eprintid: 17116 rev_number: 16 eprint_status: archive userid: 1256 dir: disk0/00/01/71/16 datestamp: 2014-07-10 07:54:07 lastmod: 2014-08-21 10:27:02 status_changed: 2014-07-10 07:54:07 type: doctoralThesis metadata_visibility: show creators_name: Goch, Caspar Jonas title: Expanding Graph Theoretical Indices to Include Medical Knowledge - An Assessment of Classification Accuracy in the Case of Autism Spectrum Disorders subjects: ddc-004 subjects: ddc-530 subjects: ddc-610 divisions: i-130001 divisions: i-850300 adv_faculty: af-13 keywords: Network analysis cterms_swd: Medical Imaging abstract: Using graph theory to analyse the architecture of the human brain, the connectome, has gained increasing interest in the last decade. In this work we extend graph measures, which have previously been developed for other applications, to include prior medical information. These extended measures are then evaluated on a collective of Autism Spectrum Disorder patients. Examining their performance in comparison to other traditionally used measures we show, that they are a valuable new tool in the analysis of the human connectome. We then further evaluate their performance over a range of network densities in order to determine the range at which they supply the most valuable information. By doing an in depth evaluation of these measures we aim to reduce the amount of guesswork in choosing variables in the analysis of the connectome and help to improve the comparability of different studies. date: 2014 id_scheme: DOI id_number: 10.11588/heidok.00017116 ppn_swb: 1658818547 own_urn: urn:nbn:de:bsz:16-heidok-171166 date_accepted: 2014-07-02 advisor: HASH(0x558eaa7a5778) language: eng bibsort: GOCHCASPAREXPANDINGG2014 full_text_status: public citation: Goch, Caspar Jonas (2014) Expanding Graph Theoretical Indices to Include Medical Knowledge - An Assessment of Classification Accuracy in the Case of Autism Spectrum Disorders. [Dissertation] document_url: https://archiv.ub.uni-heidelberg.de/volltextserver/17116/1/Dissertation_Goch.pdf