L Garnero

Publications of L Garnero

• Automatic Segmentation of the Hippocampus and the Amygdala Driven by Hybrid Constraints: Method and Validation.

  Authors: M. Chupin, A Hammers, R S N Liu, O. Colliot, J Burdett, E. BARDINET, J S Duncan, L Garnero, L Lemieux

  NeuroImage. 03/2009;

  The segmentation from MRI of macroscopically ill-defined and highly variable structures, such as the hippocampus (Hc) and the amygdala (Am), requires the use of specific constraints. Here, weThe segmentation from MRI of macroscopically ill-defined and highly variable structures, such as the hippocampus (Hc) and the amygdala (Am), requires the use of specific constraints. Here, we describe and evaluate a fast fully automatic hybrid segmentation that uses knowledge derived from probabilistic atlases and anatomical landmarks, adapted from a semi-automatic method. The algorithm was designed at the outset for application on images from healthy subjects and patients with hippocampal sclerosis. Probabilistic atlases were built from 16 healthy subjects, registered using SPM5. Local mismatch in the atlas registration step was automatically detected and corrected. Quantitative evaluation with respect to manual segmentations was performed on the 16 young subjects, with a leave-one-out strategy, a mixed cohort of 8 controls and 15 patients with epilepsy with variable degrees of hippocampal sclerosis, and 8 healthy subjects acquired on a 3T scanner. Seven performance indices were computed, among which error on volumes RV and Dice overlap K. The method proved to be fast, robust and accurate. For Hc, results with the new method were: 16 young subjects {RV=5%, K=87%}; mixed cohort {RV=8%, K=84%}; 3T cohort {RV=9%, K=85%}. Results were better than with atlas-based (thresholded probability map) or semi-automatic segmentations. Atlas mismatch detection and correction proved efficient for the most sclerotic Hc. For Am, results were: 16 young controls {RV=7%, K=85%}; mixed cohort {RV=19%, K=78%}; 3T cohort {RV=10%, K=77%}. Results were better than with the semi-automatic segmentation, and were also better than atlas-based segmentations for the 16 young subjects.

• Multivariate Reconstruction of Functional Networks From Cortical Sources Dynamics in MEG/EEG

  Authors: A. Dossevi, D. Cosmelli, L. Garnero, H. Ammari

  Biomedical Engineering, IEEE Transactions on. 09/2008;

  In this paper, we present a simple method to find networks of time-correlated brain sources, using a singular value decomposition (SVD) analysis of the source matrix estimated after any linearIn this paper, we present a simple method to find networks of time-correlated brain sources, using a singular value decomposition (SVD) analysis of the source matrix estimated after any linear distributed inverse problem in magnetoencephalography (MEG) and electroencephalography (EEG). Despite the high dimension of the source space, our method allows for the rapid computation of the source matrix. In order to do this, we use the linear relationship between sensors and sources, and show that the SVD can be calculated through a simple and fast computation. We show that this method allows the estimation of one or several global networks of correlated sources without calculating a coupling coefficient between all pairs of sources. A series of simulations studies were performed to estimate the efficiency of the method. In order to illustrate the validity of this approach in experimental conditions, we used real MEG data from a visual stimulation task on one test subject and estimated, in different time windows of interest, functional networks of correlated sources.

• Non-invasive classification of cortical activities for brain computer interface: A variable selection approach

  Authors: M. Besserve, J. Martinerie, L. Garnero

  Biomedical Imaging: From Nano to Macro, 2008. ISBI 2008. 5th IEEE International Symposium on; 06/2008

  We propose to carry out a classification method for electro-encepfialographic signals (EEG), using the activities of cortical sources estimated with an EEG inverse problem. To overcome theWe propose to carry out a classification method for electro-encepfialographic signals (EEG), using the activities of cortical sources estimated with an EEG inverse problem. To overcome the difficulties caused by the high number of sources (approximately 10000), we use a multivariate variable selection algorithm: the zero norm Support Vector Machine (L0-SVM). This technique allows to extract a small subset of sources, which are the most useful to allow for the discrimination of the mental states. The whole approach is applied to an asynchronous Brain Computer Interface (BCI) experiment from our lab. It outperforms a method based on the direct measurement of EEG electrodes' activities.

