What Can Drawing Expertise Tell Us About Visual and Memory

THESE DE DOCTORAT DE L’UNIVERSITE PARIS DESCARTES Cognition, Comportements et Conduites Humaines (ED 261)

Laboratoire Psychologie de la Perception (UMR 8242) 45 rue des Saints Pères, 75006 Paris, France

WHAT CAN DRAWING EXPERTISE TELL US

ABOUT VISUAL AND MEMORY MECHANISMS?

QU'EST-CE QUE L'EXPERTISE EN DESSIN PEUT NOUS DIRE A PROPOS DU

FONCTIONNEMENT DE LA VISION ET DE LA MEMOIRE VISUELLE?

Florian Perdreau Pour l’obtention du grade de docteur de l’Université Paris Descartes en Sciences Cognitives Patrick Cavanagh Directeur de thèse

DATE DE LA SOUTENANCE 7 novembre 2014

MEMBRES DU JURY Pr. Patrick Cavanagh Université Paris Descartes Examinateur Dr. Thérèse Collins Université Paris Descartes Examinateur Pr. Pieter Medendorp Radboud University Nijmegen Rapporteur Pr. Robert Pepperell Cardiff School of Art and Design Examinateur Dr. Jérôme Sackur Ecole Normale Supérieure Examinateur Pr. Johan Wagemans K.U. Leuven Rapporteur

Résumé

La précision en dessin a été considérablement étudiée chez l’enfant, mais la raison pour laquelle certains adultes sont bien plus précis que d’autres à copier des objets ou des scènes demeure un mystère. Un facteur possible serait l’entrainement : les artistes passent des milliers d’heures à faire des dessins. Le but de cette thèse a été d’explorer dans quelle mesure cet entrainement intensif a pu affecter certains processus de la vision et de la mémoire visuelle. Dans une série d’études, nous avons tout d’abord démontré que l’expertise en dessin n’est pas liée à une perception plus véridique du monde. En effet, les artistes professionnels et les étudiants en Art que nous avons testé n’étaient pas davantage capables de défaire les mécanismes perceptifs automatiques qui corrigent d’ordinaire les effets dus au contexte visuel. Ils étaient autant affectés par les constances visuelles que les novices. Ceci suggère qu’un entrainement en dessin ne pourrait pas affecter des mécanismes perceptifs déjà bien établis, mais plutôt des processus de plus haut ordre, tels que l’analyse visuelle de structure d’objets. Dans deux études, nous avons ensuite cherché à déterminer comment les dessinateurs encodent et intègrent les informations structurelles lorsqu’ils analysent un objet pour le dessiner. Tout d’abord, afin de tester si les artistes avaient une meilleure représentation de formes complexes, nous avons élaboré une tâche de fenêtre contingente dans laquelle les participants devaient classer un objet comme structurellement possible ou impossible alors qu’ils ne pouvaient voir qu’une portion de l’objet centrée sur la position du regard. Les experts étaient capables de faire cette tâche avec de plus petites portions de l’objet. La compétence en dessin serait ainsi liée à la capacité d’intégrer des échantillons d’informations extraits lors de chaque fixation en une représentation interne plus robuste. Nous avons ensuite voulu savoir si la précision en dessin pouvait aussi être liée à l’efficacité de l’encodage des informations structurelles à partir d’une seule fixation (sans mouvements oculaires autorisés), avec l’objet centré soit sur la position de la fixation ou en périphérie visuelle. Nous avons trouvé que les sujets entrainés étaient capables de discriminer des objets impossibles d’objets possibles avec des durées de présentation plus courtes, que ce soit en vision centrale ou périphérique. Enfin, nous avons étudié le rôle de la mémoire visuelle pendant le processus de dessin et cherché à déterminer si les dessinateurs avaient une représentation plus précise de la position des traits de l’objet. Pour cela, nous avons développé une expérience couplant une tâche de dessin sur tablette graphique et une tâche de détection de changement durant laquelle les participants devaient copier une figure sur une tablette graphique. Tout au long du processus de copie, des changements pouvaient intervenir à la fois sur la figure originale et sur la copie, et les participants devaient corriger tout changement détecté, (la figure et le dessin n’étaient visibles qu’en alternance). Nos résultats ont montré que tous nos participants détectaient mieux les changements présents dans la figure originale que dans leur propre dessin. De plus, les experts en dessin était bien meilleurs à détecter les changements, mais seulement lorsque le dessin était impliqué (contrasté avec une simple tâche de détection de changement sans dessin). Pris ensembles, ces résultats démontrent qu’un entrainement intensif en dessin peut affecter des mécanismes perceptifs de haut niveau ainsi que des mécanismes de mémoire visuelle, et non les mécanismes perceptifs basiques déjà bien établis par la longue expérience perceptive que nous partageons tous.

