Abstract
A single line, eccentrically-located and roll-tilted-from-vertical within a frontal plane, induces systematic changes in both the elevation of a target perceived to correspond to eye level (VPEL) and the roll-tilt of a small line perceived to correspond to vertical (VPV). When two simultaneous such inducers are located bilaterally symmetrically: (a) they influence VPV but not VPEL when they are parallel; (b) they influence VPEL but not VPV when they are oppositely oriented (Spat Vis, 1995). Thus, the opponent processes serving 2 line inducers of opposite orientations make use of combining rules that are reversed for the two discriminations, indicating different higher order processes. The present study measured the covariation between the two discriminations in 20 unselected subjects. Measurements were made with monocular viewing in otherwise total darkness with each of 5 roll-tilted inducer orientations, 0 (erect), +/−7.5 deg, +/−15 deg, in each of four conditions: left line alone (1L); right line alone (1R); 1L and 1R parallel (2P); 1L and 1R at equal-and-opposite orientations (2C). Each 70 deg-long line was centered at 25 deg horizontal eccentricity. The average slope of illusion magnitude-vs-roll-tilt (measuring sensitivity to induction) for VPEL was −.01 for 2P and +.22 for 2C, with the reverse for VPV, +.26 for 2P and .00 for 2C; these values are consistent with previous work. The important new result is the significant +.45 correlation between the slopes of 2P for VPV and 2C for VPEL, indicating that an individual's sensitivity to the visual influence tends to be similar for the two discriminations. Considering the differences in combining rules, we suggest that the correlation is due to commonality in lower order processes (e.g., in V1). which distribute to the two different higher order processes.
Support: NIH grant EY10534.