Visual motion due to eye movements helps guide the hand

Abstract

Movement of the body, head, or eyes with respect to the world creates one of the most common yet complex situations in which the visuomotor system must localize objects. In this situation, vestibular, proprioceptive, and extra-retinal information contribute to accurate visuomotor control. The utility of retinal motion information, on the other hand, is questionable, since a single pattern of retinal motion can be produced by any number of head or eye movements. Here we investigated whether retinal motion during a smooth pursuit eye movement contributes to visuomotor control. When subjects pursued a moving object with their eyes and reached to the remembered location of a separate stationary target, the presence of a moving background significantly altered the endpoints of their reaching movements. A background that moved with the pursuit, creating a retinally stationary image (no retinal slip), caused the endpoints of the reaching movements to deviate in the direction of pursuit, overshooting the target. A physically stationary background pattern, however, producing retinal image motion opposite to the direction of pursuit, caused reaching movements to become more accurate. The results indicate that background retinal motion is used by the visuomotor system in the control of visually guided action.

Fig. 1

A–C Stimulus and protocol used in the experiment. A The target (black rectangle) was presented on a blank screen while subjects fixated on an object (white square). After the target disappeared, the white object began to move and the subjects were instructed to smoothly pursue the object with their eyes. In one condition, there was a physically stationary background grating visible as soon as the eye movement was initiated. In two additional conditions, the background grating could either move in the same direction as the pursuit object (B) or in the opposite direction (condition not shown). The background grating (either moving or stationary) did not appear until after the pursuit eye movement began, so any differences between the results for the three conditions cannot be explained by a difference in the stimulus when the eye was stationary. C The sequence of events in each trial. Subjects fixated on the pursuit object that was initially stationary and then began to move. While subjects fixated the stationary pursuit object, a stationary target was briefly presented. When the pursuit object began to move, a background grating was simultaneously presented. The grating was either stationary (solid line), moving in the direction of pursuit (dashed line), or moving in a direction opposite the pursuit (dotted line). At an unpredictable moment during the pursuit eye movement, an audible beep was presented, signaling that the subject should reach to the remembered location of the target

Fig. 2

Results for the first experiment: the error in pointing movements (ordinate) as a function of the background velocity. Negative values along the ordinate indicate that the endpoint of the reaching movement deviated in the same direction as pursuit (0 represents an accurate reach). The direction of background motion (with or against the direction of pursuit) is indicated by the positive and negative values along the abscissa, respectively. The physical motion of the grating is plotted on the bottom; the retinal speed of the gratings, due to the pursuit eye movement, is plotted along the top. The vertical dashed line shows the physical speed of the grating that created a retinally stationary image (no retinal slip). The data show a roughly linear relationship between the background velocity and the error in the pointing movement; the error in the pointing movement varied as a function of the background velocity. Interestingly, when the grating moved at the same velocity as the pursuit object, producing a retinally stationary image, the reaching movements were strongly shifted in the direction of the eye movement. However, when there was a physically stationary background, producing retinal slip opposite to the direction of pursuit, the reaching movements were more accurate

Fig. 3

A–C Smooth pursuit eye movements while the background moves in the same direction as or the opposite direction to the target. A Solid and dashed lines show the average eye position for one subject across all trials when the background moved with or against the direction of pursuit, respectively (data are merged such that pursuit direction is always rightward). All missing data (for example, due to blinks) and saccades (defined as instantaneous eye velocities greater than 40 deg/s) were removed; missing frames accounted for less than 2% of the data. Data were collected at and normalized to pursuit onset, which occurred after an initial saccade (because the onset of the target’s motion was unpredictable). B The difference between the eye position traces for the two conditions in (A) in which the background moves with or against the direction of pursuit. The average difference in eye position is 0.15 deg (~1.5 mm, dashed line), indicating that when the background moves in the same direction as the pursuit object, the eye is located about 1.5 mm in front of its location when the background moves opposite to the direction of pursuit. This effect, however, is less than one standard error (s.e.m≈0.3 deg, indicated by the error bar), and is not significant (P>0.05). Also, note that the difference in the pointing error (>6 mm, see Fig. 2) is over four times the eye position error found here, suggesting that the pointing error is not entirely due to an error in pursuit. C The difference between the velocity of the smooth pursuit in the two conditions. A positive score indicates that tracking speed was higher when the background moved with pursuit. The instantaneous velocity of the eye was calculated using a running average of the eye position over a 40 ms window, yielding a smoothed estimate of eye speed; a difference score was then subtracted for each frame (every ~4 ms). The average velocity difference is ~0.16 deg/s, but the large variability shows that there is no systematic or significant difference in eye speed as a function of the direction of background motion. The average velocity difference of ~0.16 deg/s is equivalent to a ~1.8% modulation of pursuit speed as a function of background motion direction (so that, depending on the background motion, the pursuit sped up or slowed down by ~1.8%). Data for two other subjects showed modulations of 2.7% and 1.7%. These are insignificant effects, and are consistent with previous reports that pursuit is essentially accurate, especially when the moving background is separated from the target (Howard and Marton 1992; Schweigart et al. 2003; Goltz and Whitney 2004)