University of Nevada, Reno The Effects of Motion on Perceived Size and Other Perceptual Processes A thesis submitted in partial fulfillment of the requirements for the degree of Master of Science in Psychology by Lauren N. Gregg Dr. Gideon Caplovitz/Thesis Advisor May, 2021 THE GRADUATE SCHOOL We recommend that the thesis prepared under our supervision by Lauren Gregg entitled The Effects of Motion on Perceived Size and Other Perceptual Processes be accepted in partial fulfillment of the requirements for the degree of Master of Science Dr. Gideon Caplovitz Advisor Dr. Ryan Mruczek Committee Member Dr. Lorainne Benuto rte ereettie ee rte May, 2021 i Abstract Optical illusions provide important insights into how we process visual information and illusions that alter the perceived size of an object are a valuable tool to study size perception. Studied for over a century, the classic size illusions have informed us about the complex mechanisms underlying how our brains derive the experience of how big or small objects appear to be. However, these illusions have all been static in nature and thus have ignored motion’s effect on size perception. This review discusses observations of novel dynamic versions of these illusions. Motion has a profound impact on the strength of the illusions tested, with added motion typically creating a stronger effect. Some dynamic versions of these images create an illusion twice as strong as the classic static version. Motion-related manipulations lead to uncertainty in the image size representation of the target, specifically due to added noise at the level of retinal input. We propose a hypothesis that each visual cue involved in size perception is reweighted based on the level of precision or uncertainty in their neural representation. Thus, more weight is given to contextual information when the stimulus and/or eye is moving. Biologically accurate models of size perception need to be able to account for the observed effects of motion. ii Acknowledgements I would like to express deep gratitude to my advisor and mentor Dr. Gideon Caplovitz, whose encouragement and support led me down this rewarding path of inquiry. I also want to thank every member past and present of Dr. Caplovitz’s C-Lab who supported me, especially my colleague Taissa Lytchenko. Much gratitude goes to Dr. Ryan Mruczek for his continued guidance through this material and assistance in bringing it all together for this review. I would also like to thank the entire Cognitive & Brain Science and Neuroscience Departments at University of Reno, Nevada, for the opportunity to study here and for all the knowledge imparted to me by my professors and colleagues. Thank you as well to Dr. Lorraine Benuto for your participation on my graduate committee. And finally, thank you to my family and friends for your support and encouragement through this process. iii Table of Contents ! Introduction 1 Principles Revealed by Classic Size Illusions 2 The Precision Uncertainty Hypothesis 9 Observed Effects 9 Bayesian modeling and multisensory integration 13 Related Illusions 14 Future Directions and Open Questions 16 iv List of Figures ! Figure 1 Classic Size Illusions pg. 6 Figure 2 Muller-Lyer Fins pg. 7 Figure 3 Dynamic Moving Ebbinghaus pg. 8 Figure 4 Static vs. Dynamic Corridor Illusion pg. 12 Figure 5 Bayesian Cue Integration pg. 14 1 The Effects of Motion on Perceived Size and Other Perceptual Processes Introduction Illusions have long been thought to provide important insights about how we process visual information. “Can it be that illusions arise from information processing mechanisms that under normal circumstances make the visible world easier to comprehend?” (Gregory, 1968, p.66). In this review, we will discuss how motion affects the magnitude of size illusions and what these observations reveal about size perception in general. Size perception can be defined as the process by which our brains derive the subjective experience of how big something we are looking at appears to be. Estimating the size of an object across the room may feel simple, but this process is not as straightforward as it first may seem. Although the world around us is three dimensional, we do not have direct visual access to this 3D world. Instead, perception is constrained by the flat, two dimensional images projected on the retinas in the back of our eyes. The fundamental constraint of size perception is that the size of an object’s retinal image scales with viewing distance. An elephant far off in the distance appears as a tiny dot on the savannah. Therefore, at the level of the retina, a small object viewed from up close projects the same retinal image as a large object viewed from far away. With perfect information about an object’s image size on our retina plus the exact distance of that object, the visual system would be able to compute the exact size of the object. However, in real world visual experiences, perfect information about viewing distance and image size is simply not available, as our visual system