Author's response

Reto Fiolka, Bingying Chen, Tonmoy Chakraborty, Stephan Daetwyler, James Manton, and Kevin Dean shared

Open questions  

1.The main advantage of your work is rapid volumetric imaging. In your discussion, you mention as an example, that the increased depth of focus could be helpful in scenarios of significant sample motion. What is the maximum imaging speed you could achieve, and is this enough to visualize most phenomena occurring in vivo- for instance blood circulation?

We have not applied this technology to very rapid raster scanning microscopy, as our microscope does not possess resonant galvanometric scanners. However, in theory, the EDF is compatible with resonant scanning, in particular also with recently developed Lissajous scanning in two dimensions. Thereby, we anticipate frame rates as fast as 100-1000 Hz are feasible, provided the sample is bright enough.

2. You discuss also that your method is advantageous when multiple objects at different depths must be monitored simultaneously. What range of fluorophores/wavelengths can you visualize simultaneously?

Currently, we employ a Ti:Sapph laser in our microscope with a tuning range from 650-1050 nm, and we typically use 900 nm illumination to excite GFP. Unfortunately, tuning the laser is quite slow, so if we want to image a second channel, we typically use a fluorescent protein that can be excited at the same wavelength as GFP, but emits at a wavelength that can be separated with dichroics and emission filters.  Two great examples include CyOFP and LSS-mOrange.

If tuning the laser to a different excitation wavelength is not a problem, e.g. if you have two ultrafast lasers (each emitting at a different wavelength), or if you have an optical parametric oscillator, then in principle, you can use any fluorophore so long as it possesses a good two-photon absorption cross-section.  This is expensive, but feasible.

3. What were your findings in terms of phototoxicity and photobleaching compared with conventional two-photon imaging?

This is an interesting question that we have not yet studied. There are conflicting reports in the literature about the effects of phototoxicity and photobleaching, and their relation to peak power and the laser repetition-rate.  In general, it is our suspicion that decreasing the instantaneous laser power and integrating over a larger volume will decrease photobleaching and phototoxicity.  However, it is not clear if this effect would be measureable in our setup.

4. You mention in your discussion that existing two-photon raster scanning microscopes can be easily retrofitted without adding any other optical components than the mask. A big advantage of some commercially available microscopes is the possibility to easily switch modalities. Can the microscope be used in either mode, or how easy is it to switch?

To switch between the modes in an automated fashion, most likely a mechanical actuator could be used to swing the phase mask in and out of the optical train of the microscope.

5.Can you expand further why Bessel beams are robust to aberrations and optical occlusions, and how the EDF beam design can achieve this?

The assumption is that a Bessel beam is self-healing, as there is always new light entering the focal region from the periphery. Our EDF beam can be interpreted as a superposition of multiple Bessel beams, so the assumption is that each would inherit some of these self-healing properties. However, in what regimes this self-healing property really works is contested for Bessel beams themselves. For example, if you aberrate the plane wave that precedes the Bessel beam with strong aberrations, the Bessel beam itself will be severely aberrated. Also, it does not appear that a Bessel beam would be immune to strong light-scattering in tissues.  Thus, we believe that in the presence of strong light-scattering and aberrations, our focus will be as severely affected as normal laser foci.

6.How do the images acquired using your setup compare in terms of resolution to conventional confocal microscopy and conventional two-photon microscopy?

The most obvious difference is the increased depth of focus – which will cause more objects to appear sharply. However, there is a slight lateral deterioration of image resolution, theoretically 10%. So, a conventional two-photon volume will be slightly sharper laterally. Going to confocal microscopy, this will depend on the sample. A cell on a coverslip will likely be sharper imaged with a confocal. However, deeper inside a tissue, I expect that the EDF two-photon image will appear sharper. We expect that this cross-over point happens at 1-2 scattering mean free paths.