• Fully automatic hippocampus segmentation discriminates between early Alzheimer’s disease and normal aging

  Authors: M. Chupin, G. Chetelat, L. Lemieux, B. Dubois, L. Garnero, H. Benali, F. Eustache, S. Lehericy, B. Desgranges, O. Colliot

  Biomedical Imaging: From Nano to Macro, 2008. ISBI 2008. 5th IEEE International Symposium on; 06/2008

  The hippocampus is among the first structures affected in Alzheimer's disease (AD); hippocampal MRI volumetry is a potential biomarker for AD but is hindered by the limitations of manualThe hippocampus is among the first structures affected in Alzheimer's disease (AD); hippocampal MRI volumetry is a potential biomarker for AD but is hindered by the limitations of manual segmentation. We propose a fully automatic method using probabilistic and anatomical priors for hippocampus segmentation. Probabilistic information is derived from 16 young controls and anatomical knowledge is modeled with automatically detected landmarks. The results were compared with manual segmentation on data from 16 young healthy controls, with a leave-one-out strategy, and 8 AD patients. High accuracy was found for both groups (volume error 6% and 7%, overlap 87% and 86%, respectively). The resulting volumes were used to discriminate between 25 elderly subjects, 25 early AD patients and 24 patients with amnestic mild cognitive impairment (MCI). The classification proved accurate with 87% of the AD patients and 74% of the MCI patients correctly classified with respect to the elderly controls.

• Cross-Spectral Discriminant Analysis (CSDA) for the classification of Brain Computer Interfaces.

  Authors: M. Besserve, L. Garnero, J. Martinerie

  Neural Engineering, 2007. CNE '07. 3rd International IEEE/EMBS Conference on; 06/2007

  Mining discriminant information through the spatial and frequency domains is a challenging problem in brain computer interfaces (BCI), especially when the experimental paradigm is not as well-knownMining discriminant information through the spatial and frequency domains is a challenging problem in brain computer interfaces (BCI), especially when the experimental paradigm is not as well-known as motor imagery. We propose a new classification algorithm that extracts relevant oscillatory activities from ongoing electroencephalographic (EEG) signals without prior knowledge of the discriminant frequency bands. In this approach, a spatial filtering matrix transforms EEG signals into several components. In order to minimize the generalization error of a Bayesian classifier, the spatial filter is tuned using spectral density matrices. The algorithmic implementation of this method uses explicitly the cross-spectral properties of EEG data. Results on both realistic simulations of EEG data and real asynchronous BCI experiments show the interest of the method and its superiority over the common spatial pattern (CSP) algorithm.

• Multivariate source prelocalization (MSP): use of functionally informed basis functions for better conditioning the MEG inverse problem.

  Authors: J. Mattout, M Pélégrini-Issac, L Garnero, H Benali

  NeuroImage. 07/2005; 26(2):356-73.

  Spatially characterizing and quantifying the brain electromagnetic response using MEG/EEG data still remains a critical issue since it requires solving an ill-posed inverse problem that does notSpatially characterizing and quantifying the brain electromagnetic response using MEG/EEG data still remains a critical issue since it requires solving an ill-posed inverse problem that does not admit a unique solution. To overcome this lack of uniqueness, inverse methods have to introduce prior information about the solution. Most existing approaches are directly based upon extrinsic anatomical and functional priors and usually attempt at simultaneously localizing and quantifying brain activity. By contrast, this paper deals with a preprocessing tool which aims at better conditioning the source reconstruction process, by relying only upon intrinsic knowledge (a forward model and the MEG/EEG data itself) and focusing on the key issue of localization. Based on a discrete and realistic anatomical description of the cortex, we first define functionally Informed Basis Functions (fIBF) that are subject specific. We then propose a multivariate method which exploits these fIBF to calculate a probability-like coefficient of activation associated with each dipolar source of the model. This estimated distribution of activation coefficients may then be used as an intrinsic functional prior, either by taking these quantities into account in a subsequent inverse method, or by thresholding the set of probabilities in order to reduce the dimension of the solution space. These two ways of constraining the source reconstruction process may naturally be coupled. We successively describe the proposed Multivariate Source Prelocalization (MSP) approach and illustrate its performance on both simulated and real MEG data. Finally, the better conditioning induced by the MSP process in a classical regularization scheme is extensively and quantitatively evaluated.

