Data mining brain regions with neuroimaging databases

Helgi Mar Sigurdsson

AbstractThis thesis presents methods of automatic meta-analysis of neuroimaging articles that report brain region coordinates. Two lists were gathered, a list of brain regions and a list of cognitive terms. 1.646 articles were downloaded and the coordinates in them were linked to the brain regions in the list that had been gathered, which made the articles connected to the brain regions as well. For each brain region the abstracts of all the articles that had been linked to it were data mined with the list of cognitive terms. Non-negative matrix factorization was then used to discover topics in each of those brain regions. Each topic had a certain set of articles linked to it, and thus also a certain set of coordinates. The topics in each brain region were then compared with a statistical test to see if the distributions of their respective coordinates were similar or not. If the statistical test showed that the distributions were dissimilar that could mean that two topics within the same brain region were functionally segregated. The results were then compared to results from previous studies of chosen brain regions.
TypeMaster's thesis [Academic thesis]
PublisherTechnical University of Denmark, DTU Informatics, E-mail:
AddressAsmussens Alle, Building 305, DK-2800 Kgs. Lyngby, Denmark
NoteSupervised by Associate Professor Finn Årup Nielsen,, and Professor Lars Kai Hansen,, DTU Informatics
Electronic version(s)[pdf]
Publication link
BibTeX data [bibtex]
IMM Group(s)Intelligent Signal Processing

Back  ::  IMM Publications