The Airyscan Detector: Confocal Microscopy Evolution for the Neurosciences

Part of the Progress in Optical Science and Photonics book series (POSP, volume 5)


First introduced in August 2014, the Airyscan detector from ZEISS represents a new detector concept for laser scanning microscopy (LSM) that enables a simultaneous resolution and signal-to-noise ratio (SNR) increase over traditional LSM imaging. The Airyscan detector innovation replaces the conventional LSM detector and pinhole scheme for an array of 32 sensitive GaAsP detector elements placed in the pinhole plane to generate an optical section. The new detection geometry allows for the collection of the spatial distribution of light originating from every point of a microscopic fluorescent object at the pinhole allowing access to higher frequency information and while additionally collecting more light for ultra-efficient imaging. In May 2015, the fast mode innovation for Airyscan combines illumination shaping with pinhole-plane imaging that enhances acquisition speeds by four times while simultaneously increasing SNR and resolution overcoming the traditional compromises of LSM imaging.

4.1 Introduction

Over the last 25 years, the technique of confocal imaging has grown to become the standard choice for most fluorescence microscopy applications. The increase in utilization of confocal imaging systems in basic biomedical research can be in attributed to the technique’s ability to produce optically sectioned images with high contrast while providing acquisition versatility to address many sample and application demands [1]. Hence, during the last decades in order to meet the growing utilization of confocal microscopy across many different fields of application and model systems, most commercially available confocal imaging systems have developed novel approaches and options to increase image contrast and instrument flexibility. While novel approaches of commercially available systems have certainly enabled researchers to answer new questions and address new model systems across, it is within the field of neuroscience that the confocal principal was born and is still driving the evolution of the most fundamental aspect of confocal microscopy: the creation of an optical section.

In 1955, as a means to study the circuitry of the central nervous system and alleviate the issue of light scattering from brain tissue, Marvin Minsky invented the confocal approach while a Junior Fellow at Harvard [2]. In order to image brain samples prepared by the Golgi method, a white light source was focused into the sample via an objective and the transmission was collected by a condenser behind which Minsky place a field stop, the so-called pinhole, in a conjugate image plane in front of a detector. The sample was then moved laterally to create an image on a point by point basis. Upon opening and closing the pinhole, a researcher could restrict the amount of out of focus light from reaching the detector and blurring the resultant image. If the pinhole was sufficiently closed, out of focus light collected by the objective will be blocked from reaching the detector creating an optically sectioned image (Fig. 4.1) and thus solving the scattering problem resulting from brain tissue. The foundation laid by Minsky’s work and the revolution of fluorescence microscopy has led to the growth and development of the laser scanning microscopy industry over the past 25 years (The LSM 800 and LSM 880 systems are ZEISS’s eight generation of LSMs). In order to meet current and future research needs, most laser scanning microscopy systems in the neurosciences are designed to help researchers study two fundamental features of the brain: structure and function.
Fig. 4.1

a Schematic of traditional laser scanning fluorescence microscope. b Neuronal fluorescence image acquired using a wide-field microscope (no optical section). c Optically sectioned neuronal fluorescence image acquired using a laser scanning confocal microscopy

4.2 Airyscan for Neurosciences

With the aim of understanding structure and function of the brain, confocal imaging was an integral part of extensive analysis and experiments, which would drive the development of the molecular tools as well as the microscopy systems over the decades. The growing number and increased functionality of fluorescent compounds would drive the integration of laser lines, spectral imaging, and photomanipulation options. Advanced analysis tools such as fluorescence-lifetime imaging (FLIM) and fluorescence correlative spectroscopy (FCS) give insights into physiological processes. High-end LSM systems have grown into multimodal imaging and analysis tools. Over this time, the constant hallmark of every commercial laser scanning confocal system has been the use of a physical aperture for a pinhole (as Minsky described in 1955) in combination with a unitary detector (Fig. 4.1). Unfortunately, as current neuroscience research needs advance and new molecular biology capabilities expand, the combination of a traditional pinhole and unitary detector are starting to serve as the limiting factor for utilizing confocal microscopy in the neurosciences. The conventional design choices of modern laser scanning confocal microscopy systems limit the achievable image resolution, signal-to-noise, and speed subsequently limiting the research questions that can be answered. For the neurosciences, these limitations ultimately restrain neuroscience research centered on brain function and structure.

