Summer Project Proposal

Visible light and UV-fluorescence microscopy offer many advantages for examining biological samples and are very widely used. The spatial resolution, however, is usually bounded by the diffraction limit at too high a level to resolve many important details of the cell. Structures such as the dense core vesicles that package peptides for release by neuroendocrine cells are just a bit too small to image their activities in detail.

Structured illumination microscopy is a relatively new variant that attempts to get around this constraint by imposing a signal structure on the incident light and then integrating multiple images of the subject taken with slightly differing signals. The subject's interactions with the information encoded in the illumination provide clues from which greater spatial resolution can be derived.

The technique has been successfully applied to improve both lateral resolution and depth sectioning, and some commercial implementations are available. However, the spatial resolution comes at a significant cost in time. Current systems are slow, undermining what ought to be a key advantage of observing under physiological conditions: that the specimen can be alive, and thus in motion. Since the process involves the use of multiple images over time, its ability to handle moving targets is constrained by the speed at which those images can be captured.

In this project, we aim to investigate whether the cumbersome method currently used for repositioning the light source between frames -- physically moving a glass element in the illumination path -- could be replaced by using a high-resolution liquid crystal on silicon (LCOS) device of the kind found in data projection equipment. This should also provide greater flexibility for the projected pattern than the present fixed optical gratings permit, which may open up unexplored options for resolution improvement.

The initial stages of the project will investigate the optical and computational requirements of current structured illumination techniques, and compare these with the properties of available LCOS components. If the requirements can be met, we will then define the basic processes and architecture for an LCOS-based microscope, in terms of both acquiring the input frames and the computation required to reconstruct a higher-resolution image. If, on the other hand, the technology is not viable, we will identify the problem areas and attempt to suggest future avenues of development that might change this.

If time allows, we may conduct some experiments with actual LCOS optics to support the analysis; we may also investigate more sophisticated approaches such as using doubly-periodic illumination patterns.

Given the short duration of the project, no attempt to construct the microscope is likely. Rather, the goal is to provide a foundation of modelling and analysis on which future development efforts can build.

As the project is primarily theoretical, there should not be any substantial resource requirements other than supervisors' time and existing computational resources, although some additional software libraries may be needed.