Purchase this article with an account.
Barbara Dillenburger, Christian Wehrhahn; Backward masking of illusory contours or their inducers depends on timing. Journal of Vision 2005;5(8):570. doi: https://doi.org/10.1167/5.8.570.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
Objects in visual scenes may be only partly visible, rendering the extraction of context-induced illusory contours (ICs) an essential process in object and scene perception.
Physiological and psychophysical studies show interaction and overlap between real and illusory processes in the first visual areas, suggesting a feedforward-feedback mechanism with the IC being developped first in V2, but fed back to V1. In such a mechanism real lines would interact early with inducers only, but later predominantly with the IC.
We tested this hypothesis in an orientation discrimination task with an oblique abutting line pattern as inducer. The backward mask grating was identical to the inducers, oriented either the same, horizontal, vertical, or opposite oblique. The stimulus was presented for 100 or 130 msec, followed by a 100 msec mask. Orientation discrimination thresholds for the IC under each masking condition were obtained and compared to an unmasked condition. Real lines are masked best by parallel real lines. We thus expected strongest masks to be oriented parallel to the inducers if they interact with the inducers, but parallel to the IC if interfering with the illusory percept. In the hypothesized feedforward-feedback mechanism, we would expect only masking of the IC after completion of the illusory percept. Earlier, however, masks parallel to the inducing stimuli would interfere strongest with perception.
At 100 msec presentation time all line patterns were found to mask with varying strength, suggesting either a fragile state in processing, or an overlap between real and illusory processes. We found at 130 msec only masking by lines parallel to the IC, indicating IC-completion. We speculate that even earlier interaction would show solely inducer masking.
Our results support the hypothesized mechanism by distinguishing at least two steps of IC processing: mixed processing of inducers and IC at 100 msec, but later processing and neural representation of the IC alone.
This PDF is available to Subscribers Only