The fundamental limitations of the pinhole and unitary detector arrangement were recognized almost 30 years ago. In 1987 and 1988, Bertero [3] and Sheppard [4] both described approaches for pinhole-plane imaging systems where the detector is placed in the pinhole plane allowing researchers to improve spatial resolution and detection efficiency improving on Minsky’s original design. Much more recently, 2013, Sheppard et al. described the extension of so-called pixel-re-assignment technique to fluorescent dyes [5]. Further in 2012, York et al. describe a technique to parallelize pinhole-plane imaging to increase resolution, signal-to-noise, and speed [6, 7]. The first commercial implementation of the pinhole-plane imaging concept was in August 2014 when the Airsycan detector from Carl Zeiss Microscopy was introduced.

The Airyscan detector from Carl Zeiss Microscopy replaces the physical confocal pinhole aperture and unitary detector assembly with a hexagonally packed detector array (Fig. 4.2). Just like the traditional confocal pinhole, the Airyscan detector is positioned in a conjugate focal plane to the excitation spot and utilizes a zoom optic arrangement to project a defined number of Airy unit orders onto the detector to create an optical section. By collecting the additional information of a pinhole-plane image at every excitation scan position, the new Airyscan detector offers substantial and immediate benefit compared to traditional confocal microscopy by offering a much more efficient use of collected fluorescence to increase both the signal-to-noise ratio (SNR) and spatial resolution all while maintaining the optical sectioning ability of a traditional confocal microscope (Fig. 4.2) [8]. More recently in May 2016, the Airyscan detection approach has been extended with illumination shaping to increase acquisition speeds of standard confocal microscopy [9]. The following sections will discuss how the Airyscan detection concept represents the next substantial improvements for the study of neuroscience structure and function.
Fig. 4.2

a Beam path schematic of Airyscan detection principal where the hexagonally packed 32 channel GaAsP PMT Array is positioned in the conjugate focal plane (where the traditional pinhole aperture would reside) and adjustable zoom optics adapt to project 1.25 Airy Units onto the detector. b Schematic of Airyscan detector showing that 1.25 AU are projected onto the detector where each individual detector element represents a 0.2 AU pinhole

4.3 Airyscan Detection for Enhanced Structural Information

For Minsky, the introduction of a pinhole in a conjugate image plane was used to create the optical sectioning capabilities of a scanning microscope by rejecting out of focus light to improve the study and understanding of the underlying circuitry of the central nervous system. However, in addition to improving the optical sectioning capability of a confocal system, closing the pinhole holds the potential additional benefit of increasing the lateral and axial resolution of the resulting image. For modern fluorescence laser scanning confocal microscopy, the pinhole is traditionally set to the standard size of 1 airy unit (1 AU) and already improves the spatial resolution compared to wide-field imaging by a typical factor of 1.06. The resolution can be further improved by closing the pinhole beyond the traditional 1 airy unit size; however, this comes with a drastic reduction of light that can still reach the detector [4] (Fig. 4.3). The full resolution potential, a factor of 2× enhancement, can only be achieved theoretically when the pinhole is almost completely closed (0.2 AU). However, closing the beyond the pinhole beyond the 1AU limit dramatically reduces amount of collected fluorescent signal and in turn decreases the achievable image signal-to-noise resulting in unusable image quality.
Fig. 4.3

Three graphs representing how the pinhole diameter affects the resulting image in confocal microscopy. a Optical slice dependence on pinhole diameter. b Lateral resolution dependence on pinhole diameter. c Light transmission dependence on pinhole diameter

Like all fluorescence imaging techniques, image quality in point-scanning confocal microscopy is directly related to ratio of the amount of signal (i.e., photons) and the amount of noise of an image, the so-called SNR. In point-scanning confocal microscopy, the number of detected photons (signal) are generally extremely small due to labeling densities and/or the optical section created by the pinhole. The resulting statistic variation in the number of detected photons, the so-called shot noise or photon noise, will then become a dominant factor for image contrast; i.e., the available gray levels per image pixel. As shot noise follows Poisson statistics, it will be equal to the square root of the signal. Hence, the SNR will also be proportional to the square root of the signal or synonymous to the square root of the number of photons (N): S/N ~ \( \sqrt N \).

