@MASTERSTHESIS\{IMM2013-06647,
    author       = "M. Agn",
    title        = "Surface-based mapping of the serotonin transporter binding in cerebral cortex - filtering and modeling of {PET}/{MRI} data",
    year         = "2013",
    school       = "Technical University of Denmark, Department of Applied Mathematics and Computer Science / {DTU} Co",
    address      = "Matematiktorvet, Building 303B, {DK-}2800 Kgs. Lyngby, Denmark, compute@compute.dtu.dk",
    type         = "",
    note         = "{DTU} supervisor: Koen Van Leemput, Ph.D., kvle@dtu.dk, {DTU} Compute",
    url          = "http://www.compute.dtu.dk/English.aspx",
    abstract     = "The aim of this project was to improve the filtering and modeling of {PET} data, to better handle the high noise level present in such data. It should lessen the artifacts of conventional volume-based spatial filtering with a Gaussian kernel. The focus was on the cerebral cortex due to its highly-folded and thin structure, which makes it particularly unsuited for the conventional approach to filtering.
A surface-based approach was developed, which took the highly folded intrinsic geometry of cerebral cortex into account in the filtering and modeling by the multilinear reference tissue method MRTM2. A surface representation of the cerebral cortex was obtained by the software package FreeSurfer. By smoothing across the surface of the cortical layer the data was less affected by edge artifacts, as the smoothing is done between more functionally connected regions with similar neuronal density. The approach was contrasted with the conventional volumebased approach with good results. In addition, the surface-based statistical tools of FreeSurfer for group analysis between brains was evaluated with the use of this model.
Furthermore, a Bayesian framework was used to directly incorporate the data filtering into the mathematical model. This model was based on MRTM2 and assumed that close regions have similar parameters and regularized the data on this assumption. This model was shown to handle high levels of noise better than the ordinary surface-based approach, while at the same time retaining a higher resolution and detail. In addition, it resulted in a higher repeatability between scans on a vertex level in a test-retest setting. Furthermore, attempts were made to treat the data in a fully Bayesian approach by including optimization of the hyperparameters of the model."
}