Abstract

Drawing accuracy has been extensively studied in children, but very little is known of what would make some adults more accurate in copying objects or scenes than many others. One factor may simply be practice: artists have often spent thousands of hours making drawings. The focus of this thesis has been to explore how this intensive practice has affected visual and memory processes. In a series of studies, we first demonstrated that drawing expertise does not relate to a more veridical perception of the world: professional artists and art students were no better than novices at seeing scenes accurately – at undoing the automatic perceptual mechanisms that ordinarily correct for visual context like shadows and depth (visual constancies). This suggests that intensive training in drawing may not affect already well-established perceptual mechanisms, but might affect higher-order processes such as visual analysis of object structure. In a number of studies, we next investigated how trained draftspersons visually encode and integrate structural information when analyzing an object. First, to test whether artists had a more advanced ability to represent complex shapes, we designed a gaze-contingent moving window task where participants had to classify an object as structurally possible or impossible, while only being able to see a portion of the object centered on the gaze position. Experts were able to perform this task with smaller samples of the object. This result suggests that skill in drawing relates to the ability to integrate local samples from each fixation into a more robust internal representation. We then asked whether drawing accuracy could also be related to the encoding efficiency of structural information from a single fixation (no eye movements allowed), with the test object centered at fixation or located in peripheral vision. In this case, we found that experts could discriminate possible vs impossible objects with shorter presentation durations and this was true whether the object was presented at fixation or in the periphery. Finally, we investigated the role of visual memory during the drawing process and whether more skilled participants have a better representation of feature locations. To do so, we designed an interactive pen tablet experiment coupled with a change detection task where participants had to copy a figure on a pen tablet. Throughout the copying process, changes could occur in both the original figure and the copy and participants had to correct any changes they noticed (the figure and the drawing were visible in alternation). We found that all participants detected changes better when they occurred in the original than in their own drawing. Moreover, experts were better at detecting changes, but only when drawing was involved (contrasted with a simple change detection task without drawing). Taken together these results demonstrate that intensive training in drawing affects higher-order perceptual and visual memory mechanisms but not basic perceptual mechanisms that already well-grounded on the life-long perceptual experiences that we all share.

1 TABLE OF CONTENTS

1 TABLE OF CONTENTS 7

2 ACKNOWLEDGMENTS 11

3 INTRODUCTION 15

3.1 PROLOGUE 17

3.2 POTENTIAL SOURCES OF DRAWING SKILL 19

3.2.1 EXTERNAL SOURCES OF DRAWING SKILL: TECHNIQUES, RULES AND KNOWLEDGE 20

3.2.2 DEVELOPMENT OF PERCEPTUAL-MOTOR ABILITIES LINKED TO DRAWING SKILL 20

3.3 VISUOMOTOR ASPECTS OF DRAWING SKILL 22

3.3.1 FINE MOTOR SKILL 22

3.3.2 COORDINATE TRANSFORMATION 23

3.3.3 EFFICIENCY OF EYE-HAND COUPLING STRATEGY IN DRAWING 24

3.4 PERCEPTUAL MECHANISMS RELATED TO DRAWING SKILL 25

3.4.1 PERCEPTUAL MODE INVOLVED IN DRAWING 26

3.4.2 INFLUENCE OF OBJECT KNOWLEDGE ON DRAWING SKILL 28

3.4.3 GLOBAL VS. LOCAL VISUAL PROCESSING OF OBJECTS 31

3.5 VISUAL ANALYSIS AND MEMORY OF OBJECTS STRUCTURES 33

3.5.1 PERCEPTUAL CONSTRAINTS IN DRAWING 33

3.5.2 ENCODING EFFICIENCY AND THE CHUNKING HYPOTHESIS 34

3.5.3 MEMORY OF OBJECT STRUCTURE 36

4 EXPERIMENTS 39

4.1 COULD THE “INNOCENT EYE” HYPOTHESIS ACCOUNT FOR DRAWING SKILL? 41

4.1.1 OBJECTIVES AND SUMMARY OF RESULTS 41

4.1.2 DO ARTISTS SEE THEIR RETINAS? 43

4.2 DRAWING EXPERTS ARE BETTER ABLE TO ACCUMULATE OBJECT STRUCTURE ACROSS EYE-

MOVEMENTS. 67

4.2.1 OBJECTIVES AND SUMMARY OF RESULTS 67

4.2.2 THE ARTIST’S ADVANTAGE: BETTER INTEGRATION OF OBJECT INFORMATION ACROSS EYE-

MOVEMENTS 68

4.3 ENCODING OF STRUCTURAL INFORMATION FROM INDIVIDUAL FIXATIONS 93

4.3.1 OBJECTIVES AND SUMMARY OF RESULTS 93

4.3.2 DRAWING SKILL RELATES TO THE EFFICIENCY OF ENCODING OBJECT STRUCTURE 95

4.4 SPECIALIZED VISUAL MEMORY ENGAGED ONLY DURING THE DRAWING PROCESS 127

4.4.1 OBJECTIVES AND SUMMARY OF RESULTS 127

4.4.2 DRAWING EXPERTS HAVE BETTER VISUAL MEMORY — BUT ONLY WHILE DRAWING 129

5 GENERAL DISCUSSION 153

5.1 MAIN ISSUES OF THE THESIS 155

5.2 DISCUSSION OF THE MAIN RESULTS 156

5.2.1 IS AN ARTIST’S PERCEPTION MORE VERIDICAL? 156

5.2.2 DRAWING SKILL AS PERCEPTUAL EXPERTISE? 157

5.2.3 THE CHUNKING HYPOTHESIS 158

5.2.4 DRAWING IS NOT PLAYING CHESS: THE SPECIFICITY OF EXPERTISE IN DRAWING 162

5.2.5 TALENT VS TRAINING 163

5.3 CONCLUSION 164

6 BIBLIOGRAPHY 167

2 ACKNOWLEDGMENTS

First of all, I would like to thank Patrick Cavanagh for having supervised my work during these last four years. I am deeply grateful to Patrick for his inestimable help and his endless patience in the preparation and writing of this thesis and the studies it contains. Patrick provided me with so much more than what a student could have expected from his supervisor. He offered me the opportunity to work on my own projects, even if they were not directly related to his ongoing studies, and to present my work at numerous international conferences as well as at workshops and other Fests he organized in Paris.

