Contacts: C. Barillot, P. Hellier

Scientific context

The progress made in the acquisition of anatomical volumes and the precision of the functional   examinations opens the way to cortical atlases projects. An essential stage for the establishment of a  cortical anatomo-functional cartography is the development of a nonlinear inter patient registration method. This registration then makes it possible to put in correspondence cortical structures and to defer, from one structure to another, labels like probabilities of activation for instance.

Description

In this context our work is turned into two directions:

• on the one hand, the development of a generic data fusion framework  allowing the registration of cerebral data in different context:

• Anatomical registration of different subjects. This type of registration leads to the construction of computerized brain atlases: automatic segmentation of brain structures with identification of some cortical structures, with possible assignment to symbolic information systems (for probabilities of activation for instance).

 
• Registration of data coming from a same subject. Various functional information (fMRI and MEG for example) can thus be matched and compared (correction of patient motion in fMRI, correction of the deformations in EPI, ...).
• on the other hand, the construction of statistical hierarchical models for the statistical modeling of the variations of forms and position of the cortical folds. These statistical models intend to cooperate with the global registration methods in order to define adaptive and hierarchical dense deformation fields of the human brain anatomy. In this context, we collaborate as principal contractor to a "Sciences de la Cognition" GIS project entitled "Probabilistic Model of Cortical Folds: Application to the Fusion of MRI and MEG Data".
   

Illustrations

Preliminary results showing the learning set after registration of 6 right central sulci (left), and the first deformation mode around the mean model (right).

 

Example of sulci detected onto two subjects and registered based on a global referential (Talairach). This example is a good illustration of the great inter-individual variabilities which can be observed in shape and position. This can lead to important bias when trying to compare or average functional localizations between different individuals.

Useful Links

• Presentation of the "Sciences de la Cognition" GIS project
• Statistical Modeling of Cortical Sulci

G. Le Goualher, C. Barillot, Y. Bizais. Three-dimensional segmentation andrepresentation of cortical sulci using active ribbons.Int. Journal of Pattern Recognition and Artificial Intelligence, 11(8):1295-1315, 1997.

 B. Gibaud, S. Garlatti, C. Barillot, E. Faure. Computerised Brain Atlases vs Decision Support Systems : A Methodological Approach.  Artificial Intelligence in Medicine, 14(1):83-100, Janvier 1998.

C. Barillot, D. Lemoine, L. Le Briquer, F. Lachmann, B. Gibaud. Data Fusion in Medical Imaging: Merging Multimodal and Multipatient Images, Identification of Structures and 3D Display Aspects. European Journal of Radiology, 17(1):22-27, 1993

C. Barillot, S. Champmartin, G. Le Goualher, B. Gibaud. Probabilistic modeling of brain sulci based on statistical analysis of 3D surfaces.  Functional Mapping of the Human Brain, in Neuroimage, T. Paus, A. Gjedde, A. Evans (eds.), Academic Press, Volume 7(4):716, 1998.

Webmaster : vista_webmaster@irisa.fr Last modified: Wed Nov 21 17:48:39 MET 2001