Perceptual limits of common fate.

We studied the perception of a coherently moving group of collinearly arranged dots ("target dots") that traveled orthogonally to their linear orientation within a background of noise dots moving in random yet straight directions at constant speed ("random-direction noise"). Using a 2-interval forced-choice task we obtained coherence thresholds equal to a signal-to-noise ratio of 1-2%. These thresholds are lower than the 4-10% reported in the literature suggesting that the collinear arrangement of the target dots, in addition to movement, provided form information. Weber's Law was found to hold 4-7 target dots. Overall, sensitivity was constant for a broad range of dot speeds up to at least 6.5 deg/s. Lifetime required for optimal perception was 430 ms, far shorter than the threshold duration of 1 s reported for randomly distributed (i.e., nonaligned) target dots [Vis. Res. 41 (2001) 1891]. Angular deviations from parallel between adjacent motion trajectories were tolerated up to 27 deg for divergence and up to 19 deg for convergence. Diverging motion was detected earlier (after 600-800 ms) than converging motion (>1 s). Forced-choice discrimination yielded a higher proportion of correct responses than the actual (i.e., conscious) perception of the coherently moving group of dots. Our results are consistent with findings from neurophysiological recordings and neuroimaging of motion-sensitive neurons in areas V1 and MT showing broad tuning curves for speed and direction of a moving visual stimulus.

Unveiling the foveal blue scotoma through an afterimage.

The absence of short-wave-sensitive (S-) cones in the human foveola normally goes unnoticed, but the resulting foveal S-cone, or blue, scotoma can be visualized as the negative afterimage of a short-wavelength adapting field on a larger white background. The afterimage has an annular shape with a lighter inner region that corresponds to Maxwell's spot, and a small bright spot in the center corresponding to the foveal blue scotoma. We have shown that the visibility of the center spot in the afterimage approximately follows the spectral sensitivity curve of the S-cones. We further demonstrate that the central bright spot subtends a retinal area that is coincident with the tritanopic region of the foveola. The macular pigment distribution measured for the same observers also peaks in the central fovea, but has a relatively high density over a broader retinal region than the bright spot in the negative afterimage, and more closely corresponds to the lighter annular region of the afterimage. The results support the hypothesis of an active post-receptoral process for filling-in of chromatic scotomas.

Illusory illusions: the significance of fixation on the perception of geometrical illusions.

Some well-known geometrical illusions disappear when the eyes are fixating and saccades are suppressed for a period of time. This disappearance is not accompanied by fading due to stabilisation of the retinal image. Any saccade made on purpose restores the illusion immediately. The fixation time after which some illusions disappeared was measured for four illusions and four subjects each. Effects of practice have been observed after measurements were repeated on successive days. Present theories of vision cannot readily explain the effect.

Masking, persistence, and transfer in rotating arcs.

We demonstrate that the apparent length of a thin white arc on a black disk, rotating concentrically at 2.5 rps, varies with angular length and exposure duration. While short arcs (9-18 degrees ) gradually expand, long arcs (36-72 degrees ) first undergo a brief contraction, before they also expand. On average, perceived elongation asymptotes after 15 s equivalent to visual persistencies ranging from 68 to 170 ms. Using bi- and tri-colored arcs, we find that the apparent increase in length derives from the rear end of the rotating stimulus, while the initial shrinkage derives from contraction of the middle. After 15 s of adaptation, perceived length of the arc decays to actual stimulus length within an average of 6 s and, upon re-exposure of the arc, reaches its former value after only 5 s (priming). When the rotating arc is presented first to one eye and then to the other, apparent elongation transfers partially (46%), suggesting a contribution by the binocular cells in the visual cortex. A partial transfer (26%) also occurs from clockwise to counterclockwise rotation. When tested interocularly, the directional transfer is more pronounced (47%) and equals the interocular transfer under equidirectional conditions, suggesting that the directional transfer (cw versus ccw) might derive from non-directional cortical units. Whereas the initial contraction may be attributable to backward masking, the observed elongation likely reflects a cumulative build-up of after-discharge in cortical neurons over time.