Next, I would like to thank all the members of the CAVlab for their kindness, for their help and for all the very rich discussions and nice moments we had, especially around Patrick’s table during our “lab dinners”.

I am also very grateful to the Laboratoire Psychologie de la Perception for its very unique working environment. During these past five years, I had the opportunity to meet, exchange, and work with great researchers and students. I particularly thank my officemates and friends, Marianne, Margaux, Lucile, Léo-Lyuki, Cédric, Andrea, and all the others for all the pleasant times we have shared.

I thank the members of my Jury for having accepted judging the present work: Thérèse Collins, Pieter Medendorp, Robert Pepperell, Jérôme Sackur and Johan Wagemans.

Then, I would like to thank my family members for their support, and especially my wife, Emilie Perdreau, with whom I shared the greatest moments in my life and who was always present to my side, making every day a better day no matter the difficulties that I could have faced during this thesis.

Finally, I would like to dedicate this thesis to my dear grand-mother who always supported me in my projects and in this thesis until her last moment.

3 INTRODUCTION

3.1 PROLOGUE

Realistic drawing is a complex visuomotor task where strokes made on paper are interpreted by observers as objects and scenes. At a first glance, it might be surprising that many of us fail at accurately drawing objects, as this should only require one to transfer the details of one’s conscious visual percept to paper. Not surprisingly, professional artists and draftspersons can produce far more accurate depictions. Of course, professional artists have been intensively training often over many years. This improvement with practice may therefore indicate that drawing is an uncommon task for the human brain, despite its long history, starting with cave art about 40,000 BCE, and continued by almost all children nowadays. Drawing skill might therefore need extra practice to develop the required visual and cognitive processes, and investigating its sources may bring to light specialized mechanisms that are common to general vision and cognition as opposed to others that may not be engaged outside of the context of drawing. The main purpose of this thesis is to understand how normal visual, motor and cognitive processes are challenged by the demands of realistic drawing, and how intensive practice may affect these mechanisms. We will examine whether an artist’s training modifies early stages of perception as well as higher- order operations. We also ask whether expertise in drawing produces processing enhancements that are seen only in the context of drawing or in all contexts.

Observational drawing is a good example of perception-action coupling. Although this coupling is very efficient in most of daily situations – enabling us to reach, grasp or intercept objects of interest – the difficulties that many of us have in drawing indicate that this skill calls on processes not in play in ordinary circumstances. The superior performance of artists suggests that these processes can be developed with training. Investigating expert performance has proven to be an excellent method for understanding the organization of memory and cognitive processes (Ericsson & Lehmann, 1996a). Expertise in domains such as action video games or chess is known to convey long-lasting benefits on perceptual processes (Li, Polat, Makous, & Bavelier, 2009), selective attention (Green & Bavelier, 2003, 2007), visual working memory (Gobet & Simon, 1996a) and is related to changes in brain structures (Draganski & May, 2008). In this thesis, we will adopt an expertise approach, comparing drawing experts and novices in psychophysical tasks, in order to determine what perceptual, memory and motor mechanisms are involved in drawing and to what extent they could be improved by training. It is worth noting that contrary to other expertise (chess, sport, etc.), drawing expertise may not be restricted to specific object categories or spatial organization (e.g. possible vs impossible pieces organization in chess), since experts in drawing would have to deal with all objects or visual scenes they could possibly experience.

INTRODUCTION | Prologue

Therefore, expertise in drawing might have a greater transferability to other perceptual contexts and broader consequences on visual cognition.

Although drawing is present in many artistic activities, it is not necessarily linked to an aesthetic purpose, filling as well many practical roles in, for example, storyboarding, technical drawing or observational drawing in biology. Moreover, the requirements of skilled drawing have become less and less a factor in contemporary arts, which cover many different media. This may explain why many art students claim to have difficulties in drawing accurately (McManus et al., 2010). Finally, drawing is not a unitary behavior as it covers several different aspects that may themselves depend on a variety of perceptual-cognitive abilities. For instance, we can differentiate between copying from actual objects or scenes, and drawing from memory, and making creative drawings. These tasks are not necessarily linked, as one may be trained in one of these domains but not in another. Interestingly, during the Renaissance period, the ability to accurately copy drawings and sketch from living models was considered the first step for students to master before advancing to their own compositions (Roccasecca, 2009). Such progression may entail an increase in complexity of the work as well as increase in the number of cognitive abilities involved. For example, copying from actual models requires the artist to appropriately transform the 3-D structure of the original into a 2-D representation on the canvas. This transformation may rely on mental imagery (e.g. Carson, Quehl, Aliev, & Danckert, 2013) or on projective techniques – abilities not involved when copying other drawings or photographs that are already in a 2-D format. In this thesis, we will only focus on the act of realistically copying 2-D models (line-drawings or photographs) – namely observational drawing – without considering the aspects of aestheticism or creativity. This may be one of the most fundamental and simple aspects of drawing skill and hence probably the best way of starting the investigation of such a complex behavioral activity.

The four studies in this thesis will reveal that individuals with advanced drawing skill do not show any advantages in low-level perceptual or motor routines. Instead they show improvements of higher-order memory and visual processes as well as visuomotor processes very specific to the constraints of drawing. Comparing the performance of artists and non-artists on purely perceptual and on drawing-related tasks isolates a range of high- level cognitive processes not easily accessed in regular experiments on vision and memory.

