nature REVIEW ARTICLES photonics PUBLISHED ONLINE:31 JANUARY 2013DOI:10.1038/NPHOTON.2013.361 Advances in multiphoton microscopy technology Erich E.Hoover12*and Jeff A.Squier1.2* Multiphoton microscopy has enabled unprecedented dynamic exploration in living organisms.A significant challenge in bio- logical research is the dynamic imaging of features deep within living organisms,which permits the real-time analysis of cellular structure and function.To make progress in our understanding of biological machinery,optical microscopes must be capable of rapid,targeted access deep within samples at high resolution.In this Review,we discuss the basic architecture of a multiphoton microscope capable of such analysis and summarize the state-of-the-art technologies for the quantitative imag- ing of biological phenomena. ew windows in biological exploration are being opened from the entire imaging plane simultaneously.This important dis- through the continuing development of novel optical mul- tinction allows multiphoton microscopes to perform efficient,deep tiphoton microscopy (MPM)techniques.In this imaging explorations within scattering tissues6 by including data from paradigm,near-infrared (near-IR)femtosecond lasers are used to multiply scattered signal photons,which might otherwise introduce excite optical processes that can be accessed only through the appli- a background fog?s.Whole-field detection comes at the cost of con- cation of two (or more)photons.Two-photon excitation fluores- fining imaging to within a few tens of micrometres of the surface cence(TPEF)-a process driven by the simultaneous absorption of of a scattering sample;this trade-off is an important consideration two near-infrared photons by a single fluorophore-is one example when exploring biological behaviour in scattering media. of such a technique'.The probability of triggering a multiphoton The multiphoton microscope design discussed here allows MPM process,such as TPEF,is extremely unlikely to occur.Interactions to address one of the toughest challenges since the inception of are therefore restricted to the focal plane of the objective,where the optical microscopy:achieving image contrast at the cellular level in beam intensity is maximized,which provides the optical sectioning thick,scattering specimens.By their very nature,cells are quite thin necessary for the non-perturbative analysis of living systems. (a few micrometres);such small path lengths provide little absorp- Since its inception,MPM has permitted a variety of unique tion,path length differences or scattering-all of which can pro- explorations into highly scattering materials.These studies have vide a detailed view of the intricacies of the biological machinery. examined membrane potentials on the single-molecule scale2,the The use of a femtosecond laser as the light source for the micro- non-invasive observation of embryo development3 and the simul- scope has enabled the generation of an entirely new class of con- taneous multiplane imaging of calcium transportation in transgenic trast mechanisms within such specimens.Remarkably,these lasers mice.The ability to perform such explorations is a direct result of provide intensities of the order of tens of gigawatts per square cen- the inherent optical sectioning of multiphoton microscopes's and timetre with modest focusing (for example,0.65 NA)and relatively the reduction in photobleaching outside of the imaging planes-7. low average powers(milliwatts).At such intensities,the electric field Multiphoton microscopes also benefit from the ability to utilize at the focus creates a separation of positive and negative charges, longer excitation wavelengths(700 nm and greater)than confocal thus momentarily polarizing the material.In fact,the induced time- techniques,thus making them less biologically harmfuls7 and more varying polarization is significantly overdriven,resulting in the penetrating in scattering tissue2.In addition,multiphoton micro- generation of new optical signals-distinctive from the excitation scopes can frequently take advantage of the endogenous contrast beam-that can be used to visualize structure and function in an mechanisms inherent to many samples,thus permitting the explo- unprecedented fashion.Substantive reviews of the broad array of ration of untreated specimens325. nonlinear processes now used in imaging can be found in studies by As shown in Fig.1,a typical multiphoton microscope is com- Yue et al.7 and Chung et al.is. posed of a femtosecond laser,a scanning system,a low-magnifi- Today,this nonlinear polarization can be generated with pulses as cation high-numerical-aperture (NA)microscope objective,a short as 10 fs,which represents a pulse consisting of only five or six wavelength-sensitive dichroic and a single-element detector.The optical cycles.This result is remarkable given the complexity of scanning system is an intermediate optical system that is used to the optical system that is necessary to deliver such broadband light raster the excitation beam in a two-dimensional(2D)field at the to the specimen.Combining scan optics with well-corrected high- full NA of the objective.The objective is generally used both for NA optics poses significant challenges for producing a focal spot that excitation and to collect the signal photons.These signal photons is both diffraction-limited in space and transform-limited in time. are separated from the excitation beam on the return path using Whether a pulse of 100 fs or 10 fs is used as the excitation source,it is a wavelength-sensitive dichroic.Finally,the separated signal is desirable to achieve these space-time limits in order to optimize the collected by a single-element detector such as a photomultiplier detected optical signals and get the most out of each pixel. tube(PMT). In addition to maximizing information content,deep imaging This scanning behaviour and point-by-point detection is a defin-is also of great interest to the biological community.As a result,a ing characteristic of most MPM systems,and differs significantly variety of approaches have been developed to help multiphoton from more traditional 'whole-field'microscopy platforms,which microscopes overcome depth limitations.In this Review,we discuss generally use a 2D detector such as a CCD camera to collect data a number of strategies and design constraints for imaging at depth. Center for Microintegrated Optics for Advanced Bioimaging and Control,Colorado School of Mines,1523 Illinois Street,Golden,Colorado 80401,USA. 2Department of Physics,Colorado School of Mines,1523 lllinois Street,Golden,Colorado 80401,USA.e-mail:ehoover@mines.edu,jsquier@mines.edu NATURE PHOTONICS VOL 7|FEBRUARY 2013 www.nature.com/naturephotonics 号 2013 Macmillan Publishers Limited.All rights reserved
© 2013 Macmillan Publishers Limited. All rights reserved. NATURE PHOTONICS | VOL 7 | FEBRUARY 2013 | www.nature.com/naturephotonics 93 New windows in biological exploration are being opened through the continuing development of novel optical multiphoton microscopy (MPM) techniques. In this imaging paradigm, near-infrared (near-IR) femtosecond lasers are used to excite optical processes that can be accessed only through the application of two (or more) photons. Two-photon excitation fluorescence (TPEF) — a process driven by the simultaneous absorption of two near-infrared photons by a single fluorophore — is one example of such a technique1 . The probability of triggering a multiphoton process, such as TPEF, is extremely unlikely to occur. Interactions are therefore restricted to the focal plane of the objective, where the beam intensity is maximized, which provides the optical sectioning necessary for the non-perturbative analysis of living systems. Since its inception, MPM has permitted a variety of unique explorations into highly scattering materials. These studies have examined membrane potentials on the single-molecule scale2 , the non-invasive observation of embryo development3 and the simultaneous multiplane imaging of calcium transportation in transgenic mice4 . The ability to perform such