Another source of statistical noise is background fluorescence originating from out of focus signals and auto-fluorescence from the specimen or optical components. Both will limit the contrast of the signal in respect to the background and decrease the image SNR such that insufficient contrast exists for sample signals to be distinguished. The signal-to-background ratio (S/B) can be increased by closing the pinhole and would indeed be greatest if the pinhole aperture would be closed to zero. It is evident that this is not practical as the fluorescence signal would also be excluded from reaching the detector. Therefore, a pinhole aperture size must be chosen to maximize image SNR while retaining an adequate S/B to yield a good image contrast. Traditionally, a pinhole size of 1 airy unit (AU) has traditionally proven to be a good compromise to achieve this goal.

In contrast to a traditional LSM, the novel detector design of the Airyscan combines the resolution benefits of imaging with a small pinhole opening with the collection efficiency of a large pinhole opening (Figs. 4.2 and 4.3). The combination of these two attributes is obtained by projecting 1.25 AU onto the Airyscan detector (via zoom optics) where each detector element behaves as a small 0.2 AU pinhole while the collection efficiency of a 1.25 AU pinhole is maintained. Moreover, since only 1.25 AU is projected onto the Airyscan detector, the optical sectioning ability of an LSM is also maintained. In order to extend the resolution beyond what a 0.2 AU pinhole provides, a linear deconvolution is used, resulting in 2.0× resolution increases in all three spatial dimensions (120 nm x/y, 350 nm z). In addition to the utilization of new optical technologies such as Airyscan to improve the study of brain structure in the neurosciences, new sample preparation techniques are also being utilized to clear brain tissue making them less scattering and transparent (so-called clearing techniques) [10]. Thus, when Airyscan is combined with clearing techniques large increases in resolution, signal-to-noise, and imaging depth are possible and already positively influencing the study of brain structure and circuitry (Figs. 4.4 and 4.5) [11].
Fig. 4.4

Neuromuscular junction in Drosophila brain stained for Bruchpilot (BRP). Resolution comparison between standard confocal (left) and Airsycan (right)

Sample: courtesy of J. Pielage, Friedrich Miescher Institute for Biomedical Research, Basel, Switzerland

Fig. 4.5

Thy1-YFP mouse brain section (100 μ) z-stack max projection showing confocal (right) and Airyscan (left). Arrows in bottom panel highlight the neuron spine neck that is observable not with confocal (left) and is observable with Airyscan (right) when using the same acquisition parameters

4.3.1 Signal-to-Noise and Confocal Microscopy

The geometry of the Airyscan detector provides (and exceeds with deconvolution) the resolution benefits of a closed 0.2 AU pinhole with the light collection efficiency of a 1.25 AU pinhole. In this manner, the resulting image SNR is drastically improved when comparing to traditional confocal imaging with a 0.2 AU pinhole. Further, when using the Airyscan the combination of 32 detector elements and knowing the detector elements positions relative to the optical axis also offers the ability to dramatically improve the resolution and SNR when compared to imaging with the traditional 1 AU pinhole. The improvement in SNR is directly related to two effects: First, the larger overall pinhole size of 1.25 Airy Units collects up to 50% more light than the conventional 1 AU pinhole. Especially in optically denser specimens, the contribution of this light to image formation can be substantial (Fig. 4.6). Secondly, the small individual pinholes not only show an extended resolution beyond a conventional LSM’s resolution limit, but also raise the contrast of higher spatial frequencies collected by confocal microscope system. In other words, the Airyscan detector gives improved contrast without raising noise, which directly translates to a substantial SNR increase in the final image (Fig. 4.7).
Fig. 4.6

