Fast, cell-resolution, contiguous-wide two-photon imaging to reveal functional network architectures across multi-modal cortical areas

•High optical invariant was achieved with resonant-based two-photon microscopy•A 3 mm square of the mouse cortex can be scanned at 7.5 Hz for calcium imaging•Activity over 16,000 neurons across multiple areas observed•Network analysis revealed small-world properties: cost-effective cortical dynamics Fast and wide field-of-view imaging single-cell resolution, high signal-to-noise ratio, no aberrations have potential to inspire new avenues investigations in biology. However, such is challenging because inevitable tradeoffs among these parameters. Here, we overcome by combining a resonant scanning system, large objective low magnification numerical aperture, highly sensitive large-aperture photodetectors. The result practically aberration-free, fast-scanning microscopy (FASHIO-2PM) that enables from network composed ∼16,000 9 mm2 contiguous image plane, including more than 10 sensory-motor higher-order cerebral awake mice. Network based on activities brain exhibits rather scale-free behavior. FASHIO-2PM expected enable studies biological simultaneously monitoring macroscopic their compositional elements. If phenomenon certain system (e.g., biology, social science, or physics) observable, matter how complex it appears be, its mechanisms predicted subdividing into individual elements but not vice versa. Given only details elements, appearance may predictable (Anderson, 1972Anderson P.W. More Is Different.Science. 1972; 177: 393-396Crossref PubMed Scopus (1900) Google Scholar). development method monitor will driving force discoveries open up horizons any field benefits observations. In neuroscience, elementary information processing required cognitive processes thought executed within single areas, emergent properties require activity involving functionally diverse via long-/mid-range connections between (Manita et al., 2015Manita S. Suzuki T. Homma C. Matsumoto Odagawa M. Yamada K. Ota Matsubara Inutsuka A. Sato al.A Top-Down Cortical Circuit Accurate Sensory Perception.Neuron. 2015; 86: 1304-1316Abstract Full Text PDF (181) Scholar; Miyamoto 2016Miyamoto D. Hirai Fung C.C.A. Boehringer R. Adaikkan Matsuki N. al.Top-down input during NREM sleep consolidates perceptual memory.Science. 2016; 352: 1315-1318Crossref (72) Schneider 2014Schneider D.M. Nelson Mooney A synaptic circuit basis corollary discharge auditory cortex.Nature. 2014; 513: 189-194Crossref (269) Zhang 2014Zhang Xu Kamigaki Do J.P.H. Chang W.C. Jenvay Miyamichi Luo L. Dan Y. Long-range local circuits top-down modulation visual processing.Science. 345: 660-665Crossref (381) Thus, number (i.e., neurons) multi-modal limited unimodal, necessary comprehensive understanding functions, which one important challenges this field. Most gains are currently being made electrical high-density approaches (Jun 2017Jun J.J. Steinmetz N.A. Siegle J.H. Denman D.J. Bauza Barbarits B. Lee A.K. Anastassiou C.A. Andrei Aydın Ç. al.Fully integrated silicon probes recording neural activity.Nature. 2017; 551: 232-236Crossref (673) Scholar) make difficult identify select underlying cells geometrical structures. Optical issue quickly face intrinsic limitations. To optically neurons, needs microscope has view (FOV) higher spatial resolution. parameters inversely related. counteract tradeoff, an would (Mag) aperture (NA). Recent efforts achieve wide-FOV (2P) excitation microscopes take strategy increase spatially separate small FOVs (each FOV approximately 0.25 mm2) employing (Chen 2016Chen J.L. Voigt F.F. Javadzadeh Krueppel Helmchen F. population anatomically defined neocortical networks.eLife. 5: e14679Crossref (66) Rumyantsev 2020Rumyantsev O.I. Lecoq J.A. Hernandez O. Savall J. Chrapkiewicz Li Zeng H. Ganguli Schnitzer M.J. Fundamental bounds fidelity sensory coding.Nature. 2020; 580: 100-105Crossref (39) Sofroniew 2016Sofroniew N.J. Flickinger King Svoboda mesoscope subcellular resolution vivo imaging.eLife. e14472Crossref (255) Stirman 2016Stirman J.N. Smith I.T. Kudenov M.W. S.L. Wide field-of-view, multi-region, neuronal mammalian brain.Nat. Biotechnol. 34: 857-862Crossref (165) One known as Trepan2P (Twin Region, Panoramic 2-photon) (Stirman Scholar), also offers mode without stitching) (product NA; see results further details). This mode, however, significantly decreases sampling rate (3.5 0.1 Hz), resulting loss temporal neuroscience. Another does involve lenses instead off-the-shelf components increased (Bumstead 2018Bumstead J.R. Park Rosen I.A. Kraft A.W. Wright Reisman M.D. Côté D.C. Culver J.P. Designing using analysis.Neurophotonics. 