Second-order optic flow processing

• Published in: Vision research (Oxford) 2007-06, Vol.47 (13), p.1798-1808
• Main Authors: Aaen-Stockdale, Craig, Ledgeway, Tim, Hess, Robert F.
• Format: Article
• Language: English
• Subjects: Adult; Biological and medical sciences; Complex motion; Contrast Sensitivity; Depth Perception; Eye and associated structures. Visual pathways and centers. Vision; Fundamental and applied biological sciences. Psychology; Global; Humans; Motion; Motion Perception; MST; Optic flow; Photic Stimulation - methods; Psychophysics; Rotation; Sensory Thresholds; Space Perception; Time Factors; Vertebrates: nervous system and sense organs
• ISSN: 0042-6989; 1878-5646
• DOI: 10.1016/j.visres.2007.02.022

Optic flow—large-field rotational and radial motion—is processed as efficiently as translational motion for first-order (luminance-defined) stimuli. However, it has been suggested recently that the same pattern does not hold for second-order (e.g. contrast-defined) stimuli. We used random dot kinematogram (RDK) stimuli to determine whether global processing of optic flow is as efficient as processing of global translational motion for both first- and second-order stimuli. For first-order stimuli, we found that coherence thresholds for radial and rotational motion were equivalent to thresholds for translational motion, supporting previous findings. For second-order stimuli we found, firstly, that given sufficient contrast, second-order optic flow can be processed as efficiently as first-order optic flow and, secondly, that rotational and translational second-order motion are processed with equal efficiency. This contradicts the suggestion that there is a loss of efficiency between integration of second-order global motion and second-order optic flow. The third interesting finding was that the processing of radial second-order motion appears to suffer from a deficit that is dependent upon both the contrast and spatial extent of the stimulus. Further experiments discounted the possibility that the observed deficit is caused by a centrifugal or centripetal bias, but demonstrated that a longer temporal integration period for radial second-order motion is responsible for the observed difference. For durations of ∼850 ms, all three types of motion are processed with equal efficiency.