Assumed Light Source Direction in Different Populations 1 The Assumed Light Source Direction: Evidence from Different Populations Bridget Andrews A thesis submitted to Bangor University for the degree of Doctor of Philosophy 2014 Assumed Light Source Direction in Different Populations 6 Acknowledgements Many people have helped me throughout my time at Bangor. Firstly, many thanks to my supervisor Dr. Ayelet Sapir, who sparked my interest in this thesis topic from the beginning of my MSc, always encouraged my ideas, and was supportive when they did not work. Thanks to Prof. Charles Leek and Prof. Robert Rafal who provided valuable feedback in review meetings. Also thanks to Dr. Giovanni d’Avossa who was not an official supervisor but still shared knowledge and programming expertise, as well as referring stroke patients to me for testing. Thanks also to the rest of the lab who regularly discussed my work and gave thoughtful feedback during lab meetings. I have also received guidance from other colleagues outside of the lab, allowing me to use a variety of testing methods in this thesis. Thanks to Dr. Janet Bultitude and Dr. Nick Davis for enabling me to conduct prism adaptation and tDCS experiments respectively. Furthermore, a lot of this thesis would also not be possible without the staff and particularly the patients of the Ysbyty Gwynedd stroke ward. I have made some brilliant friends during my time at Bangor. Kat Burnett, Julie Kerai, Lowri Hadden, and Thandi Gilder have made me laugh so much and always known what to say during emotional times. Also as a lab member Kat Burnett has contributed ideas about my work during lab meetings and has happened to accompany me on every conference I have been to, making the experience a brilliant one every time! Abi Moore, you drank tea with me during my degree and you listened to me from afar during my PhD, allowing me to see the west coast of Australia with you in my final year, it was an amazing time. Assumed Light Source Direction in Different Populations 7 Also, without Bangor University I would not have met Matt Richins. As my PhD is ending his is just beginning in a new place where I’ve no doubt we will be just as happy. Finally, thanks to my mum and brother for their continuous support, particularly my mum Sue Andrews, who is also the greatest friend. Assumed Light Source Direction in Different Populations 8 Table of Contents Declaration………………………………….………………………………………….…2 Acknowledgements…………………………………………….…………………….…...5 Table of Contents….……………………………………………………...………………8 Table of Figures and Tables….……………………….………………………………...10 Abstract…...…………………………………………………...…………………………11 Chapter 1. Literature Review………………………………………………………..…13 Left Lighting Assumptions…….……………………………….…………...……………15 Other Left Biases in Visual Processing.…………………………………………...…….19 Specific Cerebral Location……………………………………………...........................20 Reference Frames for “Above”..………………………………………………...…...….23 Handedness……………………………………………………………….……………...25 Visual Field Preferences………………………………………………………….......…26 Experience……………………………………………………………...………………...27 Context Dependent Assumptions………………………...……………………...……….29 Summary……………………………...………………...……………………………..….31 Chapter 2. Cultural Experience…………………………………………...……………32 Chapter Overview…………………………………………………...….………….……..32 Introduction……….……………………………………………………………………...33 Experiment 1: First Language English Group................................................................37 Experiment 2: First Language Hebrew Group................................................................43 General Discussion……………..…………………………………………………..……47 Chapter 3. Effects of Lesion and Attentional Deficits ………………………………..51 Chapter Overview……………………………………………..….…………..…………..51 Introduction…………………………………………..…………………………...……..52 Method……………………………………………………………………….…………..58 Results……………………………………………………………………………………62 Discussion………………………………………………………………………………..67 Chapter 4. Disruption of Left Space in Healthy Participants……...……………...….70 Chapter Overview……………………………………………………………...……..…..70 Introduction…………………………………………………………………..…………..71 Experiment 1: Prism Adaptation…...…………………………………………..………..76 Experiment 2: Transcranial Direct Current Stimulation Over 10 Minutes.…….……..82 Experiment 3: Transcranial Direct Current Stimulation Over 20 Minutes……………87 General Discussion……..………………………………………………………………..93 Chapter 5. Effects of Ageing……………..……………………………………………..95 Chapter Overview……………...………………………………………………………....95 Introduction......................................................................................................................96 Method…………………………………………………………………………………..101 Results…………………………………………………………………………………..102 Assumed Light Source Direction in Different Populations 9 Discussion……………………………………………………………………………….103 Chapter 6. General Discussion…...……………………..………………………..……107 Physical properties of the stimuli...................................................................................107 Handedness…..……………..…………………………………………………………..109 Neural Correlates............................................................................................................110 Experience.......................................................................................................................115 Inter-observer variance...................................................................................................117 Object manipulation and internal representations.......................................................118 Limitations and Future Work…………………………………………………………..121 Concluding Remarks……………………………………...…………………………….122 References……………………………...……………………………………….………123 Assumed Light Source Direction in Different Populations 10 Figures and Tables Figures 1.1. Shaded Spheres……………………………………………………………..….…..14 1.2. Previous Light Bias Experimental Stimuli………………………………....…….17 2.1. Experimental Stimulus………………………………………………..…..……….39 2.2. Example Output from a Typical Participant……………………………..……...41 2.3. Mean Light Biases Grouped by Handedness and Language…………….….…..46 3.1. Typical Deficits of a Hemispatial Neglect Patient………………………………...53 3.2. Fixation Points and Scan Paths of a Typical Hemispatial Neglect Patient During a Visual Search Task………………………………………………………………........55 3.3. Example performances in the flower copy task………………………………..…63 3.4. Mean Light Biases Grouped by Patient Type………………………………...