Abstract
In 1981, Leviant devised a star-shaped figure that elicits rotary motion in the absence of real motion. This figure consisted of concentric sets of narrowly spaced radial lines interrupted by moat-like colored annuli. Compelling as this illusion is, there is no explanation to date. Gregory (1993) pointed out a similarity to MacKay's (1957) “complementary” afterimage and suggested accommodative fluctuations as a potential cause. Assuming a neural correlate, Zeki et al. (1993) found PET-activation by Enigma in visual area V5. To rule out accommodation, we paralyzed the lens of one eye and still perceived the rotary motion. We then produced a long-lasting afterimage of the Enigma figure to eliminate retinal image shifts due to eye movements; the rotation on the rings continued to be seen. These results left the black radial lines as a candidate factor. When the radial lines faded from the afterimage, the rotation disappeared. Motion was present, however, when the concentric sets of radial lines were phase-shifted relative to each other obviating the need for collinear lines. Rotary motion was also seen when the colored annuli were replaced by white annular zones defined only by the ends of the radial lines. This finding suggests that the Enigma illusion may be mediated by cortical neurons responding to T-junctions. Indeed, if the radial lines were tilted in opposite directions so that they were no longer orthogonal to the annuli, the illusory motion weakened and ultimately disappeared. Similarly, if the continuous black lines were replaced with dashed lines, the illusion was abolished. On the other hand, a moderately strong rotary motion was still seen when large round blobs were added to the tips of the radial lines rendering any “T”-like property ineffective. Even when there were no radial lines at all and just a number of black blobs dispersed over the white background, some rotary motion could still be perceived. Thus, the Enigma illusion remains enigmatic.
Supported by DFG-grant SP 67/8-2