• Disorganized somatotopy in the putamen of patients with focal hand dystonia.

  Authors: C Delmaire, A Krainik, S Tézenas du Montcel, E Gerardin, S Meunier, J F Mangin, S Sangla, L Garnero, M Vidailhet, S Lehéricy

  Neurology. 05/2005; 64(8):1391-6.

  OBJECTIVE: To explore the selectivity of neuronal somatotopic representation in the striatum of patients with unilateral task-specific dystonia of the right arm. MATERIALS AND METHODS: The authorsOBJECTIVE: To explore the selectivity of neuronal somatotopic representation in the striatum of patients with unilateral task-specific dystonia of the right arm. MATERIALS AND METHODS: The authors used fMRI in 14 right-handed dystonic subjects to examine putaminal organization. Subjects performed flexion/extension of the right and left fingers and toes, and contraction of the lips. RESULTS: Compared to healthy volunteer subjects, dystonic subjects had altered somatotopic organization in the left putamen, contralateral to the affected hand. Disease severity correlated with underactivation and decreased distance between right hand and lip representations. In the right putamen, ipsilateral to the affected hand, the somatotopic organization was not altered but disease severity also correlated with reduced distances between limbs. CONCLUSION: In dystonia there may be a dedifferentiation of the normally segregated cortico-subcortical sensorimotor maps in the putamen, which may contribute to the loss of functional selectivity of muscle activity observed in these dystonic subjects.

• Investigations of dipole localization accuracy in MEG using the bootstrap.

  Authors: F Darvas, M. Rautiainen, D Pantazis, S Baillet, H Benali, J C Mosher, L Garnero, R.M. Leahy

  NeuroImage. 05/2005; 25(2):355-68.

  We describe the use of the nonparametric bootstrap to investigate the accuracy of current dipole localization from magnetoencephalography (MEG) studies of event-related neural activity. The bootstrapWe describe the use of the nonparametric bootstrap to investigate the accuracy of current dipole localization from magnetoencephalography (MEG) studies of event-related neural activity. The bootstrap is well suited to the analysis of event-related MEG data since the experiments are repeated tens or even hundreds of times and averaged to achieve acceptable signal-to-noise ratios (SNRs). The set of repetitions or epochs can be viewed as a set of independent realizations of the brain's response to the experiment. Bootstrap resamples can be generated by sampling with replacement from these epochs and averaging. In this study, we applied the bootstrap resampling technique to MEG data from somatotopic experimental and simulated data. Four fingers of the right and left hand of a healthy subject were electrically stimulated, and about 400 trials per stimulation were recorded and averaged in order to measure the somatotopic mapping of the fingers in the S1 area of the brain. Based on single-trial recordings for each finger we performed 5000 bootstrap resamples. We reconstructed dipoles from these resampled averages using the Recursively Applied and Projected (RAP)-MUSIC source localization algorithm. We also performed a simulation for two dipolar sources with overlapping time courses embedded in realistic background brain activity generated using the prestimulus segments of the somatotopic data. To find correspondences between multiple sources in each bootstrap, sample dipoles with similar time series and forward fields were assumed to represent the same source. These dipoles were then clustered by a Gaussian Mixture Model (GMM) clustering algorithm using their combined normalized time series and topographies as feature vectors. The mean and standard deviation of the dipole position and the dipole time series in each cluster were computed to provide estimates of the accuracy of the reconstructed source locations and time series.