We first demonstrate that drawing skill does not correlate with enhanced perceptual abilities. Artists’ perception is not more veridical than that of non-artists (Perdreau & Cavanagh, 2011), even though it would undoubtedly be useful to be so. Professional artists and art students were not better than novices at seeing scenes accurately – at undoing the

18 automatic perceptual mechanisms that ordinarily correct for visual context like shadows and depth (visual constancies, see Figure 2). This suggests that intensive training in drawing may not affect already well-established perceptual mechanisms, but might affect higher-order processes such as the visual analysis of object structure. Drawing, of course, is not simply perception, but a combination of perception and motor output, where visual information encoded at each fixation must be integrated into an internal representation suited to motor planning. In a second study, we investigated this link between multiple local views of a target and an integrated representation constructed across multiple eye movements. Participants had to classify an object as structurally possible or impossible while only being able to see a portion of the object through a gaze-contingent window (Perdreau & Cavanagh, 2013b). Experts were able to perform this task with smaller samples of the object, suggesting that drawing skill relates to the ability to integrate local samples from each fixation into a more robust internal representation. We then asked whether drawing accuracy could also be related to the encoding efficiency of structural information from a single fixation (no eye movements allowed), with the test object centered at fixation or located in peripheral vision (Perdreau & Cavanagh, 2014). In this case, we found that experts could discriminate possible vs impossible objects with shorter presentation durations and this was true whether the object was presented at fixation or in the periphery. In a final study, we investigated the role of visual memory during the drawing process and whether more skilled participants have a better representation of feature locations when they are constructing a drawing as opposed to when they are in a non-drawing context (Perdreau & Cavanagh, VSS 2014). To do so, we designed an interactive pen tablet experiment coupled with a change detection task where participants had to copy a figure on a pen tablet. Throughout the copying process, changes could occur in both the original figure and the copy and participants had to correct any changes they noticed (the figure and the drawing were visible in alternation). We found that all participants detected changes better when they occurred in the original than in their own drawing. Critically, experts were better at detecting changes when drawing was involved but not in a simple change detection task not involving drawing.

Taken together these results demonstrate that intensive training in drawing affects higher-order memory and visuomotor mechanisms but not basic perceptual mechanisms that are already well-grounded by the life-long perceptual experiences that we all share.

3.2 POTENTIAL SOURCES OF DRAWING SKILL

Observational drawing is a very complex sequential process that implies many decisions at any moment about what to draw, where to draw it and how to render it, and failures at any of these steps could be sources of drawing inaccuracy. What would be the

INTRODUCTION | Potential sources of drawing skill specific factors accounting for drawing skill? In particular, drawing skill may involve several factors, either internal (perception, motor control) or external (techniques, rules). Before considering the brain mechanisms possibly involved in drawing, and which might guide the development of expertise or be influenced by it, we will discuss first the external factors that do not fully account for drawing proficiency.

3.2.1 EXTERNAL SOURCES OF DRAWING SKILL: TECHNIQUES, RULES AND

KNOWLEDGE Real objects or even photographs of objects are rich in information, such as shadings, edges, colors and textures. Given the medium used, which necessarily restrains the type of information that can be reproduced (Winner, 1982), one would have to depict some features and to omit others. Therefore, an accurate drawing requires knowing what visual features to represent and how to render them. Rendering of shadings or textures requires different drawing techniques (e.g. blending), types of strokes (with different pressures on the paper or using different pencils), segmentation routines (e.g. cutting the objects into regions or volumes), and explicit knowledge of how familiar objects are spatially organized (e.g. human proportions). Choosing which features to represent is probably based on an understanding of the diagnostic features needed in order to trigger a good perceptual recognition (Biederman & Kim, 2008; Cavanagh, 2005; Kennedy, 1974; Kozbelt, Seidel, ElBassiouny, Mark, & Owen, 2010; Ostrofsky, Kozbelt, & Seidel, 2012). Undoubtedly, all this knowledge may contribute to drawing skill, and in contrast, may explain why most of us fail at drawing things we can see, simply because we do not know how to render the features nor which to select. However, it is important to note that differences in drawing accuracy do not emerge only when copying complex scenes, but also when rendering geometric figures as simple as a square (Tchalenko, 2009) and even with an unfamiliar medium (Kozbelt et al., 2010) or an unfamiliar stimulus (Glazek, 2012). This may suggest that drawing skill may not only rely on rendering techniques and object knowledge, but may also depend on more fundamental perceptual- cognitive mechanisms.

3.2.2 DEVELOPMENT OF PERCEPTUAL-MOTOR ABILITIES LINKED TO

DRAWING SKILL Drawing is a behavioral activity observed in most of children that improves with age, from scribbling to more realistic representations. For these reasons, observational drawing has been extensively studied in children as a tool to investigate the development of perceptual-motor abilities thought to underlie drawing skill (e.g. Arnheim, 1954; Broderick & Laszlo, 1987; Del Giudice et al., 2000; Piaget & Inhelder, 1967; Rand, 1973; Van Sommers,

1989). These articles reveal a relationship between age and the complexity of shapes that could be accurately copied by children: circles can be accurately copied by 3 year-olds; squares by 4 year-olds; triangles, 5 years-old; and diamonds 7 years-olds (e.g. Arnheim, 1954; Piaget & Inhelder, 1967). To accurately copy an object or a figure, a child needs to appropriately represent the part-whole relations of the original object and to be able to transpose them to a different space (visuospatial and constructional ability), to select and extract the contours to be depicted (representational decision-making), to accurately judge the length, size and orientation of the selected contours (visual perceptual ability), to plan the sequence of strokes and their order according to mechanical constraints and stroke-direction preferences (planning of actions; Van Sommers, 1984), and to visually control the hand movement (visual motor control ability). Del Giudice and colleagues (Del Giudice et al., 2000) found that all of these abilities are developed by the age of 7 in healthy children, with visuospatial and constructional abilities having the fastest growth between 4 and 5 years old.

