PhD in Neuroscience  (Paris.VI University).
Ecole Normale Supérieure, Magistère de Biologie (ENS rue d’Ulm, Paris).
Institut des Systemes Complexes Paris Île-de-France (ISC-PIF)

UMR7656 CREA CNRS
57/59, rue Lhomond 75005 Paris, France
tel.: +33-(0)1-42.17.40.33
fax: +33-(0)1-45.35.79.21
 

The current statement of research and project is to give a formalisation of evolution and cognition into physical terms by the mean of a generalised information theory. Therefore, the objectives are duals: to derive a general model of cognition-evolution based on Physic and hopefully (with the optimistic enthusiasm of a biologist discovering physic and mathematic, and that they are in fact a cognitive theory) to bring new concepts (at least new interpretation or bridges) to physic.

To do so, the state variable i’ve been interested in is entropy: the one that binds microscopic and macroscopic levels, that constitutes one of the possible origins of the time arrow, and that quantifies uncertainty, disorder, as well as complexity. To complement and guide my limited competence and knowledge, this project is developed in collaboration with Daniel Bennequin (Topologist and differential geometer), whose own research motivation, namely to propose a neuro-cognitive theory of invariance structure and a theory of the form of the mind (grounded on singularity theory and homology operation), is in deep register with my own.

The central statement of my previous work, based on various literatures, electrophysiological theoretical and psychophysical results, can be resumed in tree words perception is dissipation.

The result of the empirical work states that the adaptation of biological systems to their environment comes to 3 general laws: a redundancy reduction (“intelligent”, differentiation-specialisation and diversity generating process), a determinism increase (this is a central empirical result: in natural condition the noise or variability of cortical responses is very low showing both that cortical states are highly determined by the environment and that its function e.g. the nonlinear part have been widely underestimated), a sparseness increase. This paradigm allows a unified understanding of receptive field (linear and non-linear compound), neuronal coding, neuronal assemblies and oscillations, cortical plasticity (adaptation mechanisms, plasticity notably STDP), and Psychophysic (and qualitative perception – our subjective time and space), as the results of the dissipation of the sensory flow structure in the cortical media.

However, whereas the two first laws can be understood theoretically as a direct result of an information maximisation (infomax) process between the studied system and its environment, the latest do not. The sparseness (first order redundancy) increase appears as an anomalous deviation from this principle which moreover seems to be a crucial and ubiquitous component of living computation. One of the aims of our work is to understand and account formally for those anomalous behaviours, embedding the infomax adaptation paradigm into a more physically relevant diffusion theory. Using the redundancy decomposition of entropy as a ground topological tool, we currently revisit cortical and neuronal processes-mechanisms in the view of the various nonlinear diffusive systems (notably at their critical states) proposed by physic: the brain is alternatively studied as an electrical conductor (with a complex RC adapting function) submitted to complex potential, as a turbulent fluid driven by the environmental flow, as an old glass aging in an environmental glass, an ensemble of an-harmonic coupled oscillators coupled with another environmental ensemble, a field driven by the propagation of the environmental field etc… But these are for the moment just analogies; and we are therefore concentrating our work on the formal definition of the entropy model and its physical implementation.

The first step was to develop a usual homology of information and then hopefully to derive an information expressions of the classical topological invariants such as the Betti’s numbers and Euler Characteristic.

El Boustani, S., Marre, O., Behuret, S., Baudot, P., Yger, P., Bal, T., Destexhe, A. and Fregnac, Y (Submited). Does the power law frequency-scaling of the membrane potential reflect stimulus-dependent correlations in network activity ? An in vivo, in vitro and in computo study.

Baudot P., Levy M., Marre O., and Frégnac, Y, (to be submitted), Eye movement dynamic constrain V1 code precision and sparsening through adapted non-linearities.

Baudot P., Marre O., Levy M., and Frégnac, Y, (to be submitted), Nature is the Code: High reproducibility in V1 and a model of optimal dissipation of visual structural complexity.

Fregnac Y, Monier C, Chavane F, Baudot P, Graham L., Journal of Physiology (Paris)., Shunting inhibition, a silent step in visual cortical computation. 2003 Jul-Nov;97(4-6):441-51.

Monier, C., Chavane, F., Baudot, P., Borg-Graham, L., and Frégnac, Y., Neuron 2003, Orientation and direction selectivity of synaptic inputs in visual cortical neurons: a diversity of combinations produces spike tuning. Vol. 37 . 4. 663-680.

Chavane, F., Monier, C., Bringuier, V., Baudot, P., Borg-Graham, L., Lorenceau, J. and Frégnac, Y. (2000). The visual cortical association field: a Gestalt concept or a physiological entity. Journal of Physiology (Paris). 94: 333-342.