Abstract
Purpose::
The human VOR during roll is poorly studied and controversial, largely due to difficulty delivering a roll stimulus and accurately recording eye positions. We employed a high performance test facility to resolve controversies.
Methods::
10 normal adult humans underwent randomly-directed, whole-body rotation about a mid-ocular, naso-occipital axis located between the eyes at accelerations ranging from 250 - 2000°/s2, peak velocities 50-300°/s2, and amplitudes 20-90°. Ocular and head positions were sampled in 3D at 1,200 Hz using dual-winding magnetic search coils on the sclera and upper teeth. Cruciate, central targets were located at 25 or 270 cm, and were either constantly visible, or in total darkness beginning 60 ms prior to rotation.
Results::
All subjects demonstrated a robust dynamic roll VOR with mean latency 51±6 (±SE, range 29-84) ms. Quick phases occurred 100-380 ms after rotation onset, but averaged 290±6 ms at 375°/s2 and 181±4 at 1000°/s2 and higher (p<0.01). Slow phase VOR gain before quick phases was an increasing function of roll intensity: 0.46±0.05 with peak head vel. 125°/s, accel. 375°/s2; but 0.71±0.03 with peak head vel. 280°/s, accel.1600°/s2. Ocular roll amplitude during 90° rolls was bimodal, averaging 13±1° in 7 subjects, in the remaining 3 subjects strikingly larger at 51±3°. Target visibility had no effect on slow phase VOR gain, timing of the first quick phase, or ocular roll amplitude (p>0.1). Target distance had an effect on VOR gain, but only with visible targets for head vel. of 120 - 160°/s (accel. 375 - 1000°/s2) averaged 0.48±0.03 for the 25 cm target and 0.58±0.03 for the 270 cm target (p=0.03).
Conclusions::
The roll VOR has marked nonlinear gain enhancement with increasing stimulus intensity, possibly accounting for earlier reports of absent or low gain. Roll VOR latency is 5-fold longer than yaw latency measured under similar conditions, and similar to otolith VOR latency. Prolonged roll VOR latency could markedly skew the early VOR axis during complex head rotations. A small reduction in roll VOR gain with convergence, was only observed with low stimulus intensity and a visible target.
Keywords: eye movements • ocular motor control • motion-3D