Image formation in Airyscan. a Each detector element, numbered in the diagram from 1 to 32, acts as its own pinhole and records a complete image. The images are offset from the center. Groups of detector elements form rings that are indicated from inside to outside with increasingly darker gray shadings. The central ring (1st ring, ~ 0.2 AU) consists of one element (#1), the next outward two rings (second and third rings, ~ 0.6 AU and ~ 1.0 AU) of 6 (#2–7) and 12 (#8–19) elements, respectively, whereas the outer ring (4th ring, ~ 1.25 AU) is formed by 13 elements (#20–32). From inside to outside detector elements or rings collect increasingly lesser light indicated by the progressively darker shading. However, the summed contribution of the outer rings to the total signal can be significant, especially for optically denser samples as shown in the example. b Intensity contribution of the different rings (sum of all respective detector elements of a ring). The numbers give the fraction of the total intensity (all rings summed). The sample is Drosophila melanogaster larval sections stained for the pre-synaptic protein Bruchpilot (Brp) with Alexa 488 and kindly provided by Jan Pielage, Friedrich Miescher Institute (FMI), Basel, Switzerland

Fig. 4.7

Lateral modulation transfer function (MTF) and resolution. a The upper panel shows a confocal (conf.) and a Sheppard sum (Shep.) image of 140 nm Origami sample (GATTAquant, Germany). The lower panel shows the MTF of the imaging setup. The confocal MTF already shows slightly improved contrasts as compared to the excitation MTF alone which means contrast is slightly improved, compared to a wide-field microscope or an LSM with open pinhole. Using the Airyscan detector, the MTF is both raised in intensity and extended to higher frequencies. This is depicted by the Sheppard MTF. The effective “hardware” MTF for Full DCV Airyscan will be the same as the Sheppard MTF for the lateral case. b The same images and MTFs with deconvolution. The target of deconvolution is to raise all contrasts to unity level. Which means, the contrast of structures is not altered, giving an image with optimized resolution as close to the original sample as possible. Image noise is distributed evenly over all spatial frequencies. The noise level limits the frequency range in which the MTF could be raised. Since the “raw” or “hardware” MTF of the Airyscan detector is widely extended over the confocal MTF, the final Airyscan image benefits from deconvolution far more than the deconvolved confocal image. This can be clearly seen in the images above. Plot simulations were done with λex = 488 nm light linear polarized along the x-axis as well as unpolarized light for detection with no Stokes shift; NA = 1.4 and pinhole size of 1.25 AU

The gain in SNR achieved by the Airyscan detector does not require any compromise in regards to speed, resolution, or sensitivity. However, to have the same SNR at its disposal a traditional confocal has to compromise on speed, resolution, or sensitivity (or combinations of those), which is an immediate consequence from the theory of the eternal triangle. If acquisition speed and sensitivity should match that of the Airyscan, the pinhole has to be opened to 1.25 AU, which will result in lower resolution (Fig. 4.3). Likewise, if resolution and speed should equal Airyscan at otherwise identical settings, the pinhole would have to be set to 0.2 AU reducing light level and hence sensitivity (Fig. 4.3). Finally, when resolution and sensitivity should be kept equal, again the pinhole has to be set to 0.2 AU to keep resolution and the sensitivity has to be increased by slower scanning or averaging (Figs. 4.4 and 4.5).

4.3.2 Photon Detection Considerations for Confocal Super-Resolution: Airyscan Versus LSM + DCV

As discussed earlier, the resolution provided by a traditional confocal imaging system can be increased beyond the diffraction limit at the cost of signal and image signal-to-noise by closing the pinhole to 0.2 AU (Fig. 4.3). Practically restricting the pinhole to 0.2 AU for imaging will work only for extreme cases with very bright and stable specimens using high laser powers to generate and detect enough photons. On the other, hand opening the pinhole to 1 or 1.25 AU to match the amount of photons collected by the Airyscan will inadvertently lead to unwanted resolution loss that cannot be recovered in a subsequent DCV. Therefore, the best strategy for traditional LSMs to increase resolution beyond the diffraction limit is to restrain resolution loss by closing the pinhole to a suboptimal value (around 0.6 AU), where the loss in photon detection can be compensated for by slower sampling rates, and then deconvolving the image [12]. At first glance this approach seems promising, since a quick review of theory would suggest imaging a single fluorophore with a traditional confocal system with a pinhole size of 0.6 AU would seem to maintain photon loss through the pinhole at a tolerable level such that the resulting image SNR should not degrade substantially (Fig. 4.8).
Fig. 4.8

