The effect of learning on smooth-pursuit eye movements was examined in six
healthy subjects by presenting a sequence of horizontal step-ramp target mo
tions that started with a target velocity of 10 degrees/s or 20 degrees/S f
or 100 ms and then in one direction changed to a second target speed that w
as twice as fast (learning direction). Direction, velocity, duration, and t
ime interval between the ramps changed randomly during the entire experimen
t. For the learning direction, the velocity of eye movements in the interva
l 50-80 ms after the onset of pursuit increased with the number of performe
d trials. In the second direction, the target velocity of 10 degrees/s and
20 degrees/s stayed constant during the ramp (control direction). In contra
st to the learning direction, the eye velocity did not change during the se
ssion in most of the tested subjects. The initial smooth-pursuit eye veloci
ty increased in the learning direction after about 60-80 learning trials. T
here was no generalization of the velocity increase to the control directio
n, and there was no change in the smooth-pursuit onset latency for the cont
rol or for the learning direction during the experiment. The results sugges
t that learning influences smooth-pursuit onset even when a complex paradig
m with two different target velocities is used.