Induced twinkle aftereffect as a probe of dynamic visual processing mechanisms

After viewing a blank patch surrounded by a dynamic noise stimulus (a video "snowstorm"), viewers report the prolonged perception of twinkle in the unstimulated region of the blank patch. We compare this induced twinkle aftereffect to the filling-in phenomenon, which may be seen in a small blank region, under similar test conditions but during stimulation. We found that strong induced twinkle aftereffects were seen both centrally and peripherally for blank test regions from 0.5 deg to as large as 20 deg in diameter, whereas filling-in was seen centrally only for test patch diameters smaller than 0.75 deg, becoming stronger peripherally but still limited to test regions less than about 3 deg in diameter. Lower noise density and larger noise element size facilitated filling-in but had little effect on the induced twinkle aftereffect. Conversely, noise frame rate had little effect on filling-in but had to be faster than 10 frames/sec to produce a twinkle aftereffect. Induced twinkle showed binocular superiority but no interocular transfer. The binocular superiority was partially explained by monocular blankout of the dynamic noise by the blank field in the occluded eye. These results all imply a different mechanism for the induced twinkle aftereffect than for filling-in. We consider a model in which the induced twinkle aftereffect is produced by post-inhibitory rebound in complex cells.