Adaptive Electrosensory Noise Cancellation in the Little Skate, Leucoraja erinacea

Sensory systems are constantly inundated with sensory noise, including self-generated noise (reafference) resulting from the animal's own behaviors. Selective filtering of irrelevant reafferent signals facilitates the detection of relevant, external stimuli from the environment. In the dorsal octavolateralis nucleus (DON) of the hindbrain in Leucoraja erinacea, suppression of reafferent electrosensory signals is accomplished, in part, by an adaptive filter mechanism. Ascending efferent neurons (AENs), the principal cells of the cerebellar-like DON, receive both primary sensory inputs and a wide array of centrally generated predictive inputs related to the animal's movements. The predictive inputs are transmitted by parallel fibers in the molecular layer. Through a process of associative synaptic plasticity, the weights of the parallel fiber inputs are adjusted to suppress AEN responses to predictable, reafferent signals. AENs have been shown to selectively ignore electrosensory stimuli repeatedly coupled to their own movements. AEN responses to electrosensory stimuli consistently coupled to ventilatory motor commands and passive fin movements have previously been shown to be progressively eliminated. Here, I show that swimming motor commands may also serve as a predictive signal for the suppression of coupled electrosensory stimuli. Furthermore, I demonstrate that a cancellation signal develops for a predictable, coupled stimulus but not when the same stimulus is free-running. Finally, by coupling stimuli to both swimming and ventilatory motor commands, I demonstrate that two cancellation signals can independently develop in a single AEN to suppress reafference associated with the two distinct behaviors.