However, although most individuals have engaged in drawing during their childhood, few of them continue to practice their drawing skill into adolescence and adulthood. This leads to dramatic differences across individuals in terms of drawing accuracy, from novices to professional artists and draftspersons. Again, one simple reason for such individual differences would be practice, as professional artists and draftspersons spend thousands of hours practicing their drawing skills whereas novices simply do not. But even so, we must determine which abilities are enhanced through practice, whether simply motor skill, perceptual or cognitive mechanisms. If most of the abilities underlying copying accuracy are already developed in late childhood, what would explain the differences in accuracy between adult novices and drawing experts?

Because both perceptual and motor processes are intimately involved in drawing, it is important to determine whether an intensive training in drawing would affect both domains independently or in conjunction. Interestingly, an earlier study (Rand, 1973) called into question the link between improvement in visual constructional abilities and copying skill. This study showed that training children to visually analyze part/whole relationships improved their discrimination performances of object shape but not their ability to copy them. In contrast, teaching children constructional rules to copy stimulus did improve their copying accuracy but not their discrimination performances. These results suggest an independence of visual analysis and motor production. In addition, there is a growing body of evidence showing that long training in drawing and drawing skill might only lead to structural and functional modifications of visuomotor-related brain areas (SMA, premotor and parietal areas) but not of visual areas (Chamberlain et al., 2014; Miall, Gowen, & Tchalenko, 2008; Schlegel et al., 2012; Solso, 2001).

INTRODUCTION | Visuomotor aspects of drawing skill

One of the central purposes of this thesis will be to demonstrate that drawing skill does not relate to improvements of either basic visual or motor processes, but rather to the development of a more efficient interface between both domains: a visual analysis and a memory encoding of objects more suited to the visuomotor constraints of the drawing process.

3.3 VISUOMOTOR ASPECTS OF DRAWING SKILL

The goal of the following sections is to review the motor and visuomotor skills involved in drawing. This is intended to show how a particular visual analysis and encoding of the object could facilitate the accumulation of object information across eye movements as well as the visuomotor mapping of subsequent hand-movements.

3.3.1 FINE MOTOR SKILL When making a drawing, one must be able to trace lines at the desired location using accurate hand movements. Training in writing helps develop fine motor skill in children (Laszlo & Bairstow, 1985) and, undoubtedly intensive training in drawing also improves these fine motor skills. However, even though handwriting and drawing appear to be quite similar in form (children learn to write by copying letter shapes), there are major differences that would make an expertise in one worthless in the other. For instance, handwriting is an automatic process that depends on a limited set of possible shapes, pre-established graphic templates and motor routines (Adi-Japha & Freeman, 2001; Feder & Majnemer, 2007; Van Sommers, 1984; Zesiger, Martory, & Mayer, 1997). In contrast, observational drawing must face an infinite number of possible shapes, as it has to carefully depict objects given specific viewing conditions. This may require an extra control of the hand movements and may recruit more complex visuomotor mapping.

Figure 1. Impact of drawing conditions on copying accuracy (adapted from Cohen & Bennett, 1997). Participants were asked to copy a photograph of a face and or an electrical generator in different copying conditions: either in direct tracing (the copy is superimposed onto the original), in tracing from distance (same as direct tracing but with a distance between the copy and the original) and in traditional copying (the copy and the original are not superimposed). Drawing conditions significantly affected participants’ accuracy (lower ratings indicate higher accuracy). Because the same hand movements were required in all of these conditions, the authors suggested that fine motor skill do not account for the increase in errors between tracing and drawing conditions.

Cohen and Bennett (Cohen & Bennett, 1997) tested motor skill in healthy adults. They asked their participants to perform in different drawing conditions: tracing directly from photographs, tracing from distance (the transparent canvas and the original model are aligned but distant from each other), and traditional copying. As expected, they found that drawings made by participants were judged more accurate in the direct tracing condition than in the distant copying condition, and were also more accurate in the latter condition than in the traditional condition (Fig. 3). The authors argued that, because all these conditions require the same hand movements, the increase in errors in traditional copying conditions couldn’t be accounted by the inability to make the appropriate hand movements.

3.3.2 COORDINATE TRANSFORMATION Although the same hand movements would be involved in all these drawing conditions, these conditions cannot be directly compared as tracing and drawing do not share the exact same motor constraints. Tracing is a pure externally driven motor task

INTRODUCTION | Visuomotor aspects of drawing skill wherein the original and the copy are superimposed and the eyes are closely pursuing the hand’s trajectory (Coen-Cagli et al., 2009; Gowen & Miall, 2006; Miall & Tchalenko, 2001; Tchalenko, Dempere-Marco, Hu, & Yang, 2003; Tchalenko & Miall, 2009; Tchalenko, 2009). This results in a higher synchronization between the visual and motor system with little computation demand for transposing the eye-centered reference frame to hand-centered reference frame. In contrast, when making a drawing, the to-be-drawn feature and hand position are not visually available at the same time, as one has to look away from the original to look at the copy. Therefore, drawing requires more complex computations of coordinate transformation (object-centered to eye-centered to body-centered; Crawford, Medendorp, & Marotta, 2004b; Kakei, Hoffman, & Strick, 2003; Ogawa, Nagai, & Inui, 2010) and must rely more on internal, memory guidance of hand movements than does tracing (Gowen & Miall, 2006). Any failure in one of these processes may explain the drop in accuracy observed between tracing and traditional copying (Cohen & Bennett, 1997).