2018; 025001Crossref (26) Tsai 2015Tsai P.S. Mateo Field Schaffer C.B. Anderson M.E. Kleinfeld Ultra-large microscopy.Opt. Express. 1609: 1825-1829Google demonstrated ultrawide-FOV 2P imaging. Bumstead al. 7-mm-diameter lateral direction (1.7 μm). axial 28 μm substantially slow objective’s performance commercial objective, inadequate imaging, galvo-galvo used, respectively. (6 diameter) one-photon confocal Mesolens (McConnell 2016McConnell G. Trägårdh Amor Dempster Reid E. Amos W.B. novel embryos tissue volumes sub-cellular throughout.eLife. e18659Crossref (62) owing combination lens achieves aberrations. limit depths tissues light-scattering designed use visible lasers (Hamel 2015Hamel E.J.O. Grewe B.F. Parker J.G. Cellular level behaving mammals: engineering approach.Neuron. 140-159Abstract (100) best our knowledge, optimized permit fast yet fully examined. Nonetheless, previous clarified maximizing high-performance straightforward approach toward realizing microscopes. study, were able maximize speed large-angled newly (0.8 NA, 56 pupil diameter, Strehl SR ∼0.99 FOV), pre-chirper, GaAsP photomultiplier current output (PMT, 14 50 μA). proposed new, hitherto untested fast, wide-, contiguous-FOV ratio (SNR), entire FOV. Our allowed us × animal behaviors. Using microscope, performed functional brain. Herein, describe fast-scanning, (0.6 1.2 collection) (see Table S1 acquisition modes). concurrently following several capabilities: (3 2,048 pixels, 36 larger conventional ∼0.5 0.5 microscope; Peron 2015Peron S.P. Freeman Iyer V. Guo Resolution Map Barrel Cortex Activity Tactile Behavior.Neuron. 783-799Abstract (153) NA (optical resolution: x-y, 1.61 μm; z, 7.07 below), frame (7.5 Hz) activity. aberration-free vignetting-less (Beiser, 1986Beiser Imaging laser scanners.Opt. News. 1986; 12: 10Crossref Scholar, Beiser, 1995Beiser architecture systems.Appl. Opt. 1995; 7307-7317Crossref (48) calculated height angle chief marginal rays, conserved constant value each component. light Ie all kinds ranging scanner plane 2P-microscope (LS2PM) non-descanned detector given follows:Ie=rmsinθm=rpsinθp=n rfsinθe,(Equation 1) where rm θm beam radius scan mirror scanner, rp θp incident collimated respectively; n refractive index immersion specimen; rf radius, θe cone (Figure 1A). avoid confusion, define rear near tube aperture. Note last equation Equation 1 corresponds product radius. Importantly, lowest system’s bottleneck component). increasing component contiguous-wide NA. used resonant-galvo system. sinθm 1, chose 7.2 5.0 elliptical clear 26-degree maximum angle, 8 kHz frequency (CRS8KHz, Cambridge Technology). These settings 2.5 times obtained 10-degree 12 (CRS12KHz) (Sofroniew 2-fold faster 9.25 mm, 26 degrees, 4 (CRS4KHz) We sought develop capable supporting available CRS8KHz. aimed collect much fluorescent emission possible specimen. collected Ic LS2PM determined independently follows:Ic=n rfsinθc=rssinθs,(Equation 2) θc rs θs sensor exiting collection optics, respectively 1B). twice leads > improve SNR Because and/or should maximally utilize Ic, developed (rs) PMT high-output high-sensitivity below). restricted utilizable area mirror. therefore selected high-precision flatness <0.015λ root mean (RMS) 633 nm surface. Next, expanded 2.75 (1/e2) expander projected onto (Note tradeoff upon wavefront error. outer periphery is, general, lower center, usage reflection considerably increases amount error, i.e., deviation ideal wavefront, leading decrease efficiency resolution.) finally ∼15 degrees (Nikon A1RMP) 25.3 kHz. 0.60, those (given there aberration FOV, Nikon A1RMP CFI75 LWD 16X W [0.8 ∼0.79 0.79 FOV] ∼0.318; Olympus FVMPE-RS XLPLN25XWMP2 [1.05 ∼0.267, other microscopes). ensure equal greater 0.6, fit commercial-standard Leica, Nikon, Olympus, Zeiss) 2A, S2 specifications). 0.4 along z axis (Zipfel 2003Zipfel W.R. Williams R.M. Webb W.W. Nonlinear magic: multiphoton biosciences.Nat. 2003; 21: 1369-1377Crossref (2994) = sinθe, mm2, sufficiently had 13 elemental groups diameter (dry immersion; 4.5 working distance, 35 focal length). 