Directional sensitivity of anterior, posterior, and horizontal canal vestibulo-ocular neurons in the cat.

Neurons subserving the vestibulo-ocular reflex transform the directionality and timing of input from semicircular canals into commands that are appropriate to rotate the eyes in a compensatory fashion. In order to assess the degree to which this transformation is evident in vestibular nucleus neurons of alert cats, we recorded the extracellular discharge properties of 138 second-order vestibular neurons in the superior and medial vestibular nucleus, including 64 neurons identified as second-order vestibulo-ocular neurons by antidromic responses to oculomotor nucleus stimulation and short-latency orthodromic responses to labyrinth stimulation (1.3 ms or less). Neuronal response gains and phases were recorded during 0.5-Hz sinusoidal oscillations about many different horizontal axes and during vertical axis rotations to define neuronal response directionality more precisely than in past studies. Neurons with spatial responses similar to anterior semicircular canal afferents were found to have more diverse maximal activation direction vectors than neurons with responses resembling those of posterior or horizontal canal afferents. The mean angle from neuron response vector to the axis of the nearest canal or canal pair was 19 degrees for anterior canal second-order neurons (n=28) and 20 degrees for anterior canal second-order vestibulo-ocular neurons (n=18), compared with 11 degrees for posterior canal second-order neurons (n=43) and 11 degrees for posterior canal second-order vestibulo-ocular neurons (n=25). Only two second-order vestibulo-ocular neurons (3%) showed a marked dependence of response phase on rotation direction, which is indicative of convergent inputs that differ in both dynamics and directionality. This suggests that spatiotemporal convergence is uncommon in the three-neuron vestibulo-ocular reflex arc of the cat. Neuron vectors included many that were closely aligned with canal axes and several that were better aligned with oblique or superior rectus extraocular muscle excitation axis vectors. Only single examples of second-order vestibulo-ocular neuron vectors were approximately aligned with the pitch and roll coordinate axes. We conclude that second-order vestibulo-ocular neurons do not exclusively represent either the semicircular canal sensory coordinate frame or the extraocular muscle excitation motor coordinate frame, and instead are mostly distributed on a continuum between the input and output coordinate frames, with anterior canal neurons having the widest distribution of directionality.