Spectral mixing of rhythmic neuronal signals in sensory cortex.

The ability to compute the difference between two frequencies depends on a nonlinear operation that mixes two periodic signals. Behavioral and psychophysical evidence suggest that such mixing is likely to occur in the mammalian nervous system as a means to compare two rhythmic sensory signals, such as occurs in human audition, and as a means to lock an intrinsic rhythm to a sensory input. However, a neurological substrate for mixing has not been identified. Here we address the issue of nonlinear mixing of neuronal activity in the vibrissa primary sensory cortex of rat, a region that receives intrinsic as well as sensory-driven rhythmic input during natural whisking. In our preparation, the intrinsic signal originates from cortical oscillations that were induced by anesthetics, and the extrinsic input is introduced by periodic stimulation of vibrissae. We observed that the local extracellular current in vibrissa primary sensory cortex contained oscillatory components at the sum and difference of the intrinsic and extrinsic frequencies. In complementary experiments, we observed that the simultaneous stimulation of contralateral and ipsilateral vibrissae at different frequencies also led to current flow at the sum and difference frequencies. We show theoretically that the relative amplitudes of the observed mixture terms can be accounted for by a threshold nonlinearity in the input-output relation of the underlying neurons. In general, our results provide a neurological substrate for the modulation and demodulation of rhythmic neuronal signals for sensory coding and feedback stabilization of motor output.