The Role of Feature Tracking in the Furrow Illusion

Abstract

In the furrow illusion (Anstis, 2012), the perceived path of a moving target follows the veridical path orientation when viewed foveally, but follows the orientation of the texture when viewed peripherally. These radically different motion percepts depending on whether the stimulus is viewed foveally or peripherally has led Anstis to conclude that the furrow illusion reveals "profound differences in the way that the periphery and fovea process visual motion." In the current study, we rather argue that the different percepts can be explained by reduced position acuity with eccentricity and therefore do not imply different ways of processing motion per se. If feature tracking, which is position-based, is involved in the perception of the veridical motion direction, then impairing the feature tracking motion system should strengthen the illusion. To reduce contribution of the feature tracking motion system, we used a crowding paradigm consisting in presenting many nearby targets. We found that under crowding conditions, the furrow illusion was stronger. We conclude that feature tracking was involved in the perception of the veridical motion direction, which is compatible with the hypothesis that the different motion percepts at fixation and in the periphery are due to a reduced position acuity with eccentricity affecting feature tracking, not to different ways of processing motion per se.

Keywords: attentional resolution; crowding; feature tracking; furrow illusion; motion; periphery.

Figure 1

Furrow illusion. When a spot (here a negative lens inverting contrast polarity) drifts over an oriented texture, the perceived motion path at fixation usually matches the veridical motion path of the spot (here vertical), but when the same stimulus is viewed peripherially, the spot is perceived as moving along the texture orientation. To experience the illusion, see Movie 1 in Supplementary Material.

Figure 2

Stimlus composed of five spots moving in a phase-lock configuration. When viewed peripherally, it should be harder to attentionally select one, which should compromise (or reduce) the ability to track them. To experience the lower ability to attentively select them, try counting them when viewed peripherally. See Movie 2 in Supplementary Material.

Figure 3

Strength of the illusion as a function of eccentricity. The orientation of 0° corresponded to the veridical motion orientation as viewed foveally (eccentricity = 0). The orientation of 45° corresponded to the background orientation. The perceived motion orientation was more influenced by the background orientation in the 5-spot configuration (dashed line) compared to the 1 spot configuration (solid line), i.e., the illusion was stronger. Error bars represent standard errors.

Figure 4

Strength of the illusion for the control experiment at 6° of eccentricity for 3 conditons: 1 spot (diameter of 1 dva), 5 spots (each with a diameter of 1 dva), and 1 large spot (diameter of 2.2 dva). As in Figure 3, the orientation of 0° corresponded to the veridical motion orientation as viewed foveally (eccentricity = 0). The orientation of 45° corresponded to the background orientation. Error bars represent standard errors.