Skip to Main content Skip to Navigation

Modélisation du problème direct de la magnéto et électroencéphalographie : méthodes numériques et calibration.

Sylvain Vallaghé 1
1 ODYSSEE - Computer and biological vision
DI-ENS - Département d'informatique de l'École normale supérieure, CRISAM - Inria Sophia Antipolis - Méditerranée , ENS Paris - École normale supérieure - Paris, Inria Paris-Rocquencourt, ENPC - École des Ponts ParisTech
Abstract : This thesis focuses on the forward problem of electroencephalography (EEG) and magnetoencephalography (MEG). The first part deals with the calculation of the forward problem solution, governed by a partial differential equation. We present a new finite element method (FEM) based on a regular hexahedral mesh and implicit descriptions of the domain, which allows to solve the forward problem in realistic geometries with a low computational cost. We add to this method some general reciprocal equations, derived by the adjoint method, in aim to efficiently compute the lead field of all kinds of EEG and MEG sensors. The second part is concerned with the choice of the electrical conductivities in the EEG head models. We first perform a global sensitivity analysis of the EEG topographies with respect to the conductivities for some classical head models with three or four layers. Following the results of this analysis, we then propose a method for conductivity calibration using somatosensory evoked potentials.
Complete list of metadatas

Cited literature [117 references]  Display  Hide  Download
Contributor : Vallaghe Sylvain <>
Submitted on : Tuesday, March 2, 2010 - 1:00:21 PM
Last modification on : Thursday, October 29, 2020 - 3:01:48 PM
Long-term archiving on: : Friday, June 18, 2010 - 8:32:49 PM


  • HAL Id : tel-00460772, version 1



Sylvain Vallaghé. Modélisation du problème direct de la magnéto et électroencéphalographie : méthodes numériques et calibration.. Mathématiques [math]. Université Nice Sophia Antipolis, 2008. Français. ⟨tel-00460772⟩



Record views


Files downloads