Purchase this article with an account.
J. S. Lappin, D. Tadin, B. Gaddy; Detecting spatial patterns of motion energy. Journal of Vision 2001;1(3):163. doi: 10.1167/1.3.163.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
Problem: What are the visual mechanisms for encoding relative motions of separate features? The present study investigated early visual encoding of relative motion by (a) evaluating the influence of neighboring image motions on the detectability of local motion and (b) comparing the detectability of local motion with the discriminability of relative motion. Methods: In Exp. 1, observers discriminated the direction (right/left) of a near optimal stationary drifting Gabor patch (sigma = 13 arcmin; 3 cyc/deg vertical gratings; 5 Hz; 18% contrast) whose motion energy was controlled by the duration of a Gaussian temporal window (s(t) = 9–35 msec). This target patch appeared either alone or, in other conditions, flanked by two additional Gabors with the same parameters, separated from the central Gabor by 1–4 arcdeg and aligned obliquely (so gratings were not aligned). Trials were randomly intermixed in which flankers moved either the same direction or opposite direction from the central Gabor. Exp. 2 compared direction discriminations for the central Gabor alone and for the two flankers alone with discriminations of the relative directions (same/different) of the central and flanking Gabors. Results: In Exp. 1, direction discriminations were consistently best when central and flanking motions moved in opposite directions. For the central Gabor alone, the average duration threshold was 31 msec; and, e.g., for 2 deg flanker separations, the same-direction threshold was 27 msec, and the opposite-direction threshold was 20 msec. These flanker effects were isotropic. Exp. 2 found the surprising result that thresholds for relative motions of both central and flanking Gabors were always lower than thresholds for the central Gabor alone and also lower than or similar to those for the flankers alone. Sensitivity to relative motion was unaffected by positional noise. Conclusion: Early visual mechanisms may best represent local motions in relation to neighboring image structure.
This PDF is available to Subscribers Only