Abstract

There exist several ways to study the brain in vivo (within a living organism). These include Magnetic Resonance Imaging (MRI), functional Magnetic Resonance Imaging (fMRI) and Positron Emission Tomography (PET). With the PET method, it is possible to see where different neuroreceptors are located.
Unfortunately, the resolution of the PET camera is very low, which makes it difficult to study small regions of interest in the brain. There are several methods to enhance the image quality, one of them is to coregister the image against an MR image, which can be thought of as laying the images over each other and trying to make them overlap as well as possible. When this is done, it is possible to draw regions of interest on the MR image, and then use them to see how much activity it is in the same region in the PET image.
At the Karolinska Hospital the process of drawing regions of interest is done manually, a very time consuming process which is suspectible to possible bias from the person drawing the regions. The work presented in this thesis will concentrate on evaluating different methods that can be used for the coregistration process, how different interpolation schemes affect the resulting data and compare available programs for this. Especially different interpolation schemes are looked upon, with the surprising result that the most commonly used, trilinear, is the worst performer.

Specification

Full specification in swedish

Final report

Pictures

Contact