Visuo-motor tracking during reversible inactivation of the cerebellum

Two monkeys were trained to track a continuously moving target using a joystick. One then had a cooling probe implanted in nucleus interpositus of the cerebellum ipsilateral to his tracking arm. The other had a cannula implanted in the ipsilateral cortex of the lateral cerebellum through which local anaesthetic could be infused. Both monkeys showed similar tracking deficits during temporary inactivation of the cerebellum. The main effects seen were an increase in the peak velocity of their intermittent corrective tracking movements, and a decrease in the accuracy of these movements. Linear regression analyses were undertaken of the peak velocity and amplitude of each corrective movement against a number of possible control signals (target velocity, target position, error, error velocity etc.). The initially strong correlation of the amplitude of each movement made with target velocity was severely reduced during cerebellar inactivation, and movement amplitude became better predicted by the error between target and joystick positions. The peak velocity of movements became more strongly correlated with movement amplitude and less correlated with target velocity than in the intact animal. These results are consistent with the hypothesis that intermittent tracking is achieved by the production of 'primitive' movements, that are then adjusted to the correct amplitude and velocity required to catch up with the moving target. Our findings suggest that the cerebellum may normally be responsible for these adjustments, using visual and memorised cues about the target. The velocity of each movement may be reduced, and its amplitude adjusted, by combining measures of the current error with estimates of target speed and direction. We conclude that the cerebellum has an inhibitory role in tuning movements during visuo-motor tasks and that optimal tuning using feedforward measurements of target motion cannot be made without it.