Purchase this article with an account.
Sae Kaneko, Stuart Anstis, Ichiro Kuriki; Effect of temporal modulations of dynamic inducer on tilt illusion. Journal of Vision 2017;17(10):1084. doi: https://doi.org/10.1167/17.10.1084.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
De Valois et al (1986) examined the temporal frequency properties of simultaneous contrasts when an inducing surround varied sinusoidally in luminance/color. They found that illusory modulation was induced into a static test stimulus only at rates below 2-3 Hz. They concluded the induction was a sluggish, possibly cortical process. However, recent studies showed brightness and color inductions can be fast (e.g. Blakeslee & McCourt, 2008) and stronger in flashes than steadily presented stimuli (Kaneko & Murakami, 2012). Previously, we also showed that strong tilt illusion can be seen in flashed stimuli (VSS 2016). We now measured the temporal frequency properties of tilt illusion with dynamic inducers, using a method of adjustment. A vertical test grating was centered in a larger grating pattern, the inducer (both 1.5 cycles-per-degree). The orientation of the inducer modulated back and forth sinusoidally at 0.5-8.3 Hz from -15° to +15º. A comparison grating was presented aside and its orientation was modulated in sync with the inducer, but in opposite phase. Subjects adjusted the amplitude of its modulation to match the orientations induced into the test. We used two types of modulation; continuous condition, where the inducer and the test were always visible, and intermittent condition, where the inducer and the test were flashed for 1 frame (10 ms) at the maximum tilts of the inducer. For continuous condition, the illusion seen at 0.5 Hz weakened as the modulation rate became higher and almost disappeared at 2-3 Hz, as in De Valois et al (1986). For intermittent condition, the tilt illusion persisted up to a higher rate (> 4 Hz). The difference was not explained by assuming that the induction system was a simple low-pass linear filter. Instead, we suggest the involvement of nonlinear mechanism(s).
Meeting abstract presented at VSS 2017
This PDF is available to Subscribers Only