Use this URL to cite or link to this record in EThOS:
Title: Laser scanning confocal microscope with direct wavefront sensing of volumetric backscattered light
Author: Correa Poblete, Noemi Patricia
ISNI:       0000 0004 6061 6106
Awarding Body: Imperial College London
Current Institution: Imperial College London
Date of Award: 2016
Availability of Full Text:
Access from EThOS:
Access from Institution:
Fluorescent labelling combined with confocal microscopy is a powerful tool employed in several laboratories around the world. This is based in the ability to image well only detail that arises from the region of the specimen close to the focal plane. In confocal microscopes, a pinhole situated in front of the detector leads to optical sectioning, at the cost of rejecting some signal photons together with the out-of-focus ones. The problem in this case is that photons generated in the focal volume are susceptible to scattering, changing their direction and not passing through the pinhole, thus losing information. Natural specimen composition, such as structures sizes and their refractive indices will degrade the intensity and the shape of the focus, but they also affect the imaging of the generated fluorescence onto the pinhole. This loss becomes more significant when we want to image deeper in the specimen. In this work a laser scanning confocal microscope with direct wavefront sensing of volumetric backscattered light was developed and built. It is shown for a couple of specimens, that the signal levels can be improved by correcting in real time the aberrations introduced by these samples at different depths. The advantage of this method relies on the fact that wavefront distortions are sensed by backscattered light instead of fluorescent light from the sample. Problems such as photo-bleaching and photo-toxicity in the specimen can be minimized with this approach. The problem associated with centroid estimation position when out-of-focus light forms part of the light gathered by the sensor was addressed. A centroid algorithm capable of rejecting this signal in order to get an accurate and meaningful centroid detection is proposed. The hybrid centroid algorithm that we propose is based on the optimisation of the product between the data and a spot model, was compared with one of the traditional methods employed with Shack-Hartmann sensors. Computer generated and also experimental data obtained from the system that we built was employed to test the centroid algorithm. Good centroid estimation values were obtained in both cases. Additionally, to improve the system, the implementation of an optimal reconstructor is ap- proached. The problem associated with the lack of prior knowledge for biological and non-biological samples to be used for wavefront reconstruction is addressed. This was done by generating different wavefront statistics as input wavefronts, and reconstructing these by using the same or different priors, at different signal-to-noise ratios. From the results, it was possible to find a range of values where the wavefront reconstruction error was small and gave reasonable error values along the whole range of possible input wavefront. Finally, successful wavefront corrections using samples made of leaves in agarose, leaves in agarose with glucose and cell spheroids were obtained at different depths. Even though improve- ments in image resolution were negligible, we did obtain an increase in the intensity of the confocal images that we recorded. Also, smaller values for the wavefront Zernike coefficients and root-mean- square were obtained, demonstrating that the system is able to perform wavefront corrections using just backscattered light.
Supervisor: Paterson, Carl ; Dunsby, Chris Sponsor: Government of Chile
Qualification Name: Thesis (Ph.D.) Qualification Level: Doctoral