3.3.3 EFFICIENCY OF EYE-HAND COUPLING STRATEGY IN DRAWING Several studies investigating eye-hand interactions during a drawing task (Miall & Tchalenko, 2001; Tchalenko et al., 2003; Tchalenko, 2009) found no differences between novices and experts in eye-hand coupling strategies, ruling out the hypothesis of a more accurate visuomotor transformation in drawing experts. In particular, Tchalenko and Miall (e.g. Tchalenko & Miall, 2009) observed a “just-in-time” strategy in both novices and experts: the hand starts moving while the eyes are fixating around the corresponding segment in the original. As suggested by the authors, fixating at a location in the original corresponding to the starting point of the subsequent hand movement enables a more stable reference frame shared by both the visual and the motor system. Furthermore, spatial coordinates of the to- be-drawn feature positions can be encoded relative to the fixation location, and hence relative to the hand starting position. Such strategy may reduce the complexity of the computation needed to transform eye-centered coordinates to body-centered coordinates (Ballard, Hayhoe, Pook, & Rao, 1997) and may therefore increase spatial accuracy of the motor output

Interestingly, Tchalenko and Miall (Tchalenko & Miall, 2009) compared eye-hand interactions in different drawing conditions, and particularly found that hiding the copy (direct blind copy) while leaving the original visible led to good shape accuracy but poor spatial positioning of the depicted stroke. This suggests that eye-movements made on the drawing- in-progress may only have the purpose of visually monitoring the hand positioning, and that the visuomotor mapping of the encoded information and the motor planning of the hand

24 trajectory (defined by the feature’s shape) may be programmed directly from visual information captured when looking at the original.

Altogether, these findings show that drawing skill may not be explained by an improvement of fine motor control or by more accurate visuomotor transformations. As we have seen, in contrast to tracing, drawing may depend more on an internal, memory representation in order to visually guide the hand trajectory (Gowen & Miall, 2006). It is therefore possible that drawing skill relates the ability to build a more accurate internal representation of the objects and its features in memory. However, the accuracy of a memory representation may depend on two factors: a correct initial perceptual encoding (the object needs to be accurately perceived), and the encoding efficiency of this information in memory. Therefore, before considering the role of memory in drawing skill, we will examine first whether drawing skill would relate to a more accurate perception of objects and scenes.

3.4 PERCEPTUAL MECHANISMS RELATED TO DRAWING SKILL

Drawing skill may not be explained by improvement of basic motor and visuomotor skills. Interestingly, Tchalenko (Tchalenko, 2007) tested participants with drawing levels ranging from novices to professional artists, in copying simple, isolated lines. The author varied the type of line to copy (straight or curved), the length (from 10 to 30 cm) and the orientations (horizontal, vertical and oblique). He found no qualitative differences in the accuracy of the tracing between drawing experts and novices. In a follow-up study (Tchalenko, 2009), the author again tested drawing experts and novices in a copying task of a complex figure (a standing nude). He found that drawing experts were more accurate than beginners at both a global level (proportion, aspect ratio) and at a local level (line). Consequently, the author suggested that novices and less trained artists (e.g. art students) may have difficulty to copy lines when they are parts of a more complex array. There are major differences between copying single, isolated lines and lines that are parts of more complex array is: 1/ the set of lines forms a visual context that could affect the perception of individual elements within it, and 2/ each line must be copied according to their position within this visual context. Therefore, an accurate copy of complex figures may depend on a correct perception of individual element properties, unaffected by visual context (Cohen & Bennett, 1997), and on a proper encoding of the spatial organization.

In this section, we will review perceptual mechanisms involved in observational drawing, and we will show that drawing may engage a particular way of looking at objects that may be unusual for those of us who are not trained in drawing.

INTRODUCTION | Perceptual mechanisms related to drawing skill

3.4.1 PERCEPTUAL MODE INVOLVED IN DRAWING Our everyday life is mostly characterized by goal-oriented behaviors whereby we interact with our environment, catching or avoiding objects. In this context, visual perception is mostly driven by action, selecting and processing visual information that is relevant for the subsequent interactions with or avoidance of objects (Hayhoe, Mruczek, & Pelz, 2003). Actions first require identifying and recognizing target objects and assessing their properties (size, shape, orientation, distance). Although the brain is very efficient at doing this, these tasks involve numerous implicit mechanisms and computations not accessible to consciousness (Hochstein & Ahissar, 2002). Vision starts first with the retinal image formed by the light reflected on objects’ surface and projected on the back of the eyes. Because of its projective nature, the retinal image constantly varies with viewing conditions (distance from the object, viewing angle or lightning) so that an infinite number of retinal projections may correspond to a single object. In order to solve this issue and to recognize a particular object, such as a cup of coffee on the breakfast table, the visual brain needs to extract structural information from the retinal input that can be matched to an internal abstract template of what a cup should look like, stored in long-term memory. Such internal representations may include non-accidental, viewpoint-invariant structural properties of the object and their spatial relationships (e.g. a red cylinder with a handle on its side; Biederman, 1987; Kosslyn, 1987; Marr & Nishihara, 1978), possibly extracted from specific viewpoints seen in past experiences of this same object (e.g. Hayward, 2003; Vuilleumier, Henson, Driver, & Dolan, 2002). In that sense, our everyday perception may be more biased toward distal properties of objects, more dependent on our knowledge of objects and viewpoint- invariant, rather than toward the way object properties are actually projected onto our retina (Rock, 1983).

