RAPID COMMUNICATION Central Versus Peripheral Determinants of Patterned Spike Activity in Rat Vibrissa Cortex During Whisking

Fee, Michale S., Partha P. Mitra, and David Kleinfeld. Central somatosensory (S1) cortex (Fabri and Burton 1991; Izraeli versus peripheral determinants of patterned spike activity in rat and Porter 1995; Miyashita et al. 1994; White and Deamicis vibrissa cortex during whisking. J. Neurophysiol. 78: 1144–1149, 1977) as well as central pathways at the level of brain stem 1997. We report on the relationship between single-unit activity nuclei (Erzurumlu and Killackey 1979). in primary somatosensory vibrissa cortex of rat and the rhythmic Here we address the question of peripheral versus central movement of vibrissae. Animals were trained to whisk freely in sources of input into rat S1 vibrissa cortex. The whisker air in search of food. Electromyographic (EMG) recordings from system is an inherently active sensory system: the rat rhyththe mystatial pads served as a reference for the position of the mically sweeps its vibrissae through the space around its vibrissae. A fast, oscillatory component in single-unit spike trains head to find and characterize objects of interest (Vincent is correlated with vibrissa position within the whisk cycle. The phase of the correlation for different units is broadly distributed. 1912; Wineski 1983). Our experimental paradigm involves A second, slowly varying component of spike activity correlates recording from single units in S1 vibrissa cortex as trained with the amplitude of the whisk cycle. For some units, the phase animals whisk freely in air in search of a food tube. We and amplitude correlations were of sufficient strength to allow the record single-unit spike trains from multiple electrodes along position of the whiskers to be accurately predicted from a single penetrations through vibrissa cortex along with the contralatspike train. To determine whether the observed patterned spike eral and ipsilateral mystatial electromyograms (EMGs) as activity was driven by motion of the vibrissae, as opposed to central an index of whisker position; the vibrissae move largely in pathways, we reversibly blocked the contralateral facial motor concert, so that their motion may be adequately described nerve during the behavioral task so that the rat whisked only on by a single degree of freedom (Carvell et al. 1991). the ipsilateral side. The ipsilateral EMG served as a reliable referThe EMG has a fast oscillatory component at the whisking ence signal. The fast, oscillatory component of the spike-EMG correlation disappears when the facial motor nerve is blocked. This frequency and a slow component that corresponds to changes implies that the position of vibrissae within a cycle is encoded in the amplitude of a whisk cycle. We consider correlations through direct sensory activation. The slowly varying component between spike trains in sensory cortex and each of the EMG of the spike-EMG correlation is unaffected by the block. This components. To distinguish between central versus periphimplies that the amplitude of whisking is likely to be mediated by eral origins of both fast and slow EMG signals, we ask corollary discharge. Our results suggest that motor cortex does not whether the correlations between whisking and the spike relay a reference signal to sensory cortex for positional information train persist in the absence of contralateral whisker motion. of the vibrissae during whisking. These latter experiments take advantage of the high degree of the coherence of the ipsilateral and contralateral EMG signals (Wineski 1983) and the persistence of whisking in I N T R O D U C T I O N the absence of sensory feedback (Welker 1964). The computational process of extracting a stable picture Preliminary accounts of this work have appeared (Fee et of the world with actively moving sensors is poorly underal. 1995; Kleinfeld et al. 1997). stood. A prevalent hypothesis of sensory-motor integration is that a copy of motor commands, known as corollary disM E T H O D S charge, is used by the sensory system to afford a representation of the environment free of the effects of sensor moveAnimals ment (Evarts 1971; McCloskey 1981). An alternate hypoth-