25 octobre 2012
The field of neuromorphic computation has grown from the idea that inspiration for future computational architectures can be gained from a better understanding of information processing in biological neural networks. Information coding in our brain is both digital, in terms of output spike timing, and analogue, produced by the slower, subthreshold changes in membrane voltage resulting from a continual barrage of synaptic inputs. These small and ever-changing voltage fluctuations in the neuronal membrane potential of the single neuron, control its excitability and spiking reliability. The reverse engineering analysis of these synaptic echoes allows to retrieve the functional effective connectivity of the contextual network within which each cell is embedded.
I will review work from my laboratory (UNIC-CNRS) on spatio-temporal features of the processing realized by the early visual system. Multiscale recordings in the mammalian visual cortex of ongoing and evoked dynamics have been compared using electrophysiological intracellular (single cell) and multiple electrode recording (assembly) techniques. By varying and controlling the visual statistics simulated by a virtual oculomotor exploration of our visual environment, we were able to show that the time precision of the code, the reliability of the evoked dynamics of the visual cortical network and the functional organization of visual cortical receptive fields all adapt to the statistics of the sensory signals. Our observations are best explained by an homeostatic representation principle, where complexities of the input statistics and of V1 receptive fields covary inversely. Generalized recruitment by the stimulus of center-surround interactions and local non-linearities tend to reduce the contextual noise in subthreshold dynamics of the single cortical neuron through a divisive shunt effect. Dynamic full field interactions are shown to regularize the functionally expressed organization of V1 receptive fields, making them more linear and “Simple”-like.
A second illustration of the predictive power of multiscale studies of visual processing has been obtained by comparing intracellular and network imaging (voltage sensitive dye) while exploring the “silent” periphery of visual cortical neurons. Using apparent motion noise at saccadic speed, we have inferred from the synaptic echoes (recorded intracellularly) the existence of long-distance propagation of visually evoked activity through lateral (and possibly feedback) connectivity outside the classical receptive field. VSD imaging has been used to visualize, at the mesoscopic level, the propagation patterns travelling at the speed inferred from our microscopic recordings. Our results demonstrate the propagation at the V1 map level of intracortical depolarizing waves, broadcasting an elementary form of collective “belief” to distant parts of the network. The functional features of these slow waves support the hypothesis of a dynamic perceptual association field, facilitating synaptic modulation in space and time during oculomotor exploration. They may serve as a substrate for implementing the psychological Gestalt principles of common fate and axial collinearity.
We conclude from this review that the early visual system is far from being understood, and that the functional dynamics of visual cortical networks show a much higher level of complexity than initially thought. Comparison between different levels of integration not only shows how limited is our understanding of the emergence of feature selective maps in primary visual areas, but reveals unexpected immergence processes through which collective order regulates more microscopic properties in a top down fashion.
Work supported by CNRS, the French National Research Agency (NatStats and V1-complex) and the European Community (FET integrated (BrainScales) and FET-open (Brain-i-nets) grants).
References: Chavane, F,. Sharon, D., Jancke, D., Marre, O., Frégnac, Y. and Grinvald, A. (2011). Lateral spread of orientation selectivity in V1 is controlled by intracortical cooperativity. Front. System Neuroscience, 5:4. 1-26. doi: 10.3389/fnsys.2011.00004. Marre, O., Yger, P., Davison, A.P. and Frégnac, Y. (2009). Reliable recall of spontaneous activity patterns in chaotic cortical networks. Journal of Neuroscience, 2009, 29(46): 14596-14606. Fournier, J., Monier, C., Pananceau, M. and Frégnac, Y. (2011). Adaptation of the Simple or Complex nature of V1 receptive fields to visual statistics. Nature Neuroscience. 14: 1053-1060. Frégnac, Y. (2012). Reading out the synaptic echoes of low level perception in V1. Lecture Notes in Computer Science. 7583: 486-495.
Yves Frégnac (CNRS)