SACCADES TO REMEMBERED TARGETS - THE EFFECTS OF SMOOTH-PURSUIT AND ILLUSORY STIMULUS MOTION

1. Measurements were made in four normal human subjects of the accurac y of saccades to remembered locations of targets that were flashed on a 20 X 30 deg random dot display that was either stationary or moving horizontally and sinusoidally at +/-9 deg at 0.3 Hz. During the interv al between the target flash and the memory-guided saccade, the ''memor y period'' (1.4 s), subjects either fixated a stationary spot or pursu ed a spot moving vertically sinusoidally at +/-9 deg at 0.3 Hz. 2. Whe n saccades were made toward the location of targets previously flashed on a stationary background as subjects fixated the stationary spot, m edian saccadic error was 0.93 deg horizontally and 1.1 deg vertically. These errors were greater than for saccades to visible targets, which had median values of 0.59 deg horizontally and 0.60 deg vertically. 3 . When targets were flashed as subjects smoothly pursued a spot that m oved vertically across the stationary background, median saccadic erro r was 1.1 deg horizontally and 1.2 deg vertically, thus being of simil ar accuracy to when targets were flashed during fixation. In addition, the vertical component of the memory-guided saccade was much more clo sely correlated with the ''spatial error'' than with the ''retinal err or''; this indicated that, when programming the saccade, the brain had taken into account eye movements that occurred during the memory peri od. 4. When saccades were made to targets flashed during attempted fix ation of a stationary spot on a horizontally moving background, a cond ition that produces a weak Duncker-type illusion of horizontal movemen t of the primary target, median saccadic error increased horizontally to 3.2 deg but was 1.1 deg vertically. 5. When targets were flashed as subjects smoothly pursued a spot that moved vertically on the horizon tally moving background, a condition that induces a strong illusion of diagonal target motion, median saccadic error was 4.0 deg horizontall y and 1.5 deg vertically; thus the horizontal error was greater than u nder any other experimental condition. 6. In most trials, the initial saccade to the remembered target was followed by additional saccades w hile the subject was still in darkness. These secondary saccades, whic h were executed in the absence of visual feedback, brought the eye clo ser to the target location. During paradigms involving horizontal back ground movement, these corrections were more prominent horizontally th an vertically. 7. Further measurements were made in two subjects to de termine whether inaccuracy of memory-guided saccades, in the horizonta l plane, was due to mislocalization at the time that the target hashed , misrepresentation of the trajectory of the pursuit eye movement duri ng the memory period, or bath. 8. The magnitude of the saccadic error, both with and without corrections made in darkness, was mislocalized by similar to 30% of the displacement of the background at the time th at the target flashed. The magnitude of the saccadic error also was in fluenced by net movement of the background during the memory period, c orresponding to similar to 25% of net background movement for the init ial saccade and similar to 13% for the final eye position achieved in darkness.9. We formulated simple linear models to test specific hypoth eses about which combinations of signals best describe the observed sa ccadic amplitudes. We tested the possibilities that the brain made an accurate memory of target location and a reliable representation of th e eye movement during the memory period, or that one or both of these was corrupted by the illusory visual stimulus. Our data were best acco unted for by a model in which both the working memory of target locati on and the internal representation of the horizontal eye movements wer e corrupted by the illusory visual stimulus. We conclude that extraret inal signals played only a minor role, in comparison with visual estim ates of the direction of gaze, in planning eye movements to remembered target locations during our illusory paradigms.