Abstract
A range of motion illusions can be elicited by static images, such as Op art paintings, or similar geometric patterns designed by visual scientists. Many of these illusions require the observer to make eye movements that lead to image shifts generating local motion signals (Zanker and Walker, Naturwissenschaften 91, 2004). In an attempt to test the physiological mechanisms underlying motion illusions in the absence of eye movements, we developed the ‘Spinning Disks Illusion‘: concentric rings of disks filled with greylevel gradients appear to spin around the centre when the background luminance is modulated, eliciting specific responses in area V5/MT (Williams et al, Perception 34, 2005). A two-dimensional implementation of a biologically plausible motion detector, the 2DMD model, was used to account for the perceived shift of an isolated greylevel disk under these conditions (Zanker, Perception 36, S64, 2007). This basic phenomenon was used in the present work to generate more complex motion patterns. When randomly distributed sets of disks with radial or tangential orientation of the greylevel gradient are presented in front of a background with modulated luminance, a strong percept of translational and rotational optic flow is perceived by human observers. We used the 2DMD model to assess the strength of the flowfield information as function of a variety of stimulus parameters such as the number and the size of the disks, or the slope of the greylevel gradient. We found that the saliency of optic flow is highest for fine-grain motion detectors, and that it increases with the number of disks and decreases with the size of the disks. Following from the results of such computer simulations, the discrimination of flowfield patterns can be used in psychophysical experiments to study the influence of these stimulus parameters on the strength of the perceived motion illusion.
Support from EPSRC (EP/C015061/1).