• Dynamics of parietofrontal networks underlying visuospatial short-term memory encoding.

  Authors: A C Croizé, R Ragot, L Garnero, A Ducorps, M Pélégrini-Issac, K Dauchot, H Benali, Y Burnod

  NeuroImage. 12/2004; 23(3):787-99.

  Brain imaging studies in TEP, functional magnetic resonance imaging (fMRI) and electroencephalography (EEG) have shown that visuospatial short-term memory tasks depend on dorsal parietofrontalBrain imaging studies in TEP, functional magnetic resonance imaging (fMRI) and electroencephalography (EEG) have shown that visuospatial short-term memory tasks depend on dorsal parietofrontal networks. Knowing the spatiotemporal dynamics of this network would provide further understanding of the neural bases of the encoding process. We combined magnetoencephalography (MEG) with EEG and fMRI techniques to study this network in a task, in which participants had to judge the symmetry in position of two dots, presented either simultaneously ("immediate comparison") or successively ("memorization" of a first dot and "delayed comparison", after 3 s, with a second dot). With EEG, larger amplitude was observed in the parietocentral P3b component (350-500 ms) in the immediate and "delayed comparisons" than in "memorization" condition, where topography at this time was more anterior and right lateralized. MEG provided a more accurate localization and temporal variations of sources, revealing a strong M4 component at 450 ms in the "memorization" condition, with two sources localized in parietal and right premotor regions. These localizations are consistent with both fMRI foci and EEG cortical current source densities (CSD), but only MEG revealed the strong increase in premotor region at 450 ms related to "memorization". These combined results suggest that EEG P3B and MEG M4 components reflect two different dynamics in parietofrontal networks: the parietocentral P3b indexes a decision mechanism during the immediate and "delayed comparisons", whereas the MEG M4 component, with a larger right premotor source, reflects the encoding process in visuospatial short-term memory.

• Localization of realistic cortical activity in MEG using current multipoles.

  Authors: K. Jerbi, S Baillet, J C Mosher, G Nolte, L Garnero, R.M. Leahy

  NeuroImage. 07/2004; 22(2):779-93.

  We present a novel approach to MEG source estimation based on a regularized first-order multipole solution. The Gaussian regularizing prior is obtained by calculation of the sample mean andWe present a novel approach to MEG source estimation based on a regularized first-order multipole solution. The Gaussian regularizing prior is obtained by calculation of the sample mean and covariance matrix for the equivalent moments of realistic simulated cortical activity. We compare the regularized multipole localization framework to the classical dipole and general multipole source estimation methods by evaluating the ability of all three solutions to localize the centroids of physiologically plausible patches of activity simulated on the surface of a human cerebral cortex. The results, obtained with a realistic sensor configuration, a spherical head model, and given in terms of field and localization error, depict the performance of the dipolar and multipolar models as a function of variable source surface area (50-500 mm(2)), noise conditions (20, 10, and 5 dB SNR), source orientation (0-90 degrees ), and source depth (3-11 cm). We show that as the sources increase in size, they become less accurately modeled as current dipoles. The regularized multipole systematically outperforms the single dipole model, increasingly so as the spatial extent of the sources increases. In addition, our simulations demonstrate that as the orientation of the sources becomes more radial, dipole localization accuracy decreases substantially, while the performance of the regularized multipole model is far less sensitive to orientation and even succeeds in localizing quasi-radial source configurations. Furthermore, our results show that the multipole model is able to localize superficial sources with higher accuracy than the current dipole. These results indicate that the regularized multipole solution may be an attractive alternative to current-dipole-based source estimation methods in MEG.

