Planning functional neurosurgical procedures to minimize the effects of movement disorders requires the accurate localization of small subcortical structures for the creation of lesions (in the thalamus or the globus pallidus) or the implantation of deep brain stimulating electrodes (in structures such as the thalamus, globus pallidus, and the subthalamic nucleus). Standard computed tomography (CT) and magnetic resonance imaging techniques (MRI) have limited signal, resolution, and contrast in subcortical regions, thereby complicating target localization. Modern imaging and image processing techniques have demonstrated that targets are better identified using digital atlases which can be customized to each patients unique anatomy. In this thesis, multi-modal techniques for targeting subcortical nuclei in neurosurgical interventions are presented. Chapter 3 describes the development of a new atlas of the subcortical anatomy developed from a set of manually segmented high-resolution serial histological data using novel reconstruction techniques. The atlas was warped to fit a high-resolution, high-contrast MRI template and all versions of the atlas can be viewed in register to allow enhanced visualization of the subcortical anatomy and contains 105 structures. To accurately target subcortical structures on pre-operative MRI data, the atlas must be customized to each patient. Chapter 4 contains a validation of warping techniques for atlas-to-patient warping. An overlap metric is used to compare the atlas definition of three subcortical structures with a silver standard derived from the agreement of manual segmentations of five manual raters. A functional validation was also performed to compare the coordinates locations of positive intra-operative stimulations and the atlas definition of the sensory thalamus. The manual segmentations of the clinical data were also used to compare two linear, two piecewise linear, and four nonlinear registration based atlas warping / Le planning des procedures fonctionnelles neurochirurgicales pour minimiser les troubles du mouvement demande une localisation précise des petites structures subcorticales pour la creation de lesions (dans le thalamus ou le globus pallidus) ou pour l'implementation d'electrodes profondes (dans des structures comme le thalamus, le globus pallidus ou le noyau subthalamique). Le scanner a rayons X (CT) ou l'imagerie par resonance magnétique (IRM) ont un signal, une resolution et un contraste limites dans les regions subcorticales, compliquant ainsi la localisation de la cible. Les techniques modernes d'imagerie et de traitement d'images ont demontre que les cibles sont mieux identifiees en utilisant des atlas digitaux ajuste a l'anatomie de chaque patient. Dans cette these, sont presentees des techniques multi-modales pour cibler les noyaux subcorticaux dans les interventions neurochirurgicales. Le chapitre 3 décrit la creation d'un nouvel atlas de l'anatomie subcorticale, développe a partir d'une serie de coupes histologiques consecutives, utilisant des techniques de reconstruction innovantes. Cet atlas a ete deforme pour s'ajuster a un modele IRM de haute resolution et de fort contraste. Toutes les versions de l'atlas peuvent etre recale pour permettre une visualisation amelioree de l'anatomie subcorticale contenant 105 structures. Pour cibler precisement les structures subcorticales sur les donnees IRM preopreratoires, l'atlas doit etre s'ajuster a la conformation individuelle du cerveau du patient. Le chapitre 4 contient une validation des techniques de déformations pour les déformations allant d'un atlas vers l'anatomie du patient. Une mesure de recouvrement est utilisee pour comparer trois structures subcorticales definies par l'atlas et des segmentations manuelles de ces structures utilises comme reference (silver standard) et definiees par consensus entre cinq opérateurs. Une validation fonctionnelle a egalement ete realisee afin de comparer le pos
| Identifer | oai:union.ndltd.org:LACETR/oai:collectionscanada.gc.ca:QMM.18755 |
| Date | January 2008 |
| Creators | Chakravarty, Megha |
| Contributors | Louis Collins (Internal/Supervisor) |
| Publisher | McGill University |
| Source Sets | Library and Archives Canada ETDs Repository / Centre d'archives des thèses électroniques de Bibliothèque et Archives Canada |
| Language | English |
| Detected Language | English |
| Type | Electronic Thesis or Dissertation |
| Format | application/pdf |
| Coverage | Doctor of Philosophy (Department of Biomedical Engineering) |
| Rights | All items in eScholarship@McGill are protected by copyright with all rights reserved unless otherwise indicated. |
| Relation | Electronically-submitted theses. |
Page generated in 0.0026 seconds