Detection efficiencies in confocal imaging. The upper panels in (a) and (b) show the detection efficiencies (DE) in normed units with the pinhole (PH) open (op.), or at 1.25 and 0.60 AU in the case of a single emitter (a) or four emitters (b) in simulations for a stationary beam (no scanning). The lower panels in (a) and (b) show the positions of the emitters in the object plane and the image or detection plane. The cross section of the emitter signal (thick blue line) in the image plane (indicated by the thin blue line) is plotted with intensity (I) in arbitrary units (au) versus distance (d) in micrometer (µm). The extent of the covered signal at pinhole sizes of 1.25 and 0.60 Airy Units (AU) are indicated by the dark and light gray boxes, respectively. (C) Available energy in confocal microscopy. The graph shows simulations of the available energy; i.e., the number of photons, used for imaging in dependence of the pinhole diameter size (in AU) for the case of a single emitter (blue line) and four emitters (red line) in close proximity. Energy was normed to the value at 1 AU of the four emitter case. The energy distribution in case of one single emitter reflects the energy distribution of the Airy disk of a microscopes point spread function (PSF) in the lateral plane. In case of four emitters, the curve is shifted to higher pinhole values until the point of saturation that occurs at around 1.25 AU in both cases. The energies available for 0.6 Airy detection, used in a confocal, and 1.25 Airy detection used routinely in Airyscan are indicated by arrows

For a single fluorophore, the energy loss from closing the pinhole from 1.25 AU to 0.6 AU is approximately 20% (Fig. 4.8). This can be compensated for by 2× averaging. However, since most neuroscience sample conditions have many fluorophores in close proximity to one another, a more informative theoretical example would be to consider the situation with multiple fluorophores. In the case of four fluorophores, the loss in photon detection is dramatically higher amounting to around 65%, which would mean that one has to average 4×–8× in order to match the SNR of a 1.25 AU Airyscan image (Fig. 4.9).
Fig. 4.9

Peripheral nervous system of an embryo of Drosophila melanogaster labeled with Cy5. The left and middle images were taken with a conventional LSM detector combined with DCV processing; the right image was taken with Airyscan detector. For all images acquisition settings were, but the middle image was acquired with four times averaging; exposing the sample to four times the laser dosage and took four times longer

Sample courtesy of Dr. Julia Sellin, AG Hoch, LIMES Institute, Bonn

Hence, the downside to utilizing a pinhole below the 1 AU limit on a traditional confocal system is the dramatic reduction in signal reaching the detector (95% loss at 0.2 AU). Subsequently, if signal reaching the detector is decreased, the resulting image quality will also decrease. This fact negatively impacts the utilization of smaller pinhole diameters in most neuroscience imaging applications as most neuroscience samples and fluorophores cannot supply enough (due to photodamage and/or phototoxicity) fluorescence to yield images with sufficient SNR. As a result, the performance gains of the Airyscan detector afford researchers to study brain structure and circuitry with increased resolution and signal-to-noise all while maintaining the optical sectioning ability of confocal microscopy as initially conceived by Minsky [13].

