of activity from identified neurons distributed in 3D space. Here we describe an acousto-optic lens two-photon microscope that performs high-speed focusing and line scanning within a volume spanning hundreds of micrometers. We demonstrate its random-access functionality by selectively imaging cerebellar interneurons sparsely distributed in 3D space and by simultaneously recording from the soma, proximal and distal dendrites of neocortical pyramidal cells in awake behaving mice.
Subscribe to Journal
Get full journal access for 1 year
only $20.17 per issue
All prices are NET prices.
VAT will be added later in the checkout.
Rent or Buy article
Get time limited or full article access on ReadCube.
All prices are NET prices.
Svoboda, K. & Yasuda, R. Neuron 50, 823–839 (2006).
Grewe, B.F., Voigt, F.F., van 't Hoff, M. & Helmchen, F. Biomed. Opt. Express 2, 2035–2046 (2011).
Kong, L. et al. Nat. Methods 12, 759–762 (2015).
Botcherby, E.J. et al. Proc. Natl. Acad. Sci. USA 109, 2919–2924 (2012).
Kaplan, A., Friedman, N. & Davidson, N. Opt. Lett. 26, 1078–1080 (2001).
Kirkby, P.A., Srinivas Nadella, K.M. & Silver, R.A. Opt. Express 18, 13721–13745 (2010).
Reddy, G.D. & Saggau, P. J. Biomed. Opt. 10, 064038 (2005).
Fernández-Alfonso, T. et al. J. Neurosci. Methods 222, 69–81 (2014).
Katona, G. et al. Nat. Methods 9, 201–208 (2012).
Froudarakis, E. et al. Nat. Neurosci. 17, 851–857 (2014).
Roorda, R.D., Hohl, T.M., Toledo-Crow, R. & Miesenböck, G. J. Neurophysiol. 92, 609–621 (2004).
Chen, X. et al. Nat. Protoc. 7, 1818–1829 (2012).
Evans, G.J., Kirkby, P.A., Naga Srinivas Nadella, K.M., Marin, B. & Angus Silver, R. Opt. Express 23, 23493–23510 (2015).
Ozden, I., Dombeck, D.A., Hoogland, T.M., Tank, D.W. & Wang, S.S. PLoS One 7, e42650 (2012).
Chen, T.W. et al. Nature 499, 295–300 (2013).
Yang, W. et al. Neuron 89, 269–284 (2016).
Cheng, A., Gonçalves, J.T., Golshani, P., Arisaka, K. & Portera-Cailliau, C. Nat. Methods 8, 139–142 (2011).
Akemann, W. et al. Opt. Express 23, 28191–28205 (2015).
Konstantinou, G. et al. Opt. Express 24, 6283–6299 (2016).
Emiliani, V., Cohen, A.E., Deisseroth, K. & Häusser, M. J. Neurosci. 35, 13917–13926 (2015).
Kaifosh, P., Zaremba, J.D., Danielson, N.B. & Losonczy, A. Front. Neuroinform. 8, 80 (2014).
Feng, L., Zhao, T. & Kim, J. eNeuro 2, 1 (2015).
This work was funded by the ERC (294667), the UCL impact PhD programme and the Wellcome Trust (095074). R.A.S. is in receipt of a Wellcome Trust Principal Research Fellowship in Basic Biomedical Science (095667; 203048). C.B. was funded by the Wellcome Trust PhD programme (097266). V.A.G. was funded by an EPSRC CASE studentship. We thank D. Farquharson and A. Hogben from the UCL Biosciences mechanical engineering workshop for the design and fabrication of mechanical components and G. Keller for sharing acquisition code. We acknowledge the GENIE Program and the Janelia Research Campus, Howard Hughes Medical Institute for making the GCaMP6 material available; the Mutant Mouse Resource Research Centre (MMRRC) for the C57BL/6 mGluR2-Cre-IRES-eGFP mouse strain. We thank D. Coyle for excellent technical assistance and D. DiGregorio, T. Fernandez-Alfonso, T. Margrie, A. Valera and T. Younts for comments on the manuscript.
Patents on the AOL technology have been filed (WO/2008/032061, WO/2011/131933).
Integrated supplementary information
Supplementary Figure 1 Random-access pointing and scanning performance of a 3D acousto-optic lens (AOL) microscope, its FPGA control system and the measured effect of dwell time on imaging volume.
a) Imaging duty cycle as function of voxel dwell time and number of voxels per line-scan assuming 24.5 μs AOL fill-time (where duty cycle = number of voxels x voxel dwell time / (number of voxels x voxel dwell time + AOL fill-time). b) Schematic diagram of custom FPGA logic for the AOL control system showing four channels that generate the radio frequency acoustic waves functions (F(t)) and their amplitudes (A(t)). c) Pollen grain imaged with AOL line-scanning using a 380 μm field of view at different dwell times. Top row shows imaging at -300 μm focal depth with dwell times from 50 ns to 800 ns. Middle row shows imaging at natural focal plane of the objective (0 μm) and bottom row shows imaging at a Z depth of +300 μm at different dwell times. For dwell times longer than 200 ns a reduction in brightness at the edges of the pollen grain can be observed when focusing 300 μm above or below the natural focal plane due to a fall-off in AOL diffraction efficiency. This effect is not evident for the 50 ns, 100 ns or 200 ns dwell times as predicted from the model (Fig. 1b). d) Estimated cycle rates for various imaging modes with AOL scanning and pointing. In the case of single plane and full volumes, 256 x 256 and 512 x 512 are considered. Sub-volume cycle rate was computed assuming 5 planes per sub volume with patch sizes of 10 x 10, 25 x 25, and 50 x 50 voxels. Arrows indicate a range of acquisition rates as the number (N) of planes, sub-volumes, patches and points are increased from 1.
