The Bernstein Center for Computational Neuroscience Freiburg


Announcement for the next
Informal Seminar
Dr. Pierre Baudot and Cyril Monier
Unité de Neurosciences Intégratives et Computationnelles (UNIC)
Institut de Neurobiologie Alfred Fessard, CNRS
Paris
Nature is the code: reliable-efficient dissipation and eye-movement processing in V1

Friday, June 30th, 2007
09:00h (c.t.)
BCCN Lecture Hall
Hansastr. 9A

79104 Freiburg
Abstract:

The specificity of sensory network organization in living systems, achieved through evolution, development and short-term plasticity, can be viewed as an ultimate form of memory collection resulting from the adaptation to the environmental statistics. Here we show that the reliability of the neural code in adult primary visual cortex (V1) reflects in a mirror way the complexity and natural relevancy of the sensory input statistics. Using intracellular recordings in the anaesthetized mammal, we find that the activity evoked in V1 during exposure to natural scenes continuously updated by eye-movements, displays highly reproducible dynamical states at the subthreshold membrane potential (Vm) level and a temporal impulsional code at the spiking output level. In contrast, responses to simple artificial stimuli (“optimal” gratings) are highly unreliable, which supports the prevalence of rate coding for unadapted stimuli. In natural-like condition s, the contrast between the temporally dense informative synaptic input with the sparse spiking output shows that cortical computation removes input redundancies by detecting transient precisely coactive assemblies. Introducing a statistical definition of complexity and ordered redundancies, we propose that both noise and redundancy reduction observed in natural-like conditions are a direct consequence of the principle of mutual-information maximisation, suggesting a universal framework for environmental adaptation. This modulation of the code by the relevancy of the inputs statistics, expressed as a balance between externally imposed states and internal ongoing states, may correspond to the well known self-generative property of recurrent networks. From the computational point of view, the irreversible dissipation of the input constraints operated by the cortex is interpreted as the entropic cost to pay for observing and engramming (or forgetting) the information present in the environment.

In order to assess the biological mechanisms generating this impulsional code, we further tried to identify the RFs nonlinearities and their obvious dependence on eye-movements. The stereotyped dynamics imposed on the retinal flow by eye-movements are a prerequisite for visual perception. However, their impact on the processing of visual information in primary visual cortex (V1) remains largely unknown. By analysing the visual responses of V1 neurons to various simplifications of the natural optic flow, we show that saccadic and fixation eye-movements recruit intracortical non-linearities which increase the reliability of membrane potential dynamics and the temporal precision of spiking responses. Our data suggest that the Receptive fields process visual input at different spatio-temporal scales and directional axis during fixational and saccadic eye movements . I n particular, a switch of direction preference is observed during saccadic-like motion for the collinear axis of the receptive field. At the mechanistic level, the cortical effects of natural visuo-motor statistics during full field stimulation are best explained by the sequential recruitment of facilitatory and suppressive processes. This biphasic spike-timing dependent non-linearity promotes Beta-Gamma frequency fast subliminal events further converted into sparse and precise spiking activity, suggesting a generic neo-cortical filtering mechanism. As a result of this dynamic modulation by eye-movements, most of the simple cells responses has to be credited to nonlinear process even at the synaptic input level. We conclude that cortical nonlinearities and eye-movements cooperate during the viewing of natural scenes to achieve a temporal, selective and motor-invariant encoding of high order spatial input statistics.

 

The talk is open to the public. Guests are cordially invited!
www.bccn.uni-freiburg.de