In contrast, observational drawing does not need to consider the object to be drawn according to its identity (e.g. a chair) any more than a photograph does. The object is a particular set of spatially related shapes and properties seen under specific viewing conditions. This requires considering and analyzing the very specific viewing angle, lightning condition and visual appearance of the object. Artists must dissect their own visual percept: considering for example relative sizes of objects as they would need to be represented on a drawing with nearer objects drawn larger than far objects of the same real size; spatial layout of the scene and of objects (relative positions of their features); particular shadings, textures, etc. Hence, they may scrutinize the visual scene at a lower level than the object-level description that characterizes our ordinary conscious percept (Hochstein & Ahissar, 2002). Of course, any human is able to access these lower-level descriptions of visual angle or local luminance when required (e.g. in psychophysical tasks), but making explicit perceptual

26 judgments about them is not a common pastime. In contrast, these analyses of the visual scene and the techniques that successful depict them give artists an extra knowledge of how visual perception may work (Cavanagh, 2005). But does it also modify the way they perceive the scene? Several art historians, psychologists and philosophers have argued that artists may see the world differently, in a more veridical fashion (see Kozbelt & Seeley, 2007 for a review). For instance, the art historian John Ruskin (Ruskin, 1912) was encouraging artists to look at the world through childish eyes, to draw objects as they see them and not according to their knowledge of the object’s properties (the “innocent-eye” hypothesis). Doing so, artists’ perception may be less influenced by the perceptual mechanisms that ordinarily correct for changes in viewing conditions to provide us with a remarkably stable perception of the environment instead of the ever-changing image that falls on the retina – a stability called visual constancy (Carlson, 1962; Rock, 1983; Todorović, 2002). In contrast, artists have to represent 3-D objects on a 2-D plan, depicting for example projected size rather than distal size of object, with more distant objects smaller than closer objects. Does drawing skill reflect a modification of perceptual routines that would enable artists to access more directly their proximal, uncorrected percept? Are they less affected by the visual constancies (Figure 2) that normally correct for distance and lighting, for example, to give us access to shapes and surfaces as they are in the external world?

INTRODUCTION | Perceptual mechanisms related to drawing skill

Figure 2. Perceptual constancies. 1) Size constancy. Left panel: the man standing in the background does not appear as being much smaller than the woman in the foreground. In contrast, when the same man is moved beside the woman, he then appears much smaller. Right panel: both individuals have the same retinal size but are standing at different perceived distances. In this case, the visual system automatically interprets that the farther one should be bigger. This interpretation results in a visual illusion in the present case: the man present in the left image appears taller than the man standing in the right image. 2) Lightness constancy. If a participant were asked to compare the luminance of the squares A and B on the tiled surface, he or she would likely say that square B is brighter than square A, although they have the exact same luminance. To the visual system, the presence of the cylinder’s cast shadow may indicate a decrease in luminance for surfaces present within its limits (square B included). The visual system compensates this decrease by increasing the apparent surface lightness of the square B in order to recover its actual surface property: it is made of light gray material while square A is made of dark gray material. Because both squares A and B have the same actual luminance, square B appears brighter due to this compensation.

3.4.2 INFLUENCE OF OBJECT KNOWLEDGE ON DRAWING SKILL Mitchell and colleagues (Mitchell, Ropar, Ackroyd, & Rajendran, 2005), tested participants in a perceptual and a drawing task both using a parallelogram and a table version of the Shepard illusion (see Fig. 2). First, they found that perceptual adjustment

28 errors were greater when participants had to match the width and length of tables than of similar parallelograms (the table’s surface without the legs). As demonstrated by this perceptual illusion, the presence of pictorial depth cues (occlusion of the table’s legs by the table’s top) automatically results in a 3-D interpretation of the line-drawing by the visual system, which in turn results in a perspective context effect, such as size constancy (Shepard & Judd, 1976; Tyler, 2011). The same effect was present when participants had to copy those same stimuli. Interestingly, they also found that participants who made smaller perceptual errors made also fewer drawing errors when copying those same stimuli, suggesting that the way we perceive objects does have an impact on the way we are going to depict them.

Consistent with this account, studies have shown that drawing accuracy was affected by object knowledge in both children and adults, with objects represented from canonical views (Cutzu & Edelman, 1994) rather than from the actual point of view (Carlson, 1966; Chatterjee, 2004; Glazek, 2012; Mitchell & Taylor, 1999; Picard & Durand, 2005; Taylor & Mitchell, 1997). For instance, Taylor and Mitchell (Taylor & Mitchell, 1997) asked non-artists and art students to copy a slanted disc presented in an absolute dark room (without any perspective cues) and found that shapes drawn by all participants were closer to a disc than to an ellipse (projected shape on the retina). Interestingly, this bias toward circularity was not found when the authors tested a control group of participants unaware of the actual shape of the object (it could have been an actual ellipse). However, the absence of difference between non-artists and art students may suggest that shape constancy might be a fundamental perceptual mechanisms too well established to be overcome by intensive training in drawing.