explorations is a direct result of the inherent optical sectioning of multiphoton microscopes1,5 and the reduction in photobleaching outside of the imaging plane1,5–7. Multiphoton microscopes also benefit from the ability to utilize longer excitation wavelengths (700 nm and greater) than confocal techniques, thus making them less biologically harmful6,7 and more penetrating in scattering tissue7–11. In addition, multiphoton microscopes can frequently take advantage of the endogenous contrast mechanisms inherent to many samples, thus permitting the exploration of untreated specimens3,12–15. As shown in Fig. 1, a typical multiphoton microscope is composed of a femtosecond laser, a scanning system, a low-magnification high-numerical-aperture (NA) microscope objective, a wavelength-sensitive dichroic and a single-element detector. The scanning system is an intermediate optical system that is used to raster the excitation beam in a two-dimensional (2D) field at the full NA of the objective. The objective is generally used both for excitation and to collect the signal photons. These signal photons are separated from the excitation beam on the return path using a wavelength-sensitive dichroic. Finally, the separated signal is collected by a single-element detector such as a photomultiplier tube (PMT). This scanning behaviour and point-by-point detection is a defining characteristic of most MPM systems, and differs significantly from more traditional ‘whole-field’ microscopy platforms, which generally use a 2D detector such as a CCD camera to collect data Advances in multiphoton microscopy technology Erich E. Hoover1,2* and Jeff A. Squier1,2* Multiphoton microscopy has enabled unprecedented dynamic exploration in living organisms. A significant challenge in biological research is the dynamic imaging of features deep within living organisms, which permits the real-time analysis of cellular structure and function. To make progress in our understanding of biological machinery, optical microscopes must be capable of rapid, targeted access deep within samples at high resolution. In this Review, we discuss the basic architecture of a multiphoton microscope capable of such analysis and summarize the state-of-the-art technologies for the quantitative imaging of biological phenomena. from the entire imaging plane simultaneously. This important distinction allows multiphoton microscopes to perform efficient, deep explorations within scattering tissues10,16 by including data from multiply scattered signal photons, which might otherwise introduce a background fog7,9. Whole-field detection comes at the cost of confining imaging to within a few tens of micrometres of the surface of a scattering sample; this trade-off is an important consideration when exploring biological behaviour in scattering media. The multiphoton microscope design discussed here allows MPM to address one of the toughest challenges since the inception of optical microscopy: achieving image contrast at the cellular level in thick, scattering specimens. By their very nature, cells are quite thin (a few micrometres); such small path lengths provide little absorption, path length differences or scattering — all of which can provide a detailed view of the intricacies of the biological machinery. The use of a femtosecond laser as the light source for the microscope has enabled the generation of an entirely new class of contrast mechanisms within such specimens. Remarkably, these lasers provide intensities of the order of tens of gigawatts per square centimetre with modest focusing (for example, 0.65 NA) and relatively low average powers (milliwatts). At such intensities, the electric field at the focus creates a separation of positive and negative charges, thus momentarily polarizing the material. In fact, the induced timevarying polarization is significantly overdriven, resulting in the generation of new optical signals — distinctive from the excitation beam — that can be used to visualize structure and function in an unprecedented fashion. Substantive reviews of the broad array of nonlinear processes now used in imaging can be found in studies by Yue et al.17 and Chung et al.18. Today, this nonlinear polarization can be generated with pulses as short as 10 fs, which represents a pulse consisting of only five or six optical cycles19–21. This result is remarkable given the complexity of the optical system that is necessary to deliver such broadband light to the specimen. Combining scan optics with well-corrected highNA optics poses significant challenges for producing a focal spot that is both diffraction-limited in space and transform-limited in time. Whether a pulse of 100 fs or 10 fs is used as the excitation source, it is desirable to achieve these space–time limits in order to optimize the detected optical signals and get the most out of each pixel. In addition to maximizing information content, deep imaging is also of great interest to the biological community. As a result, a variety of approaches have been developed to help multiphoton microscopes overcome depth limitations. In this Review, we discuss a number of strategies and design constraints for imaging at depth. 1 Center for Microintegrated Optics for Advanced Bioimaging and Control, Colorado School of Mines, 1523 Illinois Street, Golden, Colorado 80401, USA. 2 Department of Physics, Colorado School of Mines, 1523 Illinois Street, Golden, Colorado 80401, USA. *e-mail: ehoover@mines.edu, jsquier@mines.edu REVIEW ARTICLES PUBLISHED ONLINE: 31 JANUARY 2013 |DOI: 10.1038/NPHOTON.2013.361
REVIEW ARTICLES NATURE PHOTONICS DOL:10.1038/NPHOTON.2013.361 Pulsed near-IR laser Specimen Objective Figure 1|A typical multiphoton microscope fed by a near-IR laser.Typical multiphoton systems utilize near-IR(700-1,300 nm)light and use a raster scanning system to control the beam,either with 'close coupled'scan mirrors or with image-relayed scan mirrors (SM,and SM,as shown here).In this epi-detection configuration,a dichroic (D)is used to separate two-photon excited fluorescence from the excitation light and direct this fluorescence to a PMT.L=lens. Finally,we survey a number of technologies that can be used to basic light-matter interaction perspective.Additionally,there is a increase the frame rate of a multiphoton microscope,and thus its pragmatic side to extending this limit,as the extreme bandwidth of ability to measure dynamics.From this perspective,the most sig- short pulses(>100 nm)requires a highly achromatic imaging system. nificant issue is photon scarcity at high imaging rates(>30 Hz for a Designing for this constraint also benefits users who operate at longer 2D image-less than 1 us per pixel),as the number of laser pulses pulse durations(100 fs)but desire an efficient tunable microscope per pixel dwell time and the excitation efficiency of the nonlinearity over tens or even hundreds of nanometres.Furthermore,higher- of interest become critical issues that dictate image contrast. order dispersion compensation,even at these modest pulsewidths, After taking into account these disparate issues involved in gen- can have a quantifiable increase in the detected photon yield2. erating image contrast,MPM provides a set of dynamic tools for Once diffraction-limited focal conditions are achieved,a remark- addressing a variety of problems.This Review will help facilitate able number of multiphoton processes become accessible.An enor- an understanding of the strengths and limitations of many of the mous amount of information can be obtained from each image pixel common MPM techniques,thus allowing the reader to utilize MPM simultaneously,which typically encompasses a femtolitre volume of to its full potential for addressing a variety of real-world imaging the specimen.The most commonly exploited nonlinear processes tasks.The range of technology being developed in this field is truly so far include absorptive mechanisms,such as TPEF,and paramet- impressive and,as such,the scope of this Review is limited.More ric processes,such as second-harmonic generation(SHG)2,third- than ever,it is important to consult