hide
Free keywords:
-
Abstract:
Hand ^ eye coordination optimises oculomotor control in many natural tasks. For example, smooth
pursuit of a moving target (initiation and maintenance) is most accurate when the target is
under the active manual control of an observer (Steinbach and Held, 1968 Science 161 187 ^ 188;
Gauthier et al, 1988 Experimental Brain Research 73 127 ^ 137). We investigated how artificial
alterations of the control-target mapping, typical of computer input devices, affect the perfor-
mance of coordinated smooth-pursuit eye movements. Observers tracked a Gaussian blob target
on a CRT monitor that they moved back and forth (horizontally) with a digitizer pen in a
sinusoidal motion (horizontal digitizer range 23 cm, maximum target displacement about 208).
Eye movements were recorded with the Eyelink 1000 video tracker. We varied control-target
gain (1: 1, 1: 2, 1: 4, reverse 1 :1), speed (0.5 ^ 1.5 Hz) and target visibility (in half the trials the
target disappeared intermittently behind invisible occluders). Performance was evaluated by smooth-
pursuit gain and saccade count. The largest advantage for active tracking was observed for the
1: 1 mapping in the visible condition, suggesting that optimal smooth pursuit occurs for natural
control-target mappings. But performance was better in active control than in passive tracking
for all conditions, with the smallest performance advantage in the critical reverse condition.
This pattern of results suggests that the oculomotor system has access to the kinematic properties
of the control-target relationship, even without directed training prior to the task.
