CEREBELLAR ROLE IN ADAPTATION OF THE GOLDFISH VESTIBULOOCULAR REFLEX

1. The time course of eye velocity responses elicited by head velocity steps was compared in normal, adapted, and cerebellectomized goldfish. Vestibuloocular reflex (VOR) adaptation was induced by combined visual and vestibular stimulation that altered the ratio of eye to head velocity (VOR gain) toward values either higher or lower than the control amplitude. The velocity step consisted of alternating periods of head rotation at a constant velocity of 16% /s zero-to-peak around the vertical axis. 2. The VOR produced by head velocity steps consisted of an early acceleration-related component, the dynamic response, separated from a sustained period of constant velocity, the plateau, by a sag that occurred around 125-150 ms. Latency of the VOR averaged 18 ms for the adducting eye and 20 ms for abducting eye independent of the initial VOR gain. Adapted dynamic VOR responses diverged from the control records at the earliest detectable latency after both high and low VOR gain training. This result demonstrates modification in the shortest latency brain stem VOR pathway, presumably, the three-neuron reflex are. 3. After acute cerebellectomy the adapted dynamic response was unaltered for similar to 50 ms in the low-gain and 70 ms in the high-gain VOR states. Not less than 30% of the altered velocity was retained throughout the remaining dynamic and sustained component. These results demonstrate that the vestibulocerebellum is not necessary for the maintenance of the earliest adapted eye velocity. Hence brain stem pathways are sufficient for the expression of the modified VOR. 4. Purkinje cells identified by simple and complex spikes were recorded extracellularly in the area of the vestibulocerebellum, where electrical stimulation produced conjugate ipsiversive horizontal eye movements. Independent eye and head velocity sensitivities were determined in response to visual world motion and VOR suppression, respectively. The two signals either added, eye velocity induced by vertical axis visual-vestibular stimulation. 5. Latency of Purkinje cell discharge to either a vestibular or visual velocity step exhibited means of 43 and 70 ms, respectively. These latencies demonstrate that the earliest VOR responses produced by brain stem pathways at 18 ms operates without feed-forward cerebellar regulation after initiation of eye movement for 25 and 52 ms, respectively. Because suppression of the VOR at similar to 36 ms precedes the optokinetic response by 40 ms, a separate brain stem mechanism also exists for canceling the VOR at short latency. These observations are consistent with the latency, amplitude, and retention of adapted eye velocity after cerebellectomy as well as the observation that the cerebellum does not store adapted VOR responses. 6. In chronic cerebellectomized goldfish the dynamic response could not be altered by high-and low-gain adaptation paradigms. In contrast, the plateau response could be adapted up to 50 and 30% for high- and low-gain velocity step training, respectively. Thus a distinct brain stem location and/or mechanism may exist for VOR adaptation underlying changes in the sustained component. 7. The activity of Purkinje cells demonstrates a role in VOR execution and adaptation through the continuous integration of visual and vestibular sensory signals with an internal representation of eye velocity. We conclude that the vestibulocerebellum is necessary for the induction, but neither the storage nor the expression, of adaptive changes in the dynamic component. By contrast, a large portion of the sustained response is affected by acute cerebellectomy, yet this component can still be modified in the absence of the vestibulocerebellum.