Meeting Banner
Abstract #4786

Segmentation methods and lesion masking influence spatial normalization and the detection of task-related activity. On behalf of he HERMES Study Group

Marie-Charlotte Plichta 1 , Agnes Aghetti 2 , Thomas A Zeffiro 3 , Laurent Lamalle 4,5 , Olivier Detante 6,7 , Marc Hommel 8 , and Assia JAILLARD 5,9

1 CHU Grenoble, Grenoble, France, 2 Centre Hospitalo Universitaire, Grenoble, France, 3 Neural Systems Group, Massachusetts General Hospital, Boston, Massachusetts, United States, 4 Unit IRM 3T - Recherche, Centre Hospitalo Universitaire, Grenoble, France, 5 IRMaGe Inserm US 17 / CNRS UMS 3552 RMN biomdicale et Neurosciences, SFR de l'Universit Joseph Fourier, Grenoble, France, 6 Stroke Unit, Centre Hospitalo Universitaire, Grenoble, France, 7 Equipe 5 : Neuroimagerie Fonctionnelle et Perfusion crbrale, Grenoble Institut des Neurosciences (GIN) Inserm U 836, Grenoble, France, 8 CIC Ple Recherche, Centre Hospitalo Universitaire, Grenoble, France, 9 Unit IRM 3T Recherche. Ple Recherche., Centre Hospitalo Universitaire, Grenoble, France

We studied the effects of segmentation methods (SPM8) and lesion masking on the accuracy of MRI spatial normalization of brains and fMRI tactile task-related activation in large stroke lesions. The normalized T1-images were altered after Segment, with brain tissue outside of the MNI template. New-Segment method led to satisfactory T1-normalized images. There was no mask effect. We observed contralesional somatosensory activity using Segment and ipsilesional somatosensory activity with New-Segment. There was no mask effect with Segment, but higher activity with New-Segment. We showed that T1-image segmentation method influences spatial normalization, affecting the detection of task-related activity at the group level.

This abstract and the presentation materials are available to members only; a login is required.

Join Here