Keeping a target in memory does not increase the effect of the Müller-Lyer illusion on saccades

Abstract

The effects of visual contextual illusions on motor behaviour vary largely between experimental conditions. Whereas it has often been reported that the effects of illusions on pointing and grasping are largest when the movement is performed some time after the stimulus has disappeared, the effect of a delay has hardly been studied for saccadic eye movements. In this experiment, participants viewed a briefly presented Müller-Lyer illusion with a target at its endpoint and made a saccade to the remembered position of this target after a delay of 0, 0.6, 1.2 or 1.8 s. We found that horizontal saccade amplitudes were shorter for the perceptually shorter than for the perceptually longer configuration of the illusion. Most importantly, although the delay clearly affected saccade amplitude, resulting in shorter saccades for longer delays, the illusion effect did not depend on the duration of the delay. We argue that visually guided and memory-guided saccades are likely based on a common visual representation.

Keywords: Dorsal visual stream; Gaze; Variability; Ventral visual stream; Vision.

Fig. 1

Schematic illustration of the task. Participants had to fixate the dot at the centre of the screen and remember the position of the target dot on a briefly presented (200 ms) Müller-Lyer illusion while maintaining fixation. After a delay of 0, 0.6, 1.2 or 1.8 s, the fixation dot disappeared and a brief tone sounded to cue the participant to make a saccade to the remembered target position

Fig. 2

Horizontal saccade amplitude for the two shaft lengths and illusions as a function of the duration of the delay following the disappearance of the illusion. Error bars depict the standard errors across participants (n = 17)

Fig. 3

Interquartile range of horizontal saccade amplitudes for the two shaft lengths as a function of the duration of the delay following the disappearance of the illusion. The interquartile ranges were averaged across the ‘long’ and ‘short’ illusions. Error bars depict the standard error across participants (n = 14)