cannot measure distance. This creates ambiguities that prevent an object’s real-world size from being directly detected. Instead, the perceived size of an object must be constructed through the integration of visual cues in the scene that can be directly detected. One way vision scientists and experimental psychologists have tackled the mystery of size perception is through the study of illusions that alter the perceived size of an object. For over a century, size illusions have been used to characterize the factors that contribute to size perception. Some of the most well studied size illusions include the Ebbinghaus, Delboeuf, Müller-Lyer and Ponzo (Figure 1 A-D) illusions, discussed below. Size illusions can be caused by multiple factors- some are caused by inaccurate estimation of distance, however many misperceptions are mainly due to the misestimation of angular size (McCready, 1985). We now know that size perception involves integrating multiple sources of visual 2 information including estimates of both the image size of the object and its distance (Berryhill, Fendrich & Olson, 2009), as well as contextual cues surrounding the object (Coren & Girgus, 1978; Delboeuf, 1865; Titchener, 1901, Mruczek, Blair, Strother, & Caplovitz, 2017), and these processes of integration also contribute to illusions of size. As we will highlight in the following sections, one key aspect that classic size illusions have in common is they are all static (motionless) in nature. In reality, however, our visual environment is defined by movement. We move through the world as do many of the objects around us, and as such the images on our retinas are constantly in motion. While the classic size illusions have revealed foundational elements of size perception, it was surprising to us that the potential influence of motion on size perception has been largely ignored. This presents a problem when trying to develop a fully formed theory of size perception. To address this issue, we have taken classic size illusions and created novel variants incorporating stimulus dynamics and motion. In each case, we have found that motion significantly impacts the strength of the underlying illusion, most often making the illusion stronger but in some cases reducing its magnitude. In this review, we first discuss in more detail the classic size illusions. These are followed by descriptions of dynamic versions of these illusions. The next section reviews observations from our empirical studies, revealing the strong influence of motion on perceived size, and our interpretation of these findings. Following is a discussion of related illusions and how this concept more broadly applies to perception beyond size illusions, and finally we propose future directions and open questions. Principles Revealed by Classic Size Illusions Much of what we know about size perception has been discovered through the study of size illusions. In this section, we will discuss various classical illusions that reveal the main principle discussed in this paper: the perception of an object’s size is influenced by its surrounding context. Size-contrast illusions One of the most famous and well studied of the classic size illusions is the Ebbinghaus figure (Burton, 2001; Thiery, 1896), discovered in the 1890s by German psychologist Hermann Ebbinghaus. The illusion is also sometimes referred to as the Titchener Circles, as it was popularized in the 3 English-speaking world by Edward B. Titchener in a 1901 textbook of experimental psychology (Harris & Yates, 2005). The simple yet fascinating image clearly demonstrates that context influences our perception of an object’s size. In the best known version of the Ebbinghaus illusion (Figure 1), the perceived size of the center circle is influenced by the size of the circles surrounding it. The center target circles are identical in size, but are typically perceived as different. It is commonly explained as a “size contrast” effect, where objects of the same size are positioned next to comparator objects that are either much smaller or much larger than the target stimuli, causing the objects to appear more different than they actually are. As the size of the surrounding circles increases, the perceived size of the target circle decreases, consistent with the concept of size contrast (Roberts, Harris, & Yates, 2005). However, the illusion is not that simple. Other effects have also been shown to predict some variants of the illusion better than a size contrast account, such as contour interactions, distance-dependent attractive and repulsive interactions between neural representations of contours (Todorović & Jovanović, 2018). Such contour interaction effects are more readily observed in the Delboeuf illusion (Figure 1b), which was first presented in 1865 (Delboeuf, 1865). In this illusion the perceived size of a circle is affected by the size of a surrounding circle. If the outer circle is much larger than the inner circle, the target will be seen as smaller, another example of a size contrast effect.