4.4 Airyscan Detection for Fast Functional Imaging with Increased Resolution and SNR

The confocal microscopy concept was initially developed as a tool to help understand brain structure and how neurons are interconnected. The study of brain structure is still a primary application of traditional confocal imaging systems where the studied samples generally consist of excised tissue/brain specimens that have been fluorescently labeled and fixed on a microscope slide for repetitive study of their underlying structure. However, in addition to the advances in the commercial development of confocal microscope systems over the past decade, molecular biology tools sought to utilize the specificity of fluorescence labeling in combination with microscopy to observe and measure physiological changes as a means to study brain function in real time [14]. More recently, major advances have been made by the utilization of light sensitive proteins (channelrhodopsin and halorhodopsin) that allow the opening and closing of ion channels with light [15]. Subsequently, by combining channelrhodopsin or halorhodopsin with a suitable fluorescence reporter, a researcher will have much more precise control on brain function and response to further neuroscience function research. In order to adequately study and understand brain signaling and function, both the fluorescent reporter as well as the imaging system must respond fast enough to capture rapid and faint electrical signals as they transverse the brains neural circuitry. Thus, as molecular biology tools develop to provide better and better fluorescent reporters, confocal imaging systems have adapted to provide acquisition schemes that allow an increase in image acquisition rates to capture the change in the fluorescence emission of the reporter probes.

The acquisition speed of a conventional laser scanning confocal is determined by how fast a single diffraction-limited laser spot can be moved across a desired field of view with a desired resolution (i.e., pixel count). Therefore, to increase the achievable scan speeds of a conventional LSM, a researcher must decrease the amount of time the excitation laser spends on each pixel (pixel dwell time), reduce the resolution (i.e. pixel count), or reduce the image field of view. As a result, when using a conventional LSM, a researcher must compromise on image signal-to-noise by reducing pixel dwell time, on spatial resolution by reducing the pixel count, or on field of view by limiting structural context by restricting the image field of view by zooming into a portion of the structure of interest. Traditionally, for point-scanning LSMs, reduction of the pixel dwell time has been the preferred approach to maximize scan speeds in the form of resonant scanning.

Resonant scanning uses a sinusoidal scanning approach that varies the pixel dwell time across the field of view in order to achieve a maximum frame rate of around 30 Hz for a full field of view. When the laser spot is scanned faster across the field of view at this speed, the pixel dwell time is shortened considerably and, consequently, the amount of time per pixel spent collecting fluorescence is also decreased, which affects the resulting SNR of the image. As the acquisition speed increases, fewer and fewer photons are available, resulting in deterioration of the image SNR. The outcome is not only a noisy image but also compromised spatial resolution in which fine structures cannot be properly resolved (Fig. 4.10 right panel). To compensate for the deteriorating SNR, the laser power can be increased, but this too has disadvantages: The danger of bleaching the fluorophore and/or damage to live samples from phototoxic effects (e.g., free oxygen radicals) becomes more prevalent at higher laser powers, and thus the risk of influencing experimental outcomes is increased [16, 17, 18]. Therefore, traditional techniques to improve image acquisition speeds demand that a researcher compromise image SNR, resolution, field of view, and laser exposure where the impact of these concessions are likely to impede the research goal.
Fig. 4.10

Fast imaging of microtubules in a contracting cardiac myocyte. After a cardiomyocyte contraction was triggered by a voltage pulse, the entire contraction cycle, which takes only around 250 ms, was recorded using Airyscan in fast mode (left column, 96 FPS) and resonant scanner-based confocal microscope (right column, 80 FPS). Field of view was approximately 2× bigger with the Airyscan (indicated by the gray areas in the right column). Three selected frames out of a 130 recorded images show microtubules before the contraction, in the middle of the contraction, and just after the contraction. Detyrosinated microtubules in cardiomyocytes cells

Images and Samples Courtesy of Ben Prosser, University of Pennsylvania, USA

4.4.1 The Fast Mode for Airyscan

To overcome the trade-off between acquisition speed and image SNR, ZEISS has combined the Airyscan pinhole-plane detection concept with an illumination shaping approach for a new fast acquisition mode (Fig. 4.11). With the fast mode for Airyscan, the excitation spot is elongated along the y-axis to cover four lines in a single scan (Fig. 4.12). Through the use of a parallelization approach, acquisition speeds are increased by a factor of four while high pixel dwell times are maintained, resulting in a high image SNR. Although the imaging speed is notably increased, high SNRs can be preserved, resulting in meaningful data (Figs. 4.10 and 4.13). Moreover, because the fast mode is based on the Airyscan detector, the additional, simultaneous increases in SNR and resolution are retained. Therefore, Airyscan in fast mode offers a researcher the unprecedented combination of a 1.5× increase in resolution and a 4× increase in SNR at four times the acquisition speed (Table 4.1).
Fig. 4.11