the literature when developing a harmonic generation,sum-frequency generations2,stimulated multiphoton microscope for specific applications?2. Raman scattering and coherent anti-Stokes Raman scattering Used in combination,these techniques provide information about Getting the most out of each pixel a microscopic environment in terms of the chemical,structural and There is an incredible amount of information available at the focus operative mechanisms within living systems. of a multiphoton microscope;however,optimizing image content is only possible by paying careful attention to the production of a well- Multimodal imaging.Lasers capable of simultaneously and effi- focused pulse in terms of both the spatial wavefront and the tempo- ciently exciting a broad range of these nonlinearities are not pro- ral pulsefront.This attention is crucial because the production of a hibitively complex,as demonstrated by Chen et al..Their system well-focused pulse ensures the highest possible intensity at the focus, incorporates a femtosecond laser that pumps an optical parametric thereby maximizing the multiphoton signal generation.The qual- oscillator in tandem.The fundamental beam from the Ti:sapphire ity of this focus is reduced by linear dispersion from the refractive oscillator is tuned to 790 nm and used to drive the optical parametric optics in the microscope resulting from:first,an increase in the pulse oscillator,as well as provide an excitation source for TPEF and SHG duration;and,second,asymmetric distortion of the pulse in time2 imaging.The optical parametric oscillator signal(1,290 nm)and Most of these effects can effectively be pre-compensated through any idler(2,036 nm)beams perform several functions.The 1,290 nm number of means,including prism pairs,dispersion-compensating beam can be used for both SHG and third-harmonic generation mirrors and active pulse-shaping schemes.Work on improving imaging,whereas the frequency-doubled idler beam (1,018 nm)is compensation to enable the production of extremely short pulses used as the Stokes wavelength in conjunction with the main laser (10 fs and less)is particularly exciting.As stated earlier,with care- wavelength(790 nm)to provide a pump for coherent anti-Stokes ful attention to the net dispersion of the microscope,pulsewidths of Raman scattering,which is tuned to the vibrational CH,stretch less than 10 fs can be produced at the focus-Invariably,as the suitable for lipid detection. pulse duration limits are advanced,researchers also push the bound- Figure 2 is an image of a blood vessel in kidney tissue-an aries for new discoveries,from both an imaging perspective and a excellent example of a multimodal image that combines the 94 NATURE PHOTONICS VOL 7 FEBRUARY 2013 www.nature.com/naturephotonics 2013 Macmillan Publishers Limited.All rights reserved
© 2013 Macmillan Publishers Limited. All rights reserved. 94 NATURE PHOTONICS | VOL 7 | FEBRUARY 2013 | www.nature.com/naturephotonics Finally, we survey a number of technologies that can be used to increase the frame rate of a multiphoton microscope, and thus its ability to measure dynamics. From this perspective, the most significant issue is photon scarcity at high imaging rates (>30 Hz for a 2D image — less than 1 μs per pixel), as the number of laser pulses per pixel dwell time and the excitation efficiency of the nonlinearity of interest become critical issues that dictate image contrast. After taking into account these disparate issues involved in generating image contrast, MPM provides a set of dynamic tools for addressing a variety of problems. This Review will help facilitate an understanding of the strengths and limitations of many of the common MPM techniques, thus allowing the reader to utilize MPM to its full potential for addressing a variety of real-world imaging tasks. The range of technology being developed in this field is truly impressive and, as such, the scope of this Review is limited. More than ever, it is important to consult the literature when developing a multiphoton microscope for specific applications7,22–26. Getting the most out of each pixel There is an incredible amount of information available at the focus of a multiphoton microscope; however, optimizing image content is only possible by paying careful attention to the production of a wellfocused pulse in terms of both the spatial wavefront and the temporal pulsefront. This attention is crucial because the production of a well-focused pulse ensures the highest possible intensity at the focus, thereby maximizing the multiphoton signal generation. The quality of this focus is reduced by linear dispersion from the refractive optics in the microscope resulting from: first, an increase in the pulse duration; and, second, asymmetric distortion of the pulse in time25,26. Most of these effects can effectively be pre-compensated through any number of means, including prism pairs, dispersion-compensating mirrors and active pulse-shaping schemes. Work on improving compensation to enable the production of extremely short pulses (10 fs and less) is particularly exciting. As stated earlier, with careful attention to the net dispersion of the microscope, pulsewidths of less than 10 fs can be produced at the focus19–21. Invariably, as the pulse duration limits are advanced, researchers also push the boundaries for new discoveries, from both an imaging perspective and a basic light–matter interaction perspective. Additionally, there is a pragmatic side to extending this limit, as the extreme bandwidth of short pulses (>100 nm) requires a highly achromatic imaging system. Designing for this constraint also benefits users who operate at longer pulse durations (100 fs) but desire an efficient tunable microscope over tens or even hundreds of nanometres. Furthermore, higherorder dispersion compensation, even at these modest pulsewidths, can have a quantifiable increase in the detected photon yield27. Once diffraction-limited focal conditions are achieved, a remarkable number of multiphoton processes become accessible. An enormous amount of information can be obtained from each image pixel simultaneously, which typically encompasses a femtolitre volume of the specimen. The most commonly exploited nonlinear processes so far include absorptive mechanisms, such as TPEF1 , and parametric processes, such as second-harmonic generation (SHG)28, thirdharmonic generation12, sum-frequency generation15,29, stimulated Raman scattering30 and coherent anti-Stokes Raman scattering17,31. Used in combination, these techniques provide information about a microscopic environment in terms of the chemical, structural and operative mechanisms within living systems. Multimodal imaging. Lasers capable of simultaneously and efficiently exciting a broad range of these nonlinearities are not prohibitively complex, as demonstrated by Chen et al.32. Their system incorporates a femtosecond laser that pumps an optical parametric oscillator in tandem. The fundamental beam from the Ti:sapphire oscillator is tuned to 790 nm and used to drive the optical parametric oscillator, as well as provide an excitation source for TPEF and SHG imaging. The optical parametric oscillator signal (1,290 nm) and idler (2,036 nm) beams perform several functions. The 1,290 nm beam can be used for both SHG and third-harmonic generation imaging, whereas the frequency-doubled idler beam (1,018 nm) is used as the Stokes wavelength in conjunction with the main laser wavelength (790 nm) to provide a pump for coherent anti-Stokes Raman scattering, which is tuned to the vibrational CH2 stretch suitable for lipid detection. Figure 2 is an image of a blood vessel in kidney tissue — an excellent example of a multimodal image that combines the L L L PMT D Objective Specimen L L SMy SMx Pulsed near-IR laser Figure 1 | A typical multiphoton microscope fed by a near-IR laser. Typical multiphoton systems utilize near-IR (700–1,300 nm) light and use a raster scanning system to control the beam, either with ‘close coupled’ scan mirrors or with image-relayed scan mirrors (SMx and SMy, as shown here). In this epi-detection configuration, a dichroic (D) is used to separate two-photon excited fluorescence from the excitation light and direct this fluorescence to a PMT. L = lens. REVIEW ARTICLES NATURE PHOTONICS DOI: 10.1038/NPHOTON.2013.361
