Pergamon 0042-69W(93)EOOOl-N Vision Res. Vol. 34, No. 9, 1139-I 147, 1994 pp. Copyright 0 1994 Elsevier Science Ltd Printed in Great Britain. All rights reserved 0042-6989/94 $6.00 + 0.00 Properties of the Stereoscopic (Cyclopean) Motion AfterefTect ROBERT PATTERSON,* CHRISTOPHER BOWD,* RAY PHINNEY,* ROBERT POHNDORF,* WANDA J. BARTON-HOWARD,* MICHELLE ANGILLETTA* zyxwvutsrqponmlkjihgfedcbaZYXWVUTSRQPO Received I July 1993; in revised form 30 September 1993 Across four experiments, this study investigated properties of the stereoscopic motion aftereffect (adaptation from moving retinal disparity information). The results showed that stereoscopic motion can induce an adaptation aftereffect across a wide range of conditions and observers, provided that the duration of adaptation is suflkiently long and a perceptually salient test pattern is viewed. Motion adaptation was found to transfer between the stereoscopic and luminance domains [replicating a previous report by Fox, Patterson and Lelunkuhle (1982) Inuestigathe Ophthulmo~ogy and Vimal Science (Suppl.), 22, 1441, suggesting that motion perception from stereoscopic (second-order) and luminance (first-order) attributes is mediated by a common neural substrate. Motion perception Motion aftereffect Cyclopean Stereopsis INTRODUCTION After viewing for some time (e.g. several minutes) an object moving in a given direction, a subsequently- viewed stationary object will appear to move in the opposite direction. This illusion of motion, the so-called “motion aftereffect”, has been studied for hundreds of years (e.g. Adams, 1834; Aristotle, cited in Wohlgemuth, 1911; Purkinje, 1825). In the contemporary literature, the motion aftereffect has been interpreted as reflecting an induced shift in the distribution of neural activity encoding object motion (e.g. Sutherland, 1961; Moulden, 1980; Wright & Johnston, 1985). This paper reports the results of a study investigating the stereoscopic motion aftereffect, motion adaptation from moving retinal disparity information. Stereoscopic motion is cyclopean, which refers to information existing at binocular-integration levels of vision. Julesz (1960, 1971) pioneered the investigation of cyclopean percep- tion by developing computer-generated random-dot stereograms, dichoptic arrays of dots with embedded disparity which defines stereoscopic stimuli visible only to individuals with stereopsis. The concept of cyclopean vision is similar to the “purely binocular process”, a level of processing for which both eyes must be stimulated for activation of the process, equivalent to a logical “AND” operation (Wolfe, 1986). Changing the spatial position of the disparity information across time produces stereo- scopic or cyclopean motion. Retinal disparity is one of several stimulus attributes or features whose displacement in space and time pro- *Department of Psychology, Washington State University, Pullman, WA 99164-4820, U.S.A. vides information for motion perception. Motion can be perceived from displacement of stimulus boundaries defined by differences in luminance, contrast, texture, disparity, and possibly color (Cavanagh & Mather, 1989; Chubb & Sperling, 1989; Nakayama, 1985; Patterson, Ricker, McGary & Rose, 1992; Turano & Pantle, 1989). One classification scheme of attributes, introduced by Julesz (1971) and elaborated by Cavanagh and Mather (1989), is based on geometrical probability. First-order attributes are defined by differences in first-order stat- istics, such as luminance differences, and second-order attributes are defined by differences in second-order statistics, such as texture or disparity (which are second- order because they are defined by differences in the spatial arrangement of pattern elements, not by lumi- nance differences which define individual elements; see Cavanagh & Mather, 1989). This study investigated whether adaptation to moving disparity information (i.e. stereoscopic motion) induces a motion aftereffect. The existence of stereoscopic motion aftereffects has been controversial. Studies by Steinbach and Anstis (1976, cited in Anstis, 1980), Papert (1964), and Zeevi and Geri (1985) have reported that stereoscopic motion induces little or no motion aftereffect. These results have been incorporated into a contemporary theory of motion perception by Anstis (1978, 1980), Braddick (1974, 1980), Nakayama (1985), and Petersik (1989, 1991). According to these authors, there exist two qualitatively different processes for motion perception. A lower-level sensory process (“short-range” process) computes motion from first-order luminance features across small spatial/temporal intervals, while a higher-level cognitive process (“long-range” process) mediates 1139