Skip to Main content Skip to Navigation

Analyse spatiale de l'activité électrique cérébrale : nouveaux développements

Laurent Spinelli 1
1 TIMC-GMCAO - Gestes Medico-chirurgicaux Assistés par Ordinateur
TIMC - Techniques de l'Ingénierie Médicale et de la Complexité - Informatique, Mathématiques et Applications, Grenoble - UMR 5525
Abstract : A large number of factors can affect the estimation of the brain electrical activity based on EEG/MEG measurements. We have first investigated the effect of the number and configuration of the electrodes. This strongly influences the efficiency of the electrical source localisation. Using the properties of the resolution matrix, it is shown that increasing the number of electrodes improves the quality of the localisation in terms of dipole localisation error. However, we have found that for more than 100 electrodes, the goodness of the solution has an asymptotic behaviour and does not improve. Even more, the electrodes should cover uniformly the whole solution space to minimise the localisation errors. Secondly we have studied the effect of interpolation on the localisation of the P100 of visual evoked potentials, using three different algorithm, nearest neighbours, thin plane spline and spherical spline. Our results show that the nearest neighbours type algorithms have the greatest localisation errors. In contrast, the spherical spline minimises these errors. All these studies have been done while simplifying the head as a 3 shell spherical model, which means that no anatomical constraint could be performed. Using high resolution MRI, the BEM and FEM algorithm can handle anatomical constraints. These algorithms are increasingly used in the literature, but they are difficult to handle. In the third part we propose a new approach combining anatomical constraints and spherical head model. The surface of the scalp can be easily extracted from MRI data. This surface can be expressed with a spline function. Then this surface is projected on a best fitting sphere. The MRI volume is now a sphere in which we can define solution points depending on the brain tissues. In a first step, we have validated this new approach performing simulation localising cortical and mesial sources. In addition, we compared the localisation of seizure onset in the frequency domain and using minimum norm type algorithm with intracranial data obtained from an epileptic patient. The results show a good correlation between both localisations.
Document type :
Complete list of metadatas
Contributor : Thèses Imag <>
Submitted on : Wednesday, February 18, 2004 - 4:47:13 PM
Last modification on : Tuesday, November 24, 2020 - 4:18:04 PM
Long-term archiving on: : Wednesday, September 12, 2012 - 1:35:29 PM


  • HAL Id : tel-00004851, version 1



Laurent Spinelli. Analyse spatiale de l'activité électrique cérébrale : nouveaux développements. Modélisation et simulation. Université Joseph-Fourier - Grenoble I, 1999. Français. ⟨tel-00004851⟩



Record views


Files downloads