Supplementary Figure 2 Diffraction efficiency of acousto-optic deflectors (AOD) and acousto-optic lens (AOL) and dependence of focusing range on dwell time for full frame and patch imaging.
a) The relationship between diffraction efficiency and AOD deflection angle (centred on the central drive frequency) for 1st generation and 2nd generation crystal design. Measurements were performed at 3W acoustic drive power and at 800 nm optical wavelength. b) Light transmission efficiency (normalised to natural focal plane) of second generation AOL measured with a photodiode at 920 nm wavelength over a +/- 125 μm axial range. Error bars indicate standard deviation. c) Theoretical relationship between AOL focusing range for full frame (250 μm x 250 μm FOV) 512 voxel line-scans and dwell time (blue line, left axis). Relationship between AOL focusing range and dwell time for 40 x 40 voxel patch (green line, right axis). d) Averaged patch images (40 x 35 voxels, 19.5 μm x 17.09 μm) of a fluorescent chroma slide at +/- 125 μm AOL focal planes imaged at 50 ns to 500 ns. Top right schematic shows patch location within the imaging volume, with colour code. e) Image of AOL microscope on a 1.5 m x 1.5 m optical table showing AOL assembly and 4f relay into the microscope.
Supplementary Figure 3 Properties of the two-photon point spread function (PSF) across the imaging volume.
a) Montage of images of 200 nm fluorescent beads at the centre (X, Y, Z: 0, 0, 0) of the imaging volume and at the corners of a 245 μm x 245 μm FOV at the natural focal plane. Note the radial elongation of the PSF at the corners. b) Left: short axis of PSF measured from 200 nm beads as a function of location across the XY FOV (Z=0). Right: long axis of PSF measured from 200 nm beads as a function of location across the XY FOV (Z=0). At the centre of the natural focal plane the XY resolution is ~0.65 μm and the axial resolution is ~ 3.75 μm. The radial increase in long axis is due to chromatic aberration. c) Size of PSF in XZ axis at various lateral positions across the natural focal plane. d) Left panel shows XY PSF and right panel shows XZ PSF as a function of distance from the natural focal plane, measured at the centre of each Z plane. Green region indicates the restricted AOL focussing range that we used, where the lateral resolution is <1 μm and the axial resolution of <8 μm. Increases in the PSF away from the natural focal plane are predominantly due to spherical aberration.
Supplementary Figure 4 Imaging spines and dendrites at various lateral and axial positions within a 250 μm x 250 μm x 250 μm imaging volume in an awake mouse.
a) Left: schematic diagram of imaging volume and colour-coded regions of interest (ROIs) at different focal planes. Fluorescent images of dendrites and spines expressing tdTomato at three different axial positions: -125 μm, 0 μm, and +125 μm, with outlines corresponding to focal plane colour. b) Left: schematic diagram of XY field of view at Z=0 with positions of ROIs colour-coded. Fluorescent images of dendrites and spines expressing tdTomato at four different regions across the natural focal plane as indicated by the colour of the border. Despite some reduction in optical resolution towards the edges of the imaging volume, spines can still be resolved.
a) Schematic diagram showing basket, stellate, Golgi, granule and Purkinje cells, (BC, SC, GoC, GC, and PC, respectively) together with mossy fibres (MF) and their location within the three-layered structure of the cerebellar cortex. b) AOL Z-stack of a mouse expressing GCaMP6f in the cerebellum, showing the position of the four 256 x 256 voxel planes that were imaged.
Supplementary Figure 6 Comparison of multi-plane imaging and user selected random access patch imaging for monitoring sparsely distributed cells.
