Abstract
Take a Hermann grid of gray alleys on a black background and place white disks at the intersections; observers report a striking illusion of transient black spots on the disks (scintillation). There is controversy over whether scintillation is an effect of lateral inhibition or of the cortical processing of repeating patterns. To assess the importance of a repeating pattern, we tested a Hermann grid with a single white disk at one intersection. To our surprise, the disk completely disappeared when distant intersections were fixated. We examined the elements of the grid necessary for this effect to occur and compared them to those that produce scintillation.
Subjects fixated on a computer screen; stimulus patterns were then presented randomly above or below fixation. Stimuli, present for 300 ms to preclude changes of fixation, consisted of either a 4′2 or a 2′2 modified scintillating grid pattern that could include a single disk at one intersection. Disks of various luminances were presented with delays of various duration between the onsets of the grid and the disk.
Disks significantly darker than the grid alleys were detected with near perfect accuracy. As disk color approached alley color, detectability was reduced for disks relatively distant from fixation, while disks nearer fixation showed scintillation. Increasing the delay improved detection of light disks that otherwise would not be detectable, but did not preclude scintillation. The blanking effect apparently decays rapidly after presentation of the grid. The lack of effect upon dark disks demonstrates that this is not a result of inattention or limited visual capability. Both scintillation and complete cancellation of a light disk at an intersection are possible with minimal repetition of a pattern. In contrast, scintillation does not require simultaneous presentation of the grid and disk, implying a somewhat different mechanism.