Neuronal encoding of natural stimuli: the rat tactile system.

A major challenge of sensory systems neuroscience is to quantify the brain activity underlying perceptual experiences and to explain this activity as the outcome of elemental neuronal response properties. One strategy is to measure variations in neuronal response in relation to controlled variations in an artificial stimulus. The limitation is that the stimuli scarcely resemble those which the sensory system has evolved to process-natural, behaviourally relevant stimuli. A more recent strategy is to measure neuronal responses during presentation of natural stimuli, but such experiments have failed to predict the observed responses according to the fundamental properties of neurons. In the work described here, we focus on tactile sensation in rats, and try to bridge the gap between neurons' responses to natural stimuli and their responses to controlled, artificial stimuli. We focus on texture, a submodality in which the rat whisker sensory system excels. Because the physical characteristics of texture stimuli have not yet been studied, the first set of experiments measures textures from the whiskers' point of view. The second set of experiments describes neurons' responses to textures. The third set of experiments computes kernels (estimates of the extracted stimulus features) of sensory neurons using white noise and then tries to account for natural texture responses according to these kernels. These investigations suggest ways of using natural stimuli to assemble a more complete picture of the neuronal basis of tactile sensation.