Directional acuity for drifting plaids

Orientation discrimination thresholds were estimated for symmetrical "plaids", constructed by the super-imposition of two, independent drifting sine-wave gratings of 2.5 c/deg. Experiments were conducted when the apparent direction of drift was on the two principal meridians (vertical and horizontal), and when the apparent direction of drift was at the two oblique orientations (45 and 135 deg). Acuity for the direction of drift for these stimuli is of the same order of precision as orientation acuity for static or drifting gratings, and exhibits a meridional anisotropy that favours the principal meridians. This anisotropy follows a pattern that is determined by the apparent direction of drift, and not the orientation of the underlying elements. Acuity for orientation is lowest for oblique drift directions, even though both of the elements are oriented on the principal meridians. This finding was confirmed when the orientation of the underlying elements was subject to a random variation. The results are not easily accommodated by models that propose that the individual elements of the plaid are analysed separately. Second, the data are incompatible with models of the oblique effect of orientation discrimination that are based on an axis dependent, differential sampling of the retinal image.