Modeling the cerebral activity with dynamic probabilistic networks

Vincent Labatut (1) , Josette Pastor (1)

1 Neuro-Imagerie Fonctionnelle, Plasticite Cerebrale et Pathologie Neurologique

Understanding the clinical outcomes of brain lesions necessitates knowing the networks of cerebral structures that constitute the substratum of cognitive or sensorimotor functions. This is achieved by interpreting activation data obtained, during the performance of a task, through functional neuroimaging techniques. The difficulty is that no one-to-one correspondence between activated networks and functions can be found. Actually, neuroimaging methods aim at analyzing specifically the activation. They are used to localize spatially and temporally the activated areas, to detect the different areas participating in the same function, and to determine the role of anatomical links on the activation. Clearly, an interpretative method explaining how the activation of large-scale networks derives from the cerebral information processing mechanisms involved in the task performance is missing. Our goal is to provide such a tool. At this global level of representation, the human brain can be considered as a dynamic biological system that can be best modeled by a dynamic Bayesian network. Our modeling approach is based on the anatomical connectivity of cerebral regions, the information processing within cerebral areas and the causal influences that connected regions exert on each other. The information processing within a region is implemented by a causal network of functional primitives that are the interpretation of integrated biological properties. We use experimental results [8, 9] concerning the modulation of the striate cortex's activation by the presentation rate of visual stimuli, to show that our explicit modeling approach allows the formulation and the simulation of functional and physiological assumptions.

Mots clés: cerebral modeling; bayesian networks; neuroimaging

Domaines: Bio-informatique [q-bio.QM]; Bio-Informatique, Biologie Systémique [q-bio.QM]; Neurosciences

Vincent Labatut, Josette Pastor. Modeling the cerebral activity with dynamic probabilistic networks. 5th International Conference on Simulations in Biomedicine, 2003, Ljubljana, Slovenia. pp.459-468, ⟨10.2495/BIO030451⟩. ⟨hal-00634299⟩