NEURAL BASIS FOR MOTOR LEARNING IN THE VESTIBULOOCULAR REFLEX OF PRIMATES .1. CHANGES IN THE RESPONSES OF BRAIN-STEM NEURONS

1. We recorded from neurons in the brain stem of monkeys before and after they had worn magnifying or miniaturizing spectacles to cause changes in the gain of the vestibuloocular reflex (VOR). The gain of the VOR was estimated as eye speed divided by head speed during passive horizontal head rotation in darkness. Electrical stimulation in the cerebellum was used to identify neurons that receive inhibition at monosynaptic latencies from the flocculus and ventral paraflocculus (flocculus target neurons or FTNs). Cells were studied during smooth pursuit eye movements with the head stationary, fixation of different positions, cancellation of the VOR, and the VOR evoked by rapid changes in head velocity. 2. FTNs were divided into two populations according to their responses during pursuit with the head stationary. The two groups showed increased firing during smooth eye motion toward the side of recording (Eye-ipsiversive or E-i) or away from the side of recording (Eye-contraversive or E-c). A higher percentage of FTNs showed increased firing rate for contraversive pursuit when the gain of the VOR was high (greater than or equal to 1.6) than when the gain of the VOR was low (less than or equal to 0.4). 3. Changes in the gain of the VOR had a striking effect on the responses during the VOR for the FTNs that were E-c during pursuit with the head stationary. Firing rate increased during contraversive VOR eye movements when the gain of the VOR was high or normal and decreased during contraversive VOR eye movements when the gain of the VOR was low. Changes in the gain of the VOR caused smaller changes in the responses during the VOR of FTNs that were E-i during pursuit with the head stationary. We argue that motor learning in the VOR is the result of changes in the responses of individual FTNs. 4. The responses of E-i and E-c FTNS during cancellation of the VOR depended on the gain of the VOR. Responses tended to be in phase with contraversive head motion when the gain of the VOR was low and in phase with ipsiversive head motion when the gain of the VOR was high. Comparison of the effect of motor learning on the responses of FTNs during cancellation of the VOR with the results of similar experiments on horizontal-gaze velocity Purkinje cells in the flocculus and ventral paraflocculus suggests that the brain stem vestibular inputs to FTNs are one site of motor learning in the VOR. 5. Position-vestibular-pause cells (PVP-cells) were identified by their firing properties during eye movement and vestibular stimulation. They showed increased firing for contraversive eye motion with the head stationary and for ipsiversive head motion during cancellation of the VOR. The firing of PVP-cells paused during at least contraversive saccades. If measured either during cancellation of the VOR or during the VOR evoked by ipsiversive rapid changes in head velocity, the vestibular sensitivity of PVP-cells was not related to the gain of the VOR. If measured during the VOR evoked by contraversive rapid changes in head velocity, however, the amplitude of the responses of PVP-cells was related to the gain of the VOR. We conclude that the direct vestibular inputs to PVP-cells are not modified in association with changes in the gain of the VOR, but that PVP-cells may receive feedback signals that are 1) related to eye velocity and 2)modified in conjunction with learning. 6. We obtained information about possible cause-and-effect relationships among the responses of different groups of brain stem neurons by measuring the latency from the onset of head motion to the onset of neuronal responses during the VOR evoked by rapid changes in head velocity. Abducens neurons, PVP-cells, and E-c FTNs responded with latencies that averaged 7.9, 7.2, and 12.9 ms after the onset of head motion. The latencies of FTNs are appropriate to drive the modified component of the VOR, which is expressed 19 ms after the onset of head motion. In contrast, the disynaptic VOR pathway through PVP-cells is not modified in association with changes in the gain of the VOR, and it appears to drive the earliest components of the VOR.