• Fusion of simultaneous fMRI/EEG data based on the electro-metabolic coupling

  Authors: P.-J. Lahaye, S. Baillet, J.-B. Poline, L. Garnero

  Biomedical Imaging: Nano to Macro, 2004. IEEE International Symposium on; 05/2004

  This paper presents an algorithm of data fusion for simultaneous EEG and fMRI recordings, which provides an estimation of 'neural' signals on the cortical surface. The technique described here isThis paper presents an algorithm of data fusion for simultaneous EEG and fMRI recordings, which provides an estimation of 'neural' signals on the cortical surface. The technique described here is applied on simulation data, and can be generically applied to any event-related paradigm. The main originality of this method relative to EEG source reconstruction techniques is to take into account the observed relationship between sources energy and the size of corresponding BOLD responses, for each source and across event occurrences, resulting in a better estimation of the sources' energy. In presence of an EEC/BOLD link, simulation studies show that the method proposed here 1) has better spatial discrimination 2) better estimates the temporal course of sources 3) better estimates the coherence of sources, relative to a similar method using EEG information only. The method is also robust to noise on fMRI and EEG. To our knowledge, this is the first attempt to take advantage of the temporal variability across events.

• Better conditioning the MEG/EEG inverse problem: the multivariate source prelocalization approach

  Authors: J. Mattout, J. Daunizeau, M. Pelegrini-Issac, L. Garnero, H. Benali

  Biomedical Imaging: Nano to Macro, 2004. IEEE International Symposium on; 05/2004

  The recently proposed multivariate source prelocalization (MSP) proved efficient and robust for restricting the solution space of the highly under-determined MEG/EEG inverse problem through theThe recently proposed multivariate source prelocalization (MSP) proved efficient and robust for restricting the solution space of the highly under-determined MEG/EEG inverse problem through the estimation of a probability-like coefficient of activation for each cortical area. It makes MSP a good candidate for initializing conventional inverse methods in order to improve electromagnetic source reconstruction. In this paper, we evaluate the benefit of MSP as a preprocessing step for a classical iterative weighted minimum norm algorithm, the FOCUSS approach, whose results highly rely upon its initialization. By using Monte Carlo simulations, we demonstrate the usefulness of such a coupling. Moreover, we quantify the benefits due on the one hand, to the reduction of the solution space and, on the other hand, to the explicit use as functional prior knowledge, of the activation probabilities inferred by MSP.

• Imaging cortical oscillations during sustained visuomotor coordination in MEG

  Authors: K. Jerbi, J.-P. Lachaux, S. Baillet, L. Garnero

  Biomedical Imaging: Nano to Macro, 2004. IEEE International Symposium on; 05/2004

  Cortical oscillations have been shown to play an important role in a wide range of neural activities. In particular, task-related changes in spectral power and task-related modulations of couplingCortical oscillations have been shown to play an important role in a wide range of neural activities. In particular, task-related changes in spectral power and task-related modulations of coupling both within and between neuronal assemblies in specific frequency bands have been reported. In the current study, we use magnetoencephalography (MEG) recordings to produce maps of task-related changes of spectral power in the alpha (8-15Hz), beta (15-30) and gamma (30-50Hz) range during a sustained visuomotor coordination task in which 4 subjects were asked to use a track-ball device to continuously counter the quasi-random rotations of a cube in a 3D environment. We use a wavelet-based method to compute and compare the time-frequency maps corresponding to the visuomotor task to those of the control conditions. We find that, compared to rest conditions, visuomotor coordination is associated with a sustained depression of spectral power in mu and beta bands over the sensorimotor cortex throughout the whole duration of the task (8 seconds) followed by a wide-band increase in power starting within 1 to 2 seconds after the end of the coordination task. The spatial distribution of the relative power changes in relevant frequency bands are computed and displayed as topographic images on a flattened sensor geometry.

• [Dystonia: contributions of functional imaging and magnetoencephalography]

  Authors: S Lehéricy, S Meunier, L Garnero, M Vidailhet

  Revue neurologique. 11/2003; 159(10 Pt 1):874-9.

