By Professor Kishan Dholakia
Seminar date: 24 March 2021
Prof. Kishan Dholakia, from the University of St Andrews, runs a vibrant, international group exploring advanced optical beam shaping for biomedicine, precision measurement and optical manipulation. He has hundreds of refereed papers including over twenty in Nature and Science titled journals. He is the recipient of the R.W. Wood Prize of the Optical Society (2016), the Institute of Physics Thomas Young Medal and Prize (2017) and the SPIE Dennis Gabor Award (2018).
The 4 key points we walked away with
- Optical Imaging
Optical imaging holds a fundamental role in the medical sphere contributing to 64% of all medical imaging techniques. There are two distinct types of optical imaging with different challenges, lateral imaging, where the super-resolution can be achieved, and the depth imaging domain. Scientists are trying to merge these two domains to create an imaging technique that can have full 3D resolution and allow to see, e.g. the depth of a whole brain in super-resolution.
- Point Scanning in Optical Imaging
Point-by-point scanning uses high numerical aperture (NA) optics to focus a laser beam on a sample and move the beam around. This process can use one, two or three photon processes. During one photon microscopy, only one high-energy photon is used to excite the target to illuminate. This causes the excitation to happen everywhere across the illuminating beam aperture. During two photon microscopy, two low-energy photons are needed to simultaneously excite the target. This process has an intensity threshold, and excitation only occurs at highest photon flux, thus making it possible to have the excited area much smaller than the illuminating beam diameter.
- Light Sheet Fluorescent Microscopy
In contrast to point scanning a light sheet, e.g. produced by a cylindrical lens, and imaging emissions orthogonally (at ninety degrees) holds value. Through this method, there is the capability to illuminate part of a sample, a layer, so as to obtain particular information about a specific area. By using a broad beam with a gentle focus, a wide view of the field is obtained, however there is low resolution. A narrow, tightly focused beam delivers high resolution though the penalty is a narrow field of view.
By using Bessel beams with two-photon (or three photon) excitation, one is able to achieve high resolution yet obtain a wide field of view. Through multiphoton excitation, one can detect nonlinear signals and filter out the source beam. Bessel beams are particularly good at illuminating the sample at depth and creating nonlinear signals. The ring structure is needed for the propagation of the beam. the outer rings have much lower power and produce lower nonlinear signals compared to the central maximum, this ensures the desired area is being detected with minimal noise.
- Airy Beams
Airy beams are a diffractionless solution of the Schrödinger equations and a different approach to the Bessel beam. When used in fluorescent microscopy, each lobe of an Airy beam contributes to the generated and consequently detected signal, and for the recreation of the image, a deconvolution is needed. The Airy beams can be used to ensure high axial resolution over an extended field of view without increasing the sample exposure or the number of images requires. In this situation, Airy beams are superior to the Bessel beams for one photon light sheet.
The work being done
Kishan and his group have realised both attenuation-compensated Bessel beams and Airy beams, as well as using this attenuation-compensated Airy light sheet microscopy in a mouse brain section. This showed great signal to noise ratio improvements of the medical imaging between 20% and 150%. They have also realised three-photon light sheet microscopy using Bessel beams. In this three-photon method, they had increased penetration depth at high resolution and better signal to noise ratio compared to two photon methods.
What does this mean for the future?
Light sheet microscopy is developing towards obtaining super-resolution of an extended field of view with low photo-toxicity. This would be extremely useful for medical imaging. Scientists will continue to work with varying techniques, such as three-photon microscopy and machine learning in order to continue improving these activities.
Questions from the audience
A (Kishan Dholakia): No. We’ve done studies of the modulational optical transfer function for various beam shapes and the Bessel beam wins. We compared them side by side. For single photon, the Airy seems to win. For two and three photons, the Bessel wins.
Q (Dragomir Neshev): My question is regarding machine learning. You effectively collect these images and put them into a computer, and the computer uses a neuron-network to reconstruct the images. Do you think this is possible to do all optically without the computer?
A (Kishan Dholakia): I guess I don’t see any reasons why one couldn’t. What advantages that might give you, at the moment I don’t know, but in principle, it can be done all optically and optically computed. It is an outstanding challenge for microscopy to do that. But if you see an advantage, yes, I think there are no barriers to seeing it done all optically.
Q (Steve Lee): In the acoustic trap, with the light sheet coming down, is there any aberration in the imaging?
A (Kishan Dholakia): Yes, there would be aberration due to refractive index mismatches. We didn’t take that into account in much detail, because it was just low-res images we wanted to take for our study. But you would have to, and you can design some kind of more appropriate chambers to do that.
Further reading and enquiries
For more information, please refer to Prof. Kishan Dholakia’s research group website: https://opticalmanipulationgroup.wp.st-andrews.ac.uk/