…..65 3.5. Individual Patient Biases as Z-scores………………………………………….…..66 4.1. Prism Adaptation Equipment……………………………..…………...…………..77 4.2. Individual Participants’ Light Biases Pre and Post Prism Adaptation.….……..80 4.2. Mean line deviation Grouped by Real and Sham, Pre and Post Brain Stimulation……………………………………………………………………………….84 4.3. Mean Light Biases Grouped by Real and Sham, Pre and Post Brain Stimulation…………………………………………….…………………………………85 4.4. Mean Target RT Grouped by Real and Sham, Pre and Post Brain Stimulation…………………………………………………………………………….....90 4.5. Mean Light Biases Grouped by Real and Sham, Pre and Post Brain Stimulation……………………………………………………………………………….91 4.6. Scatter Plot Displaying the Relationship between Light Bias and Visual Angle……………………………………………………………………………………...93 5.1. Individual Participants’ Light Biases Grouped by Age…………………..…….101 Tables 3.1. Patient Demographics…………...……………………………………………….…64 6.1. Multiple Regression Predictors…………………………………………………...107 Assumed Light Source Direction in Different Populations 11 Abstract The experiments in this thesis measured light source biases in different populations to investigate why observers assume that light originates from above and to the left. All the experiments use the same greyscale stimulus, the “honeycomb”, which consists of a hexagon surrounded by six hexagons. Light and dark edges give the impression of the stimulus being lit from one side, although there is no explicit light source in the picture. Participants viewed the honeycomb presented at different orientations and stated whether they perceived the central hexagon as pushed in or out compared to the surrounding hexagons. Participants’ light biases were calculated from these judgements, the angle at which they assumed the light was originating. The effect of long term experience is explored in Chapters Two and Three, showing that the cultural experience of language modulates the left lighting bias as biases were significantly reduced in a group of first language Hebrew speakers, who read and write from right to left, compared to English speakers. However, the experience of hemispatial neglect, an attentional disorder commonly seen after stroke, does not cause a shift in the light bias. The role of hemispheric asymmetry is examined in Chapters Three, Four, and Five by measuring the effects of lesions after stroke, simulated lesions in healthy participants, and the natural decline of hemispheric asymmetry with ageing. These experiments show that the lighting bias is modulated by hemispheric asymmetry; however the specific role of the right hemisphere is not clear. Chapter Five shows the lighting bias shifts to the right with age, as the right hemisphere degenerates, yet conversely Chapters Three and Four found that a disruption of the right hemisphere causes a leftward shift in participants’ light biases, regardless of lesion location. Assumed Light Source Direction in Different Populations 12 These findings make a significant contribution to the light assumptions literature, showing that the left lighting bias is caused by an interaction between fixed factors, hemispheric asymmetry, and also environmental experience. Assumed Light Source Direction in Different Populations 13 Chapter One. Literature Review The retinal image is two dimensional; observers must use various depth cues in order to recover the three dimensional shape of surfaces. There are a multitude of cues including binocular disparity, occlusion, motion, surface reflectance, texture, and shading. However, these cues can often be interpreted in multiple ways. For instance Figure 1.1 shows two sets of stimuli which are equally consistent with two lighting directions separated by 180˚ degrees; directly above, and directly below. Therefore each sphere can either be perceived as either a bump or a crater. In these cases interpretation is guided by extra-retinal prior knowledge about the environment. This involves making assumptions based on prior experience in order to quickly interpret the shape of objects (Cavanagh & Lecerc, 1989; Mamassian, Landy, & Maloney, 2001). Shading is a cue that indicates depth by showing which areas of a surface are oriented toward a light source (O’Shea, Banks, & Agrawala, 2008); a convex surface will be lighter on the side which is oriented toward a light source. As such, in order to use shading to interpret the shapes in Figure 1.1, observers must use knowledge and prior experience to assume the light source is originating from above or below the scene. As such, prior assumptions aid observers to quickly recover the three dimensional scenes around them. Assumptions include the judgement that there is a single light source and that the light source is located above, rather than below the scene (Cavanagh & Lecerc, 1989; Kleffner & Ramachandran, 1992; Mamassian, Knill, & Kersten, 1998). The assumption of a single light source is evident by the fact that two groups of spheres with opposite shading (Figure 1.1) cannot be perceived as simultaneously convex or concave. If an observer were to mentally reverse the perception of one group, from convex to concave, the group with opposite shading would reverse to convex automatically. Additionally, the Assumed Light Source Direction in Different Populations 14 assumption of the light source placed above the scene means a shaded grey sphere will usually be reported as convex when lighter at the top and concave when lighter at the bottom. These assumptions are thought to reflect environmental regularities as sun light, a single light source, as well as most artificial lights are placed above the observer (Ramachandran, 1988). Therefore it makes ecological sense for observers to assume there is a single light source above the scene, as statistically, that source of light is the most likely. However, some evidence suggests these assumptions may not be, at least exclusively, ecological but also innate. Hershberger (1970) reared chickens in cages lit from below and trained them to discriminate between convex and concave dents. When shown photographs of dents the chickens interpreted the depth in the photographs as though the dents were lit from above, despite never experiencing overhead lighting. Figure 1.1: A group of spheres with two opposite shading directions, one as though lit from above and the other as though lit from below (adapted from Vision Research Lab, UCL, n.d.).