Time Series Analysis of Optical Topography Data

Functional imaging is the study of changes in oxygenation locally in the brain in response to certain stimuli such as visual or sensory tasks. It is a large research area which is targeted at the general area of brain mapping. One new method is using optical signals which have the advantage of being cheap as well as very fast (in comparison for example to Magnetic Resonance Imaging).

One group who have developed a system for this approach is Hitachi. They have provided us with sample data from this system with a view to improving their analysis.

Up to now, data is simply presented as a 2D map rather than being reconstructed onto the cerebral cortex. We will use simple spherical shell models as a simplified model of the head, and investigate the use of Kalman filters and other time-series analysis methods to improve the reconstruction of the images, taking into account the time-varying nature of the signal. The work will therefore build on existing expertise at UCL in non-linear image reconstruction for static images.

The project will suit someone interested in numerical methods, signal processing and medical imaging.

Click here for Simon's Home page.