NATURE PHOTONICS DOL:10.1038/NPHOTON.2013.361 REVIEW ARTICLES aforementioned contrast mechanisms.In this case,the SHG signal for many efforts to use pulse shape as a contrast mechanism is the (blue)delineates collagen,the TPEF signal (green)marks the elastin formative work of Meshulach and Silberberg3.For example,by of the vascular wall in addition to intracellular nicotinamide ade- controlling the third-order spectral phase of a broadband excitation nine dinucleotide (NADH),and the coherent anti-Stokes Raman pulse,Pillai et al.demonstrated selective TPEF imaging in living scattering signal (red)shows lipids in adipose cells.The image was Drosophila embryos.In this approach,phase-only control enables taken at 0.75 NA,with a pixel dwell time of 4 us and a field-of-view selective excitation of either endogenous fluorescence or enhanced measuring300m×300m. green fluorescence protein(eGFP)-labelled bodies,and altering the pulse shape at kilohertz rates readily enables dynamic imaging. Fluorescence lifetime.Remarkably,even considering this broad New classes of contrast mechanisms can also be exploited if one array of contrast mechanisms,there is still more information to be alters the pulse amplitude,rather than relying solely on phase control. had from each pixel of a multiphoton image.The environment can For example,by reshaping the pulse such that a high-intensity fast often be further explored,for example,by measuring fluorophore component resides on a slower low-intensity background compo- lifetimes.Additionally,lifetime measurements can provide a mecha- nent,with each component consisting of equal areas,it becomes pos- nism for discriminating between different fluorescent labels that sible to measure the amount of two-photon absorption or self-phase may have spectrally similar signatures.Fluorescent lifetime imaging modulation that is accumulated by the pulse at the focal plane- lends itself quite naturally to TPEF imaging as a result of the three- This pulse shape is created by effectively masking out the central fre- dimensional (3D)confinement of the excitation.Time-correlated quency of the pulse in the spectral domain(that is,digging a hole single-photon counting is one of the most mature technologies for at the central wavelength in the pulse spectrum).At the focus,this performing lifetime measurements,and it is extremely well-suited spectral hole is refilled through two-photon absorption processes or to today's multiphoton imaging platforms In this approach,the self-phase modulation.Fortunately,these two mechanisms can be dis- lifetime of a sample is given by a histogram built from the arrival tinguished as the field at the replenished frequency is 90 out of phase times of individual signal photons collected by a fast detector (for for two-photon absorption,with respect to self-phase modulation. example,a PMT),which makes it suitable for use even within scat- Significantly,endogenous molecular tags such as melanin or haemo- tering specimens. globin",which are nominally transparent,can be distinguished by Figure 3 is an example of using lifetime measurements to dis- using two-photon absorption as the contrast agent,whereas neuronal criminate between spectrally similar fluorophoresss.In this case, activity can be tracked using self-phase modulation. cells labelled with propidium iodide and vessels labelled with Texas Red dextran are indistinguishable when measuring the intensity Imaging deep of the TPEF signal alone.This situation is particularly evident in One task that compounds the challenge of generating image contrast Fig.3a.However,if the image is reformulated based on the fluores- in a thin specimen (such as a cell)is the task of imaging cells and cence lifetime (Fig.3b),contrast between the labels becomes evi- cellular function while embedded deep(hundreds of micrometres) dent.A final image based on fluorescent lifetime and photon counts within an organism.Switching to the longer wavelengths necessary renders a composite image(Fig.3c)that enables the unambiguous to promote efficient multiphoton excitation and detection(near- determination of the fluorophore and its targeted structure. infrared,750-1,100 nm)can increase image depths by a factor of two or three in multiphoton systems,when compared with their tra- Pulse shaping.Fluorophores can also be selectively excited or dis- ditional confocal counterparts.These wavelengths are intrinsically tinguished by altering the shape of the excitation pulse The basis more penetrating owing to their increased scattering length,with the maximum wavelength being limited by the absorption properties of the materials in the specimen.In neuronal tissue-a common MPM application-this limit is set by the blood and water in the brain and therefore limits the excitation wavelength to around 1,300 nm (ref.10).However,it is important to note that the two-photon cross- section for any fluorophore is spectrally dependent and can therefore also limit the excitation wavelength when performing TPEE High-energy lasers.Different strategies can be employed to push the maximum imaging depth,which has now exceeded 1 mm.To maintain sufficient intensity at the focus when reaching signifi- cant depths in scattering media,one of the primary tactics is to increase the energy of the excitation pulse44.For example,using ~150 fs pulses,amplified to the microjoule level (at a repetition rate of 200 kHz)and centred at a wavelength of 953 nm,Theer et al.6 used TPEF to image GFP-labelled neurons at depths of up to a mil- limetre within a sample.This strategy functions as a result of the signal dependence on unscattered (or ballistic)excitation light.As the focus is pushed deeper into the specimen,the excitation beam is depleted of these ballistic photons,primarily as a result of scattering, and the excitation efficiency is subsequently reduced.Increasing the pulse energy therefore results in more ballistic photons at depth,but this approach has its limits. In fact,it has been shown that in biological tissue the ballistic power decreases exponentially with depth as a result of scattering? Consequently,as the input power is increased to counteract this Figure2 Multimodal image of a blood vessel in kidney tissue.SHG effect,a new problem emerges.The beam intensity becomes high (blue),TPEF (green)and coherent anti-Stokes Raman scattering (red). enough such that tissue at the surface of the sample-outside of Image courtesy of Eric Potma,University of California,Irvine,USA. the perifocal region-can fluoresce.This out-of-focus fluorescence NATURE PHOTONICS VOL 7|FEBRUARY 2013 www.nature.com/naturephotonics 2013 Macmillan Publishers Limited.All rights reserved