(Upper panel) Excitation in laser scanning microscopy. a Standard confocal microscope illuminates the smallest possible spot in the focal plane by filling the back aperture of the objective lens with an expanded laser beam. b If the laser beam does not fill the back aperture, the illuminated spot is bigger and the specimen can be covered with less scans, but suboptimal image resolution. c Airyscan in fast mode uses a slit aperture to fill the back aperture of the objective lens only in one direction (y-axis), and thus extend the illumination spot in the other dimension (x-axis). Specimen can again be imaged with a reduced number of scans, but in combination with an Airyscan detector, this does not negatively affect the image resolution. d. Elements of Airyscan detector unit. Detection spectrum is first selected using a filter-wheel mounted emission filter. Size of the elliptical emission spot is then matched to the size of the Airyscan detector using motorized zoom optics

Fig. 4.12

Confocal volume motion in resonant scanning and Airyscan in fast mode. a Resonant scanner requires long deceleration/acceleration motion (red arrows) to achieve high scan speed required for high imaging frame rates. It therefore scans an area in the specimen that is 60% bigger than the resulting image and spends 60% of the time scanning outside the image. b Airyscan in fast mode spends only 15% of the time decelerating/accelerating and the scanning beam barely leaves the area of the image, during which time it is switched off. Moreover, as four pixels are imaged in parallel, the number of turnaround motions required for an image is reduced fourfold

Fig. 4.13

For example, when imaging calcium signaling, speed is obviously a key issue but imaging at speed over a large field of view gives the experimental data much more context. In order to match the FOV and resolution of provided by Airyscan fast mode (right), the time resolution of a standard confocal microscopy (left) would be reduced to 3.3 FPS. This has an impact on the usefulness of the data as shown in the delta f/f0 plots where the shape of the mean ROI response curves is negatively impacted by not having an enough temporal resolution to match the physiological response. Calcium signaling in Zebrafish expressing GCaMP6

Sample Courtesy of Claire Oldfield Univ of California Berkeley

Table 4.1

Imaging time, dead time, and pixel dwell time for resonant scanner and Airyscan in fast mode


Resonant scanner

Airyscan in fast mode

Image acquisition time (7.5 FPS)

133 ms

133 ms

Imaging time

53 ms

113 ms

Dead time

80 ms

20 ms


Accumulated pixel dwell time (imaging time × parallelization)

53 ms

452 ms

Average pixel dwell time (1024 × 1024 pixel image)

51 ns

431 ns

A 1024 × 1024 image was recorded using a resonant scanner and Airyscan in fast mode at 7.5 frames per second. Due to parallelization and 4× shorter dead time, Airyscan in fast mode achieves almost 9× longer pixel dwell. Resonant scanner was used with 2× averaging (accumulation), otherwise pixel dwell time would be only 25 ns

4.5 Conclusion

The growth and development of the laser scanning microscopy industry over the past 25 years has led to development improvements that have been designed to increase image contrast and instrument versatility. Utilizing the foundation laid by Bertero [3] and Sheppard [4], the Airyscan detector from Carl Zeiss Microscopy replaced the most fundamental part of a traditional laser scanning microscope: the pinhole. The Airyscan’s novel detector design dispenses with the classical physical pinhole and unitary detector assembly and utilizes a new pinhole-plane image detection assembly based on a collection of 32 detection elements. In the new assembly, each of the 32 detector elements acts as its own small pinhole with positional information. The positional information gained by the new approach allows for increased contrast of high spatial frequency information previously not available in traditional confocal systems. The increase in spatial frequency contrast enables the Airyscan to produce images with substantially increased SNR and resolution compared to an LSM acquiring images with a 1 AU pinhole without having to increase laser exposure or sampling. The afforded resolution increase is 2.0× in all three spatial dimensions and the obtained SNR increase is 4×–8× over traditional LSM images acquired with a 1 AU pinhole. Further, as a confocal detector Airyscan benefits from all advantages of confocal microscopy, above all out of focus light reduction. Hence, the Airyscan detection concept outperforms the traditional pinhole and unitary detector setup in every imaging metric for the study of neuroscience samples for both the study of structure and function.


