Abstract
The perceived visual direction of a monocular target is displaced toward the cyclopean visual direction of surround targets. The monocular target seems to be particularly susceptible to the visual direction of the surround when stimulus conditions favor the recruitment of "local-sign" based position mechanisms. This study extends this observation by measuring the alignment bias of a monocular pair of vertically separated Gabor targets presented within a 10' random dot stereogram depth edge. METHODS: Alignment bias was measured for conditions in which the relative offset between the monocular Gabor pair was defined either by horizontal shifts of the envelope only (EO) (carrier comprised horizontal cosine gratings) or by horizontal phase shifts of the carrier only (CO) which comprised vertical cosine gratings (1, 2, 4 & 8 cycles per degree). Alignment bias was also measured for a Gabor target comprising a vertical 1-cpd square-wave grating (SQ) and a 1-cpd vertical missing fundamental (MF) square-wave grating. Alignment bias was measured for each condition (CO, EO, SQ and MF) across four (8, 30, 60, 120 arc minute) vertical separations. RESULTS: Alignment bias increased with vertical separation and scaled proportionally with alignment threshold for the EO condition but not for the CO condition. Alignment bias also increased with vertical separation and scaled proportionally with alignment threshold for the MF condition but not for the SQ condition. CONCLUSIONS: These results strongly suggest that higher-level feature-based position mechanisms are more susceptible to capture by the surround visual direction compared to first-order spatial frequency based position mechanisms. When higher-level feature-based position mechanisms are recruited, positional uncertainty plays a key role in determining the magnitude of capture.
Meeting abstract presented at VSS 2012