a) Schematic diagram of 250 μm x 250 μm x 250 μm imaging volume with 14 neuronal cell bodies randomly distributed at low density. Left: cell bodies intersected (green) by three imaging planes (orange shading), cell bodies not intersected (red) by imaging planes. Right: user-selected patches (orange shading) aligned to intersect each cell body (green), giving probability of intersection is 100%. b) Probability of intersecting randomly positioned neuronal compartments as function of their sparsity (defined as Z dimension of the imaging volume/Z dimension of structure of interest) with multi-plane imaging and patch/point imaging. Arrows indicate the sparsity of two common structures of interest assuming 14 fall within the imaging volume. c) Estimated cycle rate for plane (magenta line, 50 ns dwell time with 512 x 512 voxels), patch (yellow line, 50 ns dwell time with 51 x 50 voxels), and point imaging (blue dashed line, 4 μs dwell time) as function of number of planes, patches and points imaged. Examples of multi-plane imaging (squares) and patch imaging (circles). Liquid crystal spatial light modulator multi-plane (green square, Yang et al. Ref 16). Space time multiplexed multi-plane imaging (light blue square, Cheng et al. Ref 17). Multi-plane imaging with large field of view two photon mesoscope (brown square, Sofroniew et al. Ref 23). Multi-plane imaging with expanded field of view two photon microscope (dark blue square, Stirman et al. Ref 24). Multi-plane imaging with AOL (orange square, Fig. 1). AOL patch imaging speed is shown in circles, Fig. 2 (red), Fig. 3a-c (grey), Suppl. Fig. 7 (green) and Suppl. Fig. 8 (purple, note that one of the four circles is behind the red).
Supplementary Figure 7 Random-access patch imaging across layers in the cerebellum of an awake behaving mouse with post hoc movement correction.
a) AOL Z-stack of molecular and granule cell layers in the cerebellum of a mouse expressing the Ca2+ indicator GCaMP6f. b) Location of 13 X-Y patches (171 x 46 voxels, 73.5 μm x 19.8 μm) within the imaging volume in panel a, distributed between 62 μm to 119 μm below the pia (see Supplementary Video 4 for patch fluorescence during part of recording). c) Averaged images of cellular structures scanned near-simultaneously at 50 ns/voxel in the 13 patches after post hoc correction for brain movement. Traces to the right show ΔF/F responses extracted from each patch with a 50.6 Hz sampling rate for a complete 13 patch cycle. Grey trace at the bottom shows the speed of locomotion of the animal on the cylindrical wheel. Grey shading indicates periods of locomotion. Vertical bar indicates normalised change in fluorescence (∆F/F) together with speed of locomotion (cm/s) and horizontal scale bar indicates time (s).
Supplementary Figure 8 Random-access patch imaging of layer 2/3 neurons expressing only GCaMP6f in primary visual cortex of an awake behaving mouse using different dwell times and post hoc movement correction.
a) Image from a Z-stack showing cells in layer 2/3 visual cortex sparsely expressing GCaMP6f. b) Location of 14 selected X-Y patches (46 x 40 voxels, 22 μm x 19.4 μm) within the imaging volume, distributed between 76 μm to 235 μm below the pia. c) Averaged movement corrected images of cellular structures scanned near-simultaneously in the 14 patches. Traces to the right show ΔF/F responses extracted from each patch when imaged at 50, 100, 200 and 300 ns/voxel dwell time, which gave sampling rates of 67 Hz, 61 Hz, 53 Hz, 47 Hz for the complete 14 patch cycle. Numbers within brackets at bottom indicate measurement sequence. Recordings were filtered using a 5-point running average and arranged in order of data acquisition. Depth and patch numbers are indicated on the right. Dashed boxes indicate instances where two patches monitored the same soma. Grey trace at the bottom shows the speed of locomotion of the animal on the cylindrical wheel. Grey shading indicates periods of locomotion. Vertical bar indicates normalised change in fluorescence (∆F/F) together with speed of locomotion (cm/s) and horizontal scale bar indicates time (s).
a) Patch images following post hoc movement correction from Figure 2 together with the regions (masks) from which the fluorescence intensity data was extracted from the image time series (area within green contour lines). b) as for (a) but for data in Supplementary Fig. 7. c) as for (a) but for Supplementary Fig. 8.
Supplementary Figures 1–9, Supplementary Table 1 and Supplementary Notes 1–3. (PDF 1755 kb)
250x250x250 μm Z-stack of tdTomato expressing L2/3 pyramidal cells in visual cortex of awake mouse (115 planes, 1024x1024 voxels, 50 ns dwell time, with 5 averages per plane). (AVI 24080 kb)
Z-stack of 39 planes with frame size of 512x512 voxels and 100 ns dwell time. Acquisition speed 25.8 Hz per plane. (AVI 9988 kb)
High-speed simultaneous 4 plane two-photon imaging with an AOL in cerebellar cortex of an awake mouse at 19.5 Hz per frame cycle. Locomotion data is updated at frame rate. (AVI 35364 kb)
Random-access patch imaging of 13 selected Patches across cortical layers in the cerebellum of awake behaving mice imaged at 50.6 Hz per patch cycle. (AVI 24753 kb)
Random-access patch imaging of patches across sparsely labelled pyramidal cells in layer 2/3 of visual cortex
Random-access patch imaging of 14 selected patches across sparsely labelled pyramidal cells in layer 2/3 of visual cortex of awake behaving mice imaged at 52.8 Hz per patch cycle. (AVI 34887 kb)
About this article
Cite this article
Nadella, K., Roš, H., Baragli, C. et al. Random-access scanning microscopy for 3D imaging in awake behaving animals. Nat Methods 13, 1001–1004 (2016). https://doi.org/10.1038/nmeth.4033