© 2013 Macmillan Publishers Limited. All rights reserved. NATURE PHOTONICS | VOL 7 | FEBRUARY 2013 | www.nature.com/naturephotonics 95 aforementioned contrast mechanisms. In this case, the SHG signal (blue) delineates collagen, the TPEF signal (green) marks the elastin of the vascular wall in addition to intracellular nicotinamide adenine dinucleotide (NADH), and the coherent anti-Stokes Raman scattering signal (red) shows lipids in adipose cells. The image was taken at 0.75 NA, with a pixel dwell time of 4 μs and a field-of-view measuring 300 μm × 300 μm. Fluorescence lifetime. Remarkably, even considering this broad array of contrast mechanisms, there is still more information to be had from each pixel of a multiphoton image. The environment can often be further explored, for example, by measuring fluorophore lifetimes. Additionally, lifetime measurements can provide a mechanism for discriminating between different fluorescent labels that may have spectrally similar signatures. Fluorescent lifetime imaging lends itself quite naturally to TPEF imaging as a result of the threedimensional (3D) confinement of the excitation. Time-correlated single-photon counting is one of the most mature technologies for performing lifetime measurements, and it is extremely well-suited to today’s multiphoton imaging platforms33,34. In this approach, the lifetime of a sample is given by a histogram built from the arrival times of individual signal photons collected by a fast detector (for example, a PMT), which makes it suitable for use even within scattering specimens. Figure 3 is an example of using lifetime measurements to discriminate between spectrally similar fluorophores35. In this case, cells labelled with propidium iodide and vessels labelled with Texas Red dextran are indistinguishable when measuring the intensity of the TPEF signal alone. This situation is particularly evident in Fig. 3a. However, if the image is reformulated based on the fluorescence lifetime (Fig. 3b), contrast between the labels becomes evident. A final image based on fluorescent lifetime and photon counts renders a composite image (Fig. 3c) that enables the unambiguous determination of the fluorophore and its targeted structure. Pulse shaping. Fluorophores can also be selectively excited or distinguished by altering the shape of the excitation pulse36–38.The basis for many efforts to use pulse shape as a contrast mechanism is the formative work of Meshulach and Silberberg39. For example, by controlling the third-order spectral phase of a broadband excitation pulse, Pillai et al.40 demonstrated selective TPEF imaging in living Drosophila embryos. In this approach, phase-only control enables selective excitation of either endogenous fluorescence or enhanced green fluorescence protein (eGFP)-labelled bodies, and altering the pulse shape at kilohertz rates readily enables dynamic imaging. New classes of contrast mechanisms can also be exploited if one alters the pulse amplitude, rather than relying solely on phase control. For example, by reshaping the pulse such that a high-intensity fast component resides on a slower low-intensity background component, with each component consisting of equal areas, it becomes possible to measure the amount of two-photon absorption or self-phase modulation that is accumulated by the pulse at the focal plane41–43. This pulse shape is created by effectively masking out the central frequency of the pulse in the spectral domain (that is, digging a hole at the central wavelength in the pulse spectrum). At the focus, this spectral hole is refilled through two-photon absorption processes or self-phase modulation. Fortunately, these two mechanisms can be distinguished as the field at the replenished frequency is 90° out of phase for two-photon absorption, with respect to self-phase modulation. Significantly, endogenous molecular tags such as melanin or haemoglobin41, which are nominally transparent, can be distinguished by using two-photon absorption as the contrast agent, whereas neuronal activity can be tracked using self-phase modulation43. Imaging deep One task that compounds the challenge of generating image contrast in a thin specimen (such as a cell) is the task of imaging cells and cellular function while embedded deep (hundreds of micrometres) within an organism. Switching to the longer wavelengths necessary to promote efficient multiphoton excitation and detection (nearinfrared, 750–1,100 nm) can increase image depths by a factor of two or three in multiphoton systems, when compared with their traditional confocal counterparts. These wavelengths are intrinsically more penetrating owing to their increased scattering length, with the maximum wavelength being limited by the absorption properties of the materials in the specimen. In neuronal tissue — a common MPM application — this limit is set by the blood and water in the brain and therefore limits the excitation wavelength to around 1,300 nm (ref. 10). However, it is important to note that the two-photon crosssection for any fluorophore is spectrally dependent and can therefore also limit the excitation wavelength when performing TPEF. High-energy lasers. Different strategies can be employed to push the maximum imaging depth, which has now exceeded 1 mm. To maintain sufficient intensity at the focus when reaching significant depths in scattering media, one of the primary tactics is to increase the energy of the excitation pulse44–46. For example, using ~150 fs pulses, amplified to the microjoule level (at a repetition rate of 200 kHz) and centred at a wavelength of 953 nm, Theer et al.16 used TPEF to image GFP-labelled neurons at depths of up to a millimetre within a sample. This strategy functions as a result of the signal dependence on unscattered (or ballistic) excitation light. As the focus is pushed deeper into the specimen, the excitation beam is depleted of these ballistic photons, primarily as a result of scattering, and the excitation efficiency is subsequently reduced. Increasing the pulse energy therefore results in more ballistic photons at depth, but this approach has its limits. In fact, it has been shown that in biological tissue the ballistic power decreases exponentially with depth as a result of scattering7 . Consequently, as the input power is increased to counteract this effect, a new problem emerges. The beam intensity becomes high enough such that tissue at the surface of the sample — outside of the perifocal region — can fluoresce. This out-of-focus fluorescence Figure 2 | Multimodal image of a blood vessel in kidney tissue. SHG (blue), TPEF (green) and coherent anti-Stokes Raman scattering (red). Image courtesy of Eric Potma, University of California, Irvine, USA. NATURE PHOTONICS DOI: 10.1038/NPHOTON.2013.361 REVIEW ARTICLES
REVIEW ARTICLES NATURE PHOTONICS DOL:10.1038/NPHOTON.2013.361 results in undesired photons obscuring the features of interest and, Total counts 6,400 Composite once again,limits the depth at which effective imaging can be per- formed.It is this undesired fluorescence that limited the amplified ,600 approach of Theer et al.,as the features in their images became clouded at depths of around 1 mm.This loss of signal compared 600 with the noise is not a result of limited pulse energy-only 225 nl 400 of the ~3 uJ available (roughly 29%of the available laser power) 0 was used-but rather results from the out-of-focus fluorescence at the surface of the specimen.Hence,alternative strategies are now b Lifetime,r actively being pursued. 6 Long-wavelength excitation.One of the most effective tactics 4 for imaging at depth exploits a key feature that made nonlinear imaging compelling in the first place:the use of longer excitation 0 wavelengths.By moving away from 800 nm towards 1,280 nm, Kobat et al.1047 have been able to perform in vivo TPEF imaging in a mouse cortex at depths as great as 1.6 mm(Fig.4).This improve- Figure3 lllustrative fluorescence lifetime image with two similar ment in depth is a result of decreased scattering at the 1,280 nm fluorophores and comparison to TPEF imaging.Fluorescence intensity wavelength generated from their Ti:sapphire pumped optical para- and lifetime imaging of propidium iodide(Pl)-labelled cells and Texas Red metric oscillator.This choice of laser is significant,as it provides a dextran (TR)-labelled vessels in a mouse model.a,TPEF image shows that high-repetition-rate (80 MHz)pulse train with modest pulse ener- the two dyes are indistinguishable.Scale bar (right)represents photon gies(~1.5 n),which facilitates rapid imaging.Although the use ofa counts.b,Image is rescaled according to the measured fluorescent lifetime; longer wavelength compromises the resolution slightly,the benefits the Pl-label and the TR-label are now spatially distinct.Scale bar (right)is of improved depth penetration and facile multimodal detection34 in nanoseconds.c,The images in a and b are combined,thus enabling facile makes the concession worthwhile for many applications. detection of the two fluorophores.The arrowhead points to a Pl-labelled cell,whereas the arrow points to a TR-labelled vessel.Figure reproduced Imaging through gradient-index lenses.The dual complications with permission from ref.35,2011 APS. of reduced power as a function of depth and increased out-of-focus