Dynamics of vestibular neurons during rotational motion in alert rhesus monkeys.

The temporal processing in the encoding of head rotation was investigated by comparing the dynamics of vestibular nuclei neurons with those of the regularly and irregularly firing semicircular canal afferents in alert rhesus monkeys. During earth-vertical axis rotations, neurons without eye movement sensitivity differed in their response dynamics from both regularly and irregularly firing semicircular canal afferents. At high frequencies, central responses increased in sensitivity and maintained phase leads of nearly 30 degrees relative to head velocity. These persistent high-frequency phase leads resembled those of irregularly firing (but not regularly firing) semicircular canal afferents. However, at low frequencies, central responses exhibited significantly smaller phase leads than those of irregularly firing semicircular canal afferents, and dynamics resembled more those of the regularly firing afferents. The response dynamics of central non-eye movement cells were significantly different from those of position-vestibular-pause and eye-head neurons (collectively referred to as eye movement cells). In contrast to the persistent phase leads of non-eye movement neurons, all eye movement cells modulated closely in phase with head velocity at all frequencies down to 0.05 Hz during visual suppression tasks. Vertical canal non-eye movement neurons that were insensitive to both translations and static head tilts led head velocity by approximately 5-30 degrees during high-frequency earth-horizontal axis rotations. Unlike the earth-vertical axis responses that led head velocity at low frequencies by as much as 20-40 degrees, vertical canal neurons only slightly led or even lagged behind head velocity during low-frequency earth-horizontal axis rotations. Posterior canal central non-eye movement cells lagged behind head velocity significantly more than anterior canal neurons. These frequency dependencies of central vestibular neurons in comparison with those of the afferents suggest that both low- and high-pass filtering might be necessary to convert primary semicircular canal afferent response dynamics to central neuron ones.