  Functional neuroimaging using positron emission tomography (PET), and more recently functional MRI (fMRI) and magnetoencephalography (MEG), is a valuable tool to study functional anomalies inFunctional neuroimaging using positron emission tomography (PET), and more recently functional MRI (fMRI) and magnetoencephalography (MEG), is a valuable tool to study functional anomalies in dystonia. Activation studies have contributed to a better understanding of cerebral dysfunction in dystonia showing two main types of abnormalities: changes in activation levels during performance of sensory or motor tasks and disorganization of the selectivity of neuronal representations. In primary dystonia, most PET and fMRI studies have shown overactivity in premotor and prefrontal areas and underactivation of primary sensorimotor areas. In secondary dystonia, premotor and prefrontal areas are similarly overactive as well as primary sensorimotor areas. Altered selectivity of neuronal representations has been described more recently along cortico-subcortical circuits. The loss of neuronal selectivity may contribute to the loss of selectivity of muscular contractions observed in dystonia. Spectroscopic MRI may also be used to measure GABA levels, which are decreased in the cortex and basal ganglia in these patients.

• Segmentation of the amygdalo-hippocampal complex by competitive region growing [MRI analysis]

  Authors: M. Chupin, D Hasboun, F Poupon, S Baillet, L Garnero

  Biomedical Imaging, 2002. Proceedings. 2002 IEEE International Symposium on; 02/2002

  Semi-automatic segmentation of amygdala and hippocampus is of major interest to neurologists but it is a challenge to numerical image analysis techniques because of intrinsic complexity of thoseSemi-automatic segmentation of amygdala and hippocampus is of major interest to neurologists but it is a challenge to numerical image analysis techniques because of intrinsic complexity of those structures. We introduce a new method to approach this problem with Markovian region growing. As we are concerned by segmenting atrophic structures, we did not use shape priors, but constrained the growth with relational and weak geometric priors. The major relational constraint is given by the simultaneous segmentation of hippocampus and amygdala. The algorithm runs quickly on a basic workstation with satisfactory results on both structures.

• Functional MRI-derived priors for solving the EEG/MEG inverse problem

  Authors: J. Mattout, L Garnero, H Benali

  12/2001;

  Introduction Because of their excellent temporal accuracy (of the order of 1 ms), electroencephalography (EEG) and magnetoencephalography (MEG) provide the most relevant data for studying theIntroduction Because of their excellent temporal accuracy (of the order of 1 ms), electroencephalography (EEG) and magnetoencephalography (MEG) provide the most relevant data for studying the temporal dynamics of brain activity. However, difficulties arise when trying to localize the electromagnetic sources of this activity from EEG/MEG scalp recordings. This mathematical inverse problem is indeed ill-posed and largely underdetermined. An efficient way of constraining the problem and thereby reducing the solution space is to perform a regularization. By taking some anatomical and/or functional a priori knowledge into account, the regularization process may yield a more consistent localization of the electromagnetic sources. Anatomical priors have already been used (e.g., [1]) but only few regularization methods (e.g., [2]) have introduced functional information so far. In this study, we propose a new multimodal approach for solving the EEG/MEG inverse problem. This method involves a

• Human brain mapping in dystonia reveals both endophenotypic traits and adaptive reorganization.

  Authors: S Meunier, L Garnero, A Ducorps, L Mazières, S Lehéricy, S T du Montcel, B Renault, M Vidailhet

  Annals of neurology. 11/2001; 50(4):521-7.

  Dystonia has a wide clinical spectrum from early-onset generalized to late-onset sporadic, task-specific forms. The genetic origin of the former has been clearly established. A critical role ofDystonia has a wide clinical spectrum from early-onset generalized to late-onset sporadic, task-specific forms. The genetic origin of the former has been clearly established. A critical role of repetitive skilled motor tasks has been put forward for the latter, while underlying vulnerability traits are still being searched for. Using magnetoencephalography, we looked for structural abnormalities reflecting a preexisting dysfunction. We studied finger representations of both hands in the primary sensory cortex, as compared in 23 patients with unilateral task-specific dystonia and 20 control subjects. A dramatic disorganization of the nondystonic hand representation was found in all patients, and its amount paralleled the severity of the dystonic limb motor impairment. Abnormalities were also observed in the cortex coding the dystonic limb representation, but they were important only in the most severely affected patients. The abnormal cortical finger representations from the nondystonic limb appear to be endophenotypic traits of dystonia. That finger representations from the dystonic limb were almost normal for the less severely affected patients may be due to intrinsic beneficial remapping in reaction against the primary disorder.