background fluorescence can be completely obviated through the Photon counting.As the imaging depth is increased,another addi- use of gradient-index(GRIN)lenses.This technique was demon- tional complication is the scattering of the signal photons.If the sig- strated by Levene et al.s,who used needle-like(320 um diameter) nal light is collected in a non-imaging modality using single-element GRIN lenses that can penetrate directly into the specimen and per- detection,such as with a PMT,multiple scattering events en route form in vivo multiphoton imaging at depths of several millimetres. to the detector are not necessarily detrimental.Essentially,collect- Appropriately engineered GRIN lenses effectively relay the focal ing scattered light at angles or in regions beyond the cone of light plane of the microscope over tens of millimetres(even centimetre) defined by the excitation beam enforces the requirement of main- distances,as the lens is pressed into the tissue up to the layer of taining high-NA collection over a large field-of-view;hence the interest.Using 0.6-NA GRIN lenses,Levene et al.achieved a circu- drive towards high-NA,low-magnification objectivess.Having col- lar field-of-view measuring 58 um in diameter and axially scanned lected the light,it is often the case when working in this regime that over a distance of 95 um without needing to shift the GRIN lens. there is essentially less than one signal photon generated per excita- A natural extension of GRIN technology is to consider complete tion pulse.In such a situation,it becomes beneficial to incorporate endoscopic multiphoton imaging platforms.Indeed,this is a vibrant photon counting detection in order to discriminate signal photons area of developments-s and features millimetre-diameter probes from background noise.Until recently,this was considered prohibi- that are suitable for clinical applicationss tive given the repetition rates of the lasers,which are in the range of 70-100 MHz.However,with the advent of inexpensive,high-perfor- Photo-activatable fluorophores.Other approaches for imag- mance microelectronics such as field-programmable gate arrays0-3 ing at depth that are less invasive than GRIN technology include the implementation of photon-counting circuitry is not only quite the incorporation of photo-activatable fluorophoresss,as recently feasible,but also very economical.Driscoll et al.s have shown that demonstrated by Chen et al..In their technique,the fluorophores through proper implementation of photon counting,and by account- remain in a dark state(that is,a non-fluorescent state)until opti- ing for the censor period of the detector,the signal-to-noise ratio can cally triggered by multiphoton excitation.In this case,the ratio of be measurably improved.This improvement is sufficient to extend the signal-to-background fluorescence is improved by using one photon counting for operation in the high-emission-rate regime, multiphoton source,centred at 830 nm,to activate the fluoro- where analogue integration is generally thought to be requiredss phores,and a second source,centred at 920 nm,to produce TPEF signal from the activated sites.This multiphoton activation strategy Adaptive optics.A final notable consideration for improving mul- allows a larger number of fluorophores to be activated within the tiphoton imaging at depth is the incorporation of adaptive optical focal plane compared with the out-of-focus regions,thus resulting schemes.The breadth ofinnovation in terms of adaptive optical cor- in a measurable increase in the signal-to-background ratio.Indeed, rection is worthy of a review in and of itself,and is therefore only starting with control samples that are engineered to mimic the fun- briefly considered here.One of the most intriguing pathways for damental depth limits(where the signal-to-background ratio equals the implementation of adaptive optics,with respect to deep imag- unity),Chen et al.s6 have demonstrated signal-to-background ratios ing,is to incorporate a system that is capable of rapidly adjusting the of around 20 by using the photo-activation approach.In general, wavefront to accommodate aberrations induced by both the opti- customizing probess for deep imaging,as briefly discussed here,is cal delivery system and the specimen without a direct assessment a field in and of itself (see Extermann et al.for an example of a of the aberrated wavefront7.Rather than assessing the wavefront deep SHG probe-a completely different approach from the one directly,the image is corrected based on metrics derived from the discussed here)and further elaboration is outside the scope of image itself.This 'sensorless'approach has recently been analysed in this Review. detail by Facomprezet al.",who established a useful series of guiding 96 NATURE PHOTONICS VOL 7 FEBRUARY 2013 www.nature.com/naturephotonics 2013 Macmillan Publishers Limited.All rights reserved
© 2013 Macmillan Publishers Limited. All rights reserved. 96 NATURE PHOTONICS | VOL 7 | FEBRUARY 2013 | www.nature.com/naturephotonics results in undesired photons obscuring the features of interest and, once again, limits the depth at which effective imaging can be performed9 . It is this undesired fluorescence that limited the amplified approach of Theer et al.16, as the features in their images became clouded at depths of around 1 mm. This loss of signal compared with the noise is not a result of limited pulse energy — only 225 nJ of the ~3 μJ available (roughly 29% of the available laser power) was used — but rather results from the out-of-focus fluorescence at the surface of the specimen. Hence, alternative strategies are now actively being pursued. Long-wavelength excitation. One of the most effective tactics for imaging at depth exploits a key feature that made nonlinear imaging compelling in the first place: the use of longer excitation wavelengths. By moving away from 800 nm towards 1,280 nm, Kobat et al.10,47 have been able to perform in vivo TPEF imaging in a mouse cortex at depths as great as 1.6 mm (Fig. 4). This improvement in depth is a result of decreased scattering at the 1,280 nm wavelength generated from their Ti:sapphire pumped optical parametric oscillator. This choice of laser is significant, as it provides a high-repetition-rate (80 MHz) pulse train with modest pulse energies (~1.5 nJ), which facilitates rapid imaging. Although the use of a longer wavelength compromises the resolution slightly, the benefits of improved depth penetration48 and facile multimodal detection3,49 makes the concession worthwhile for many applications. Imaging through gradient-index lenses. The dual complications of reduced power as a function of depth and increased out-of-focus background fluorescence can be completely obviated through the use of gradient-index (GRIN) lenses. This technique was demonstrated by Levene et al.50, who used needle-like (320 μm diameter) GRIN lenses that can penetrate directly into the specimen and perform in vivo multiphoton imaging at depths of several millimetres. Appropriately engineered GRIN lenses effectively relay the focal plane of the microscope over tens of millimetres (even centimetre) distances, as the lens is pressed into the tissue up to the layer of interest. Using 0.6-NA GRIN lenses, Levene et al. achieved a circular field-of-view measuring 58 μm in diameter and axially scanned over a distance of 95 μm without needing to shift the GRIN lens. A natural extension of GRIN technology is to consider complete endoscopic multiphoton imaging platforms. Indeed, this is a vibrant area of development51–54 and features millimetre-diameter probes that are suitable for clinical applications54. Photo-activatable fluorophores. Other approaches for imaging at depth that are less invasive than GRIN technology include the incorporation of photo-activatable fluorophores55, as recently demonstrated by Chen et al.56. In their technique, the fluorophores remain in a dark state (that is, a non-fluorescent state) until optically triggered by multiphoton excitation. In this case, the ratio of the signal-to-background fluorescence is improved