Following Randolph Blake's furrow further

Abstract

In 1992, Randolph Blake, in collaboration with Robert Cormack and Eric Hiris, reported a strong deviation in perceived direction for a target moving over an oblique, static grating. Here we follow up on this effect, subsequently called the furrow illusion, to determine its origin. We find, unlike Cormack et al., that it is influenced by the luminance of the target and that it does not survive smooth pursuit of a moving fixation that stabilizes the target on the retina. We also introduce an inverted version of the furrow stimulus with the static grating visible only within the moving target rather than only around it. This "peep-hole" furrow stimulus shows a similar deviation in its direction and is quite similar to the well-known double-drift stimulus (Lisi & Cavanagh, 2015). Like the double-drift but unlike the furrow stimulus, its illusory direction persists when tracking a fixation that moves in tandem with the target. The main source for the illusion in both cases appears to be the terminators where the grating's bars meet the target contour. These terminators move laterally along the target's contour as the target moves vertically and the combination of these two directions creates the illusory oblique motion. However, the loss of the illusion for the tracked furrow stimulus suggests either a contribution from negative afterimages within the target or from induced motion in this case.

Figure 1.

(A) As the square moves up, the intersections slide up and to the right on the top and bottom but on the sides the intersections remain fixed in place relative to the screen. The oblique motion of the intersections along the top and bottom edges adds to the vertical motion to make the square appear to drift up to the right (Cormack et al., 1992; Anstis, 2012). (B) Because the background grating is static, a negative afterimage will build up during steady fixation and be visible inside the gray square. This afterimage may add to the strength of the edge intersections (Thornton & Riga, 2024).

Figure 2.

Illusion strength as a function of target luminance. To indicate the illusion strength, the 12 participants adjusted the orientation of a marker bar to match the perceived path. Luminance is given in HSL units of L. The illusion was strongest for 50% (mid-gray) targets and weakest for 0% (black) targets. Vertical error bars show ±1.0 SE.

Figure 3.

Illusion strength for furrow vs peep-hole furrow. To indicate the illusion strength, the 12 participants adjusted the orientation of a marker bar to match the perceived path. With static fixation, both the furrow and peep-hole furrow show a strong illusion, replicating Experiment 2. The peep-hole furrow was also strong but smaller than the regular furrow effect. Vertical error bars show +1.0 SE.

Figure 4.

Frequency of reporting illusion direction for static vs moving fixation. With static fixation, the illusion was reported on almost every trial for both the furrow (red bars) and peep-hole furrow (blue bars), replicating Experiment 2. With moving fixation, the peep-hole furrow was again frequently reported, but the furrow effect was virtually eliminated. Vertical bars show +1.0 SE of the mean, n = 4.