• A multiresolution framework to MEG/EEG source imaging.

  Authors: L Gavit, S Baillet, J F Mangin, J Pescatore, L Garnero

  IEEE transactions on bio-medical engineering. 11/2001; 48(10):1080-7.

  A new method based on a multiresolution approach for solving the ill-posed problem of brain electrical activity reconstruction from electroencephaloram (EEG)/magnetoencephalogram (MEG) signals isA new method based on a multiresolution approach for solving the ill-posed problem of brain electrical activity reconstruction from electroencephaloram (EEG)/magnetoencephalogram (MEG) signals is proposed in a distributed source model. At each step of the algorithm, a regularized solution to the inverse problem is used to constrain the source space on the cortical surface to be scanned at higher spatial resolution. We present the iterative procedure together with an extension of the ST-maximum a posteriori method [1] that integrates spatial and temporal a priori information in an estimator of the brain electrical activity. Results from EEG in a phantom head experiment with a real human skull and from real MEG data on a healthy human subject are presented. The performances of the multiresolution method combined with a nonquadratic estimator are compared with commonly used dipolar methods, and to minimum-norm method with and without multiresolution. In all cases, the proposed approach proved to be more efficient both in terms of computational load and result quality, for the identification of sparse focal patterns of cortical current density, than the fixed scale imaging approach.

• Statistical method for source localization in MEG/EEG tomographic reconstruction problem

  Authors: M. Mattout Jr, L. Garnero, M.L. Pelegrini-Issac, H. Benali

  Image Processing, 2001. Proceedings. 2001 International Conference on; 02/2001

  Characterizing the brain electromagnetic activity using magneto-encephalography (MEG) and/or electro-encephalography (EEG) requires solving an ill-posed inverse problem. This ill-posedness is due toCharacterizing the brain electromagnetic activity using magneto-encephalography (MEG) and/or electro-encephalography (EEG) requires solving an ill-posed inverse problem. This ill-posedness is due to spatial indetermination. Thus, restricting the number of the possible sources of the observed activity would improve the conditioning and facilitate the estimation of their amplitude. We developed a multivariate approach for estimating the spatial support of the electromagnetic cortical activity. Within the framework of a distributed source model and taking advantage of the linearity of the problem, we first performed an orthogonal decomposition of the space of possible source contributions. We then defined the subspace that explained the studied data best. The last step consisted in determining the sources that corresponded to this subspace. These sources defined brain activation areas associated to the observed data time window. The approach was validated using simulated MEG data

• Evaluation of inverse methods and head models for EEG source localization using a human skull phantom.

  Authors: S Baillet, J J Riera, G Marin, J F Mangin, J Aubert, L Garnero

  Physics in medicine and biology. 02/2001; 46(1):77-96.

  We used a real-skull phantom head to investigate the performances of representative methods for EEG source localization when considering various head models. We describe several experiments using aWe used a real-skull phantom head to investigate the performances of representative methods for EEG source localization when considering various head models. We describe several experiments using a montage with current sources located at multiple positions and orientations inside a human skull filled with a conductive medium. The robustness of selected methods based on distributed source models is evaluated as various solutions to the forward problem (from the sphere to the finite element method) are considered. Experimental results indicate that inverse methods using appropriate cortex-based source models are almost always able to locate the active source with excellent precision, with little or no spurious activity in close or distant regions, even when two sources are simultaneously active. Superior regularization schemes for solving the inverse problem can dramatically help the estimation of sparse and focal active zones, despite significant approximation of the head geometry and the conductivity properties of the head tissues. Realistic head models are necessary, though, to fit the data with a reasonable level of residual variance.