by using one multiphoton source, centred at 830 nm, to activate the fluorophores, and a second source, centred at 920 nm, to produce TPEF signal from the activated sites. This multiphoton activation strategy allows a larger number of fluorophores to be activated within the focal plane compared with the out-of-focus regions, thus resulting in a measurable increase in the signal-to-background ratio. Indeed, starting with control samples that are engineered to mimic the fundamental depth limits (where the signal-to-background ratio equals unity), Chen et al.56 have demonstrated signal-to-background ratios of around 20 by using the photo-activation approach. In general, customizing probes57 for deep imaging, as briefly discussed here, is a field in and of itself (see Extermann et al.58 for an example of a deep SHG probe — a completely different approach from the one discussed here) and further elaboration is outside the scope of this Review. Photon counting. As the imaging depth is increased, another additional complication is the scattering of the signal photons. If the signal light is collected in a non-imaging modality using single-element detection, such as with a PMT, multiple scattering events en route to the detector are not necessarily detrimental. Essentially, collecting scattered light at angles or in regions beyond the cone of light defined by the excitation beam enforces the requirement of maintaining high-NA collection over a large field-of-view; hence the drive towards high-NA, low-magnification objectives59. Having collected the light, it is often the case when working in this regime that there is essentially less than one signal photon generated per excitation pulse. In such a situation, it becomes beneficial to incorporate photon counting detection in order to discriminate signal photons from background noise. Until recently, this was considered prohibitive given the repetition rates of the lasers, which are in the range of 70–100 MHz. However, with the advent of inexpensive, high-performance microelectronics such as field-programmable gate arrays60–63, the implementation of photon-counting circuitry is not only quite feasible, but also very economical. Driscoll et al.35 have shown that through proper implementation of photon counting, and by accounting for the censor period of the detector, the signal-to-noise ratio can be measurably improved. This improvement is sufficient to extend photon counting for operation in the high-emission-rate regime, where analogue integration is generally thought to be required35. Adaptive optics. A final notable consideration for improving multiphoton imaging at depth is the incorporation of adaptive optical schemes. The breadth of innovation in terms of adaptive optical correction is worthy of a review in and of itself, and is therefore only briefly considered here. One of the most intriguing pathways for the implementation of adaptive optics, with respect to deep imaging, is to incorporate a system that is capable of rapidly adjusting the wavefront to accommodate aberrations induced by both the optical delivery system and the specimen without a direct assessment of the aberrated wavefront64–70. Rather than assessing the wavefront directly, the image is corrected based on metrics derived from the image itself. This ‘sensorless’ approach has recently been analysed in detail by Facomprez et al.71, who established a useful series of guiding 6,400 3,600 1,600 400 0 a Total counts c b Composite Lifetime, τ 0 ns 2 4 6 8 10 10 µm 10 µm 10 µm Figure 3 | Illustrative fluorescence lifetime image with two similar fluorophores and comparison to TPEF imaging. Fluorescence intensity and lifetime imaging of propidium iodide (PI)-labelled cells and Texas Red dextran (TR)-labelled vessels in a mouse model. a, TPEF image shows that the two dyes are indistinguishable. Scale bar (right) represents photon counts. b, Image is rescaled according to the measured fluorescent lifetime; the PI-label and the TR-label are now spatially distinct. Scale bar (right) is in nanoseconds. c, The images in a and b are combined, thus enabling facile detection of the two fluorophores. The arrowhead points to a PI-labelled cell, whereas the arrow points to a TR-labelled vessel. Figure reproduced with permission from ref. 35, © 2011 APS. REVIEW ARTICLES NATURE PHOTONICS DOI: 10.1038/NPHOTON.2013.361
NATURE PHOTONICS DOL:10.1038/NPHOTON.2013.361 REVIEW ARTICLES 400 Surface 1,150 元 200 10 450 8 400 Slope=-0.007 1,200 600 1,500 & 4 800 1,250 2 1,000 1550 0- 02004006008001,0001.2001,4001,600 1,300 1.200 Depth (um) Depth 1.090 um 1,600 Depth 1,300 um Depth 1,460 um Depth 1,600 um 1,400 1,350 1,600 um μm 1,650 um Figure 4 Example of deep invivo imaging through the use of longer excitation wavelengths.1,280 nm light from an optical parametric oscillator is used to perform TPEF imaging of mouse vasculature labelled with Alexa680-Dextran.a,Invivo two-photon fluorescence images of cortical vasculature in mouse brain.235 x-y frames from 60 um above the cortical surface to 1,110 um below are taken at depth increments of 5 um.The depth increments in the stack are 20 um in the range of 1,110-1,490 um and 30 um in the range of 1,490-1,670 um.3D reconstruction is made in Image J software using the volume viewer plug-in.Expanded 3D stacks are shown for the deepest sections (>1,130 um).b,Fluorescence intensity as a function of imaging depth for the stack shown in a.Fluorescence signal strength at a particular depth is represented by the average value of the brightest 1%of the pixels in the x-y image at that depth.Scale bars are 50 um for both a and b.Figure reproduced with permission from ref.47,2011 SPIE. principles that can be employed to optimize adaptive optical strate- foci within the sample.The emitted signal photons generated by this gies.Interestingly,they demonstrated that adaptive systems incorpo- matrix must be imaged to their conjugate positions on the detector, rating this philosophy are compatible with biological systems,both as opposed to collecting all of the photons in single-element detec- in terms of the speed at which corrections can be implemented and tion.If the signal photons are scattered,they will not be correctly the light levels that must be used to achieve accurate correction. mapped to the conjugate image position by the optical system,thus resulting in a background haze in the images.This limitation can be High-speed imaging mitigated somewhat by introducing a segmented detector and uti- Owing to the raster-scanning nature of most imaging systems in lizing descanned detection,in which emitted photons are detected MPM and the limited number of emitted signal photons available after the scan system.Kim et al have successfully established this for constructing an image,accessing dynamic behaviour in a 3D strategy.In their configuration,a multi-anode PMT is used to match volume has proven to be an interesting challenge.Several different the coordinates of the foci within the sample such that each anode strategies for approaching rapid imaging are described here,but this receives the vast majority of photons emitted from a particular is by no means a comprehensive list.Each of these techniques comes focus".This mode of operation permits the multifocal microscope with its own particular strengths and weaknesses,which should be to operate in a similar fashion to that of a single-focal-spot,single- carefully weighed in order to adopt an optimal imaging approach. element detection system.Kim et al.have successfully demonstrated that using 64 foci can extend the effective imaging depth from less Multifocal microscopy.One of the most widely used strategies for than 30 um to around 75 um in neuronal tissue?9 improving the frame rate in MPM is the use of multiple foci to paral- lelize the imaging process.Simply put,by distributing the excitation High-speed scan systems.Another important strategy in high- light over multiple foci,the time required to scan the focal plane is speed imaging is simply to raster the beam as fast as possible.As reduced accordingly.For example,when