5 groups) (10 satisfy (0.6) Figure S1, entrance through relay lenses, For collection, fluorescence sample possible, 0.8. Finally, 64 170 84 (without flange), weight 4.2 kg. 1.2. various below) lens, applied two strategies design manufacturing. First, different realize efficiency, difference effective diameters (Figures 1A produces thickness, light-projected size lens. reasons why suppressing when designed. passes relatively produce 0.8 1B; S1). Therefore, considered corrections mostly center opposed parts passes. challenge large-lens, almost perfect correction NA0.4 show supplementary figures S2–S8). Second, note manufacturing mounting cause eccentric errors unaligned axis), reduce originally simulation results. address issue, tube, so could aligned after mounting. adopted four-screws alignment technique groups. technique, close design. commonly semiconductor stepper manufacturing). case, planes 10, 4, 2A). features contribute first attempt both methods wide-field microscope. invariant, high-sensitive photodetector, (14 S9 specifications), commercially extended red multialkali (ERMA) (18 2B). ∼10× PMTs (Φ H10770A-40, Hamamatsu photonics K.K.), QE ∼2.6× ERMA (45% versus 17% 550 nm). importantly, reason compared (2 contributes upper signal dynamic range low- high-fluorescence signals same time bandwidth amplifier photodetector proportional pixels frames per second. high-speed wider amplifier). feedback register resistance noise gain (reciprocal attenuation input) increasement high-frequency region. obtain high-voltage resistance, requires high-input photodetector) maintain SNR. provides current. three (large high-sensitive, current) essential achieving pre-chirper avoids pulse stretching degradation pulses (Fork 1984Fork R.L. Martinez O.E. Gordon Negative dispersion pairs prisms.Opt. Lett. 1984; 9: 150-152Crossref (737) group delay (GDD positive-GDD) components. positive GDD significant includes thicker many lenses. chirp reduces peak intensity light. estimated ∼10.7 103 fs2 (at 920 nm), whereas ∼4.0 fs2. total 15.0 compensate GDD, generate negative dispersion. stands out respect compensation built-in MaiTai DeepSee source (∼–9.3 GDD). decided expand adding another pre-chirper. four-prisms compressor additional pre-chirper) featuring transmission configuration long-term stability (because absence reflective components) efficiency. prisms reducing dimensions device. Nevertheless, device here long distances (550 separation most), 2C). ∼–6.0 wavelength 500 separation) pre-chirpers Indeed, improved S10), critically probability shorter pixel (dwell time, text 2D 2E), infrared (IR) generator, build-in introduced unit consisting four IR then led dichroic illuminates mirrors. power ∼25%. while maintaining frequency, thus permitting Conversely, dwell (the dwells position: FASHIO-2PM, ∼18–36 ns; A1R MP, ∼70–140 ns), quantum PMT. encircled energy function (EEF), representing concentration showed 80% contained 1.1 3A 3B ; S11 system). value, even edge equivalent diffraction limit, indicating axes less 1%, thereby preserving uniform superior SR, quality point spread (PSF), objective. full widths half (FWHMs) bead images below surface cover glass 3C; STAR Methods). As shown EEF simulation, FWHMs depth 100 same, independent measurement position 3D–3H). ± 0.06 (mean SD). 0.46 μm. measured indicated neuron x, y, axes, well S12). demonstrate scale microscopic expressed genetically encoded indicators (GECIs) G-CaMP7.09 (Shiba 2016Shiba Gomibuchi Seto Wada Ichimura Tanaka Ogasawara Okada Shiba Sakamoto al.Allogeneic transplantation iPS cell-derived cardiomyocytes regenerates primate hearts.Nature. 538: 388-391Crossref (383) GCaMP6f 2013Chen T.-W. Wardill T.J. Sun Pulver S.R. Renninger Baohan Schreiter E.R. Kerr R.A. Orger M.B. Jayaraman al.Ultrasensitive proteins 2013; 499: 295-300Crossref (3133) widely distributed wide- Ca2+ Specifically, injected adeno-associated virus (AAV)-conjugated indicator (AAV-DJ-Syn-G-CaMP7.09 AAV9-Syn-GCaMP6f) ventricle wild-type postnatal day 0–8 (P0–8) 4A; Methods), opened cranial window (∼4.5 diameter), monitored P28. estimate labeled injection, stained slices, primary somatosensory cortices forelimb (S1FL) hindlimb (S1HL) cortices, motor (M1), posterior parietal (PPC), barrel antibodies against NeuN, marker, GAD67, inhibitory marker. found 85.1%–90.2% layer 2/3 (L2/3) excitatory non-GABAergic injection 4A–4C; Methods cell estimation procedure). ⨯ (2,048 pixels), area, frames/s (G-CaMP7.09; Video raw ΔF/F data representations). L2/3 ∼60–80 mW (<180 PMT) front 250 mW, initiate heating damage phototoxicity conventional-FOV (Podgorski Ranganathan, 2016Podgorski K