The study of dynamic cortical states and global perceptual performance requires a multi-scale experimental and theoretical approach. Increasing levels of integration need to be explored, from the single neuron (microscopic level) to the very large population of spatially distributed neurons (meso- and macroscopic levels). Sensory systems integrate exteroceptive afferent signals and ‘inferences’ or processes internal to the central nervous system. The latter are generated in primary sensory areas and conveyed by horizontal connectivity or in higher sensory areas and reinjected via feedback connectivity. One of the major challenges of sensory neurophysiology is therefore to understand the interrelationship between the different information flows and their interaction with local functional architectures. The use of massive recordings using electrode arrays whose geometry can be controlled means that, unlike two-photon microscopy, all the cortical layers can be sampled with optimum temporal resolution. Sophisticated ‘spike sorting’ methods allow neuronal identification and measurements at microscopic (action potentials) and mesoscopic (multi-unit activity and LFP) levels on very large ensembles of neurons, in both anaesthetised and awake in vivo preparations.

These techniques will make it possible to build bridges between the different levels of organisation of neuronal activity in identified contexts and to approach the complexity of the dynamic interactions that can be highlighted in sensory cortex, whether in response to different types of stimulation or as a result of its own internal dynamics. However, in order to implement these new techniques and to avoid wasting time redoing what has already been done by other teams (mistakes or successes), the French community will have a unique opportunity to create interactions between the different teams within this GDR.