scanning linearly,two foci such,polygonal mirrors and resonant scanners hold an important cover a fixed field-of-view in half the time,and subsequent gains in place in high-speed MPM,as these systems provide a way to image the frame-rate scale proportionately to the number of foci?.However, 2D areas at video rates-30 Hz (refs 80,81)-without losing the as the density of foci increases,the axial resolution decreases as a ability to explore deep within scattering tissue.In such systems, result of constructive interference between the foci.Fortunately,this practical image speeds are essentially dictated by the number of exci- problem can be overcome by delaying each focal spot temporally tation pulses per pixel dwell time.With lasers operating at repetition with respect to its neighbours by an amount of the order of the pulse rates of 75-100 MHz,pixel dwell times of the order of 150 ns are duration (or slightly greater).In this way,the interference is entirely needed to ensure~10 pulses per pixel.A second design consideration eliminated and the axial resolution from a 2D array of focal spots when optimizing the frame rate for systems scanned in this manner is equivalent to that of its single-focal-spot counterpart2-%6.Truly involves the scan's 'dead time.For polygonal mirrors this problem remarkable frame rates have been achieved through this approach. occurs when the laser beam hits the interface between mirror facets Bahlmann et al.7 have successfully exceeded frame rates of 600 Hz. and,for resonant scan mirrors,the nonlinear scan region where the In a multifocal microscope,a single-element detector can no mirrors are accelerating and decelerating. longer be used to collect the excited photons from the sample,so it becomes necessary to use a camera?8.The necessity for a 2D detector Acousto-optics and tunable lenses.Although polygonal mirrors stems from the implementation of a 2D spatial matrix of excitation and resonant scanners can permit rapid imaging,theylack flexibility NATURE PHOTONICS VOL 7 FEBRUARY 2013 www.nature.com/naturephotonics 97 2013 Macmillan Publishers Limited.All rights reserved
© 2013 Macmillan Publishers Limited. All rights reserved. NATURE PHOTONICS | VOL 7 | FEBRUARY 2013 | www.nature.com/naturephotonics 97 principles that can be employed to optimize adaptive optical strategies. Interestingly, they demonstrated that adaptive systems incorporating this philosophy are compatible with biological systems, both in terms of the speed at which corrections can be implemented and the light levels that must be used to achieve accurate correction. High-speed imaging Owing to the raster-scanning nature of most imaging systems in MPM and the limited number of emitted signal photons available for constructing an image, accessing dynamic behaviour in a 3D volume has proven to be an interesting challenge. Several different strategies for approaching rapid imaging are described here, but this is by no means a comprehensive list. Each of these techniques comes with its own particular strengths and weaknesses, which should be carefully weighed in order to adopt an optimal imaging approach. Multifocal microscopy. One of the most widely used strategies for improving the frame rate in MPM is the use of multiple foci to parallelize the imaging process. Simply put, by distributing the excitation light over multiple foci, the time required to scan the focal plane is reduced accordingly. For example, when scanning linearly, two foci cover a fixed field-of-view in half the time, and subsequent gains in the frame-rate scale proportionately to the number of foci72. However, as the density of foci increases, the axial resolution decreases as a result of constructive interference between the foci. Fortunately, this problem can be overcome by delaying each focal spot temporally with respect to its neighbours by an amount of the order of the pulse duration (or slightly greater). In this way, the interference is entirely eliminated and the axial resolution from a 2D array of focal spots is equivalent to that of its single-focal-spot counterpart72–76. Truly remarkable frame rates have been achieved through this approach. Bahlmann et al.77 have successfully exceeded frame rates of 600 Hz. In a multifocal microscope, a single-element detector can no longer be used to collect the excited photons from the sample, so it becomes necessary to use a camera78. The necessity for a 2D detector stems from the implementation of a 2D spatial matrix of excitation foci within the sample. The emitted signal photons generated by this matrix must be imaged to their conjugate positions on the detector, as opposed to collecting all of the photons in single-element detection. If the signal photons are scattered, they will not be correctly mapped to the conjugate image position by the optical system, thus resulting in a background haze in the images. This limitation can be mitigated somewhat by introducing a segmented detector and utilizing descanned detection, in which emitted photons are detected after the scan system. Kim et al.79 have successfully established this strategy. In their configuration, a multi-anode PMT is used to match the coordinates of the foci within the sample such that each anode receives the vast majority of photons emitted from a particular focus79. This mode of operation permits the multifocal microscope to operate in a similar fashion to that of a single-focal-spot, singleelement detection system. Kim et al. have successfully demonstrated that using 64 foci can extend the effective imaging depth from less than 30 μm to around 75 μm in neuronal tissue79. High-speed scan systems. Another important strategy in highspeed imaging is simply to raster the beam as fast as possible. As such, polygonal mirrors and resonant scanners hold an important place in high-speed MPM, as these systems provide a way to image 2D areas at video rates — 30 Hz (refs 80,81) — without losing the ability to explore deep within scattering tissue82. In such systems, practical image speeds are essentially dictated by the number of excitation pulses per pixel dwell time. With lasers operating at repetition rates of 75–100 MHz, pixel dwell times of the order of 150 ns are needed to ensure ~10 pulses per pixel. A second design consideration when optimizing the frame rate for systems scanned in this manner involves the scan’s ‘dead time’. For polygonal mirrors this problem occurs when the laser beam hits the interface between mirror facets and, for resonant scan mirrors, the nonlinear scan region where the mirrors are accelerating and decelerating. Acousto-optics and tunable lenses. Although polygonal mirrors and resonant scanners can permit rapid imaging, they lack flexibility a b Depth 1,090 µm Depth 1,300 µm Depth 1,460 µm Depth 1,600 µm In (2P Fluorescence signal) Depth (µm) 0 0 200 400 600 800 1,000 1,200 1,400 1,600 2 4 6 8 10 12 14 1,150 Slope = –0.007 1,400 1,450 1,500 1,550 1,600 1,650 200 400 600 800 1,000 1,200 1,400 1,600 Surface 1,200 1,250 1,300 1,350 µm µm µm Figure 4 | Example of deep in vivo imaging through the use of longer excitation wavelengths. 1,280 nm light from an optical parametric oscillator is used to perform TPEF imaging of mouse vasculature labelled with Alexa680-Dextran. a, In vivo two-photon fluorescence images of cortical vasculature in mouse brain. 235 x–y frames from 60 μm above the cortical surface to 1,110 μm below are taken at depth increments of 5 μm. The depth increments in the stack are 20 μm in the range of 1,110–1,490 μm and 30 μm in the range of 1,490–1,670 μm. 3D reconstruction is made in Image J software using the volume viewer plug-in. Expanded 3D stacks are shown for the deepest sections (>1,130 μm). b, Fluorescence intensity as a function of imaging depth for the stack shown in a. Fluorescence signal strength at a particular depth is represented by the average value of the brightest 1% of the pixels in the x–y image at that depth. Scale bars are 50 μm for both a and b. Figure reproduced with permission from ref. 47, © 2011 SPIE. NATURE PHOTONICS DOI: 10.1038/NPHOTON.2013.361 REVIEW ARTICLES