The current understanding of underlying mechanisms involved in neuropathologies has contributed to a better comprehension of those diseases origin, progression and resulting phenotypic and behavioral changes. This knowledge is, nevertheless, still insufficient to fully characterize and diagnose disorders at stages that are not advanced ones. There are bilateral contribution of bottom-up approaches, including genetics and molecular interactions studies, and top-down methods, such as behavioral monitoring and analysis of neuro-activity recordings, which have allowed both a realistic, physiological modeling of neuronal processing and a quantitative characterization of the brain, in healthy or impaired conditions. Our approach consists of (1) being able to extract valuable features out of neurophysiological recordings based on dynamically or statistically based methods in order to characterize brain normal and disrupted functions; (2) recreate mathematical, realistic models of neuropsychiatric disorders, and validate them using previously observed features (e.g. complexity, linear-nonlinear interdependence and nonstationarity); (3) finally, implement reliable tools inspired from combined simulated and recorded databases to explain brain activity and/or diagnose neuropathologies from early stage of the disease to more severe demented cases.
We aim at associating the power of computational neuroscience and neural recording analysis to provide models and understanding of neural substrates at several scales, for a better insight of neuropsychiatric disorders and the elaboration of their diagnosis, but also, to contribute in shedding a light on the mysteries of brain functions and cognition.