  1. 1.
    J.-A. Conchello, J.W. Lichtman, Optical sectioning microscopy. Nat. Methods 2(12), 920–931 (2005)CrossRefGoogle Scholar
  2. 2.
    M. Minsky, Memoir on inventing the confocal scanning microscope. Scanning 10(4), 128–138 (1988)MathSciNetCrossRefGoogle Scholar
  3. 3.
    M. Bertero, P. Brianzi, E.R. Pike, Super-resolution in confocal scanning microscopy. Inverse Prob. 3, 195–212 (1987)MathSciNetCrossRefGoogle Scholar
  4. 4.
    C.J. Sheppard, Super-resolution in confocal imaging. Optik 80(2), 53–54 (1988)Google Scholar
  5. 5.
    C.J. Sheppard, S.B. Mehta, R. Heintzmann, Superresolution by image scanning microscopy using pixel reassignment. Opt. Lett. 38(15), 2889–2892 (2013)CrossRefGoogle Scholar
  6. 6.
    A.G. York et al., Resolution doubling in live, multicellular organisms via multifocal structured illumination microscopy. Nat. Methods 9(7), 749–754 (2012)CrossRefGoogle Scholar
  7. 7.
    A.G. York et al., Instant super-resolution imaging in live cells and embryos via analog image processing. Nat. Methods 10(11), 1122–1126 (2013)CrossRefGoogle Scholar
  8. 8.
    T.R. Neu, J.R. Lawrence, Innovative techniques, sensors, and approaches for imaging biofilms at different scales. Trends Microbiol. 23(4), 233–242 (2015)CrossRefGoogle Scholar
  9. 9.
    J. Huff, The Fast mode for ZEISS LSM 880 with Airyscan: high-speed confocal imaging with super-resolution and improved signal-to-noise ratio. Nat. Methods 13, 1–2 (2016)CrossRefGoogle Scholar
  10. 10.
    D. Richardson, J. Lichman, Clarifying tissue clearing. Cell 162, (2015)CrossRefGoogle Scholar
  11. 11.
    M. Ke et al., Super-resolution mapping of neuronal circuitry with an index-optimized clearing agent. Cell Rep. 14(11), 2718–2732 (2016)CrossRefGoogle Scholar
  12. 12.
    J. Huff et al., ZEISS Technology Note: The Airyscan Detector from ZEISS. Confocal Imaging with Improved Signal-to-noise Ratio and Superresolution (Carl Zeiss Microscopy, Oberkochen, 2015)Google Scholar
  13. 13.
    E. Suberbielle et al., DNA repair factor BRCA1 depletion occurs in Alzheimer brains and impairs cognitive function in mice. Nat. Commun. 6(8897), 33 (2015)Google Scholar
  14. 14.
    M. Lin el al., Genetically encoded indicators of neuronal activity. Nat. Neurosci. 19, 1142–1153 (2016)Google Scholar
  15. 15.
    S. Mohanty, V. Lakshminarayananan, Optical techniques in optogenetics. J. Mod. Opt. 62(12), 949–970CrossRefGoogle Scholar
  16. 16.
    S. Wäldchen et al., Light-induced cell damage in live-cell super-resolution microscopy. Sci. Rep. 5, 15348 (2015)CrossRefGoogle Scholar
  17. 17.
    D. Li et al., ADVANCED IMAGING. Extended-resolution structured illumination imaging of endocytic and cytoskeletal dynamics. Science 349, aab3500 (2015)Google Scholar
  18. 18.
    G. Kucsko et al., Nanometre-scale thermometry in a living cell. Nature 500, 54–58 (2013)CrossRefGoogle Scholar

Copyright information

© Springer Nature Singapore Pte Ltd. 2019

Authors and Affiliations

  1. 1.Carl Zeiss Microscopy, LLCThornwoodUSA

Personalised recommendations