Abstract
For many motor behaviors, the more time devoted to planning a movement, the higher the spatial accuracy. To determine whether latency/accuracy trade-offs apply to saccadic eye movements, subjects made saccades in sequence to 4 target circles (diameter 15′–180′) located at the corners of an imaginary square (corner separation 90′–520′). Subjects were instructed to look at each target in sequence at either a fast or slow pace.
In the “fast” condition, the time to complete the sequences increased with increasing target separation and with decreasing target size, in keeping with Fitts's Law (1954). The increase in scanning time was due mainly to a greater frequency of secondary saccades, needed to correct the landing errors of the primary saccades. Secondary saccades were preferred over a strategy of increasing the latency of primary saccades in an attempt to improve accuracy.
Why were subjects reluctant to increase primary saccadic latency? Under the “slow” condition, where increases in latency were encouraged, the fixation pause duration increased by about 50%. The increase in pause duration was due to both a higher occurrence of smaller secondary saccades during the pauses, as well as to longer saccadic latencies. Yet despite the longer latencies in the “slow” condition, the scatter of saccadic landing positions did not change.
These results show that increasing the time available for saccadic planning did not improve the spatial precision of saccades. In our visual display, with fixed target locations and no distracters, latency/accuracy trade-offs did not apply. Saccades thus differ from other motor behaviors, where optimal performance depends on trading off the duration of primary movements with the occurrence of corrections (Meyer et al., 1988). For saccades, the only apparently viable strategy to improve accuracy is to issue more corrections.