INTRODUCTION | Perceptual mechanisms related to drawing skill

Figure 3. Drawing accuracy depends on how veridical we perceive the visual context (adapted from Mitchell et al., 2005). (a) Perceptual matching task. Participants were asked to adjust the width and length of a test stimulus (table or parallelogram) to make it matching the target stimulus’ dimensions. The left table appears thinner and longer than the right table, although both tables’ tops have the same width and length. This effect is stronger in the table case than in the parallelogram case. (b) Drawing task. In a second experiment, participants had to copy the same stimuli. In both tasks, perceptual errors were greater in the table condition than in the parallelogram condition, suggesting an effect of depth cues (e.g. occlusion of the table legs by the table’s top) on perceptual judgments and drawing accuracy. However, participants who made fewer perceptual errors also made fewer drawing errors.

This result may be surprising in light of other studies that found a relation between drawing skill and reduced perceptual constancies where more skilled artists were more immune to the effects of shape (Cohen & Jones, 2008; Thouless, 1932) or size constancy (Ostrofsky et al., 2012). However, these findings have not held up. For instance, McManus and colleagues (McManus, Loo, Chamberlain, Riley, & Brunswick, 2011) tested shape constancy in people of various drawing skills with a computer-generated version of Cohen and Jones’ task (Cohen & Jones, 2008), and found no advantage for more skilled participants in their task (even an opposite trend). Also, although Ostrofsky, Kozbelt and Seidel (Ostrofsky et al., 2012) found a smaller effect of size constancy in skilled participants, this was not the case in their shape constancy task. Given these conflicting results, we do not yet know whether drawing skill would be related to modifications of automatic, perceptual mechanisms – whether drawing expertise causes or is related to perceptual expertise.

A first goal of this thesis is to systematically test skilled artists (professional and art students) and non-artists in different perceptual adjustment tasks involving shape, size and brightness constancies. We use these results to determine how skilled artists undo perceptual constancies. In particular, it remains unknown whether drawing skill relates to a more direct access to earlier, uncorrected proximal visual representation, or to visual analyses of the corrected, final percept.

3.4.3 GLOBAL VS. LOCAL VISUAL PROCESSING OF OBJECTS Observational drawing is not a single copy-paste of the entire scene. Instead, it is better described as a sequential process, where the depiction is made progressively, adding new elements to the already drawn ones and sometimes erasing some of them (Gombrich, 1960; Locher, 2010). A critical aspect of drawing accuracy is the correct positioning of each individual feature according to the global organization of the depiction, which must also match that of the original object (Locher, 2010; Tchalenko, 2009). This may require visuospatial and constructional abilities to properly encode and represent part-whole relationships in addition to the ability to focus alternatively on local individual features and then their place within the object’s spatial layout (Chamberlain, McManus, Riley, Rankin, & Brunswick, 2013).

In line with the idea of a relationship between drawing skill and enhanced visuospatial abilities, Kozbelt (Kozbelt, 2001) suggested that artists would be “experts in visual cognition”. His results demonstrated that artists outperformed non-artists subjects in perceptual tasks as well as in drawing tasks (see Fig. 3). The author designed four perceptual tasks without drawing involved: object recognition from out-of-focus pictures, Gestalt completion, visual search for a specific shape embedded in another figure, and mental rotation. In all of these tasks, participants’ performances correlated with their level of drawing skill and experience. It therefore seems that training in drawing can improve visual analyses, but it remains unclear what specifically makes artists better in those tasks. In a more recent study, Ostrofsky, Kozbelt and Kurylo (Ostrofsky, Kozbelt, & Kurylo, 2013) investigated whether the artist’s perceptual advantage could be accounted for by improvements of low-level perceptual grouping mechanisms. They found no advantage for trained artists at this level, again suggesting that drawing skill may relate to an improvement of top-down visual strategies rather than of modifications of bottom-up visual mechanisms involved in object perception.

INTRODUCTION | Perceptual mechanisms related to drawing skill

Figure 4. Art students outperformed novices in high-level perceptual tasks (Kozbelt, 2001). Participants were tested in four perceptual tasks: object recognition from out-of- focus picture (“Focus”), Gestalt completion (“Gestalt”), recognition of embedded shape (“Embedded”) and mental rotation (“Rotation”).

Carlson (Carlson, 1966) tested skilled and novices participants in perspective drawing in perceptual judgment tasks involving either visual illusions (e.g. Müller-Lyer or Sander’s parallelogram illusion) or visual constancies (size and shape) and varied the type of instructions (report the proximal “apparent” or distal “actual” property of the target stimulus). The author found that experts were not overall less affected by either visual constancies or illusions, but were more responsive to changes in instructions in visual constancy tasks, although not in visual illusion tasks. These results support the idea that visual constancies can be to some extent ignored by anyone, but that people more proficient in drawing may be better able to switch between an object-level and a feature-level description of the scene on demand.

Accordingly, it has been suggested that people skilled in drawing would analyze objects and visual scenes differently than novices, focusing more on local aspects rather than on global properties of visual stimuli; which could also explain why drawing skill would relate to a lesser impact of visual context effects (Chamberlain et al., 2013; Drake & Winner, 2011; Drake, 2013). For instance, Chamberlain and colleagues (Chamberlain et al., 2013) tested participants of various drawing skills in different visuospatial tasks, including a Navon shape task, to test whether more skilled participants could better ignore the stimulus’s global aspect. As expected, they found that participants, regardless of their drawing skill, made

32 more errors and had higher response time when reporting local shapes incompatible with the stimulus’ global aspect, whereas this was not the case when they had to report the stimulus global aspect with incongruent local shapes – namely, a global interference effect (Navon, 1977). Interestingly, they found a smaller global interference effect on more skilled participants’ reaction time, although not on their error rate (possibly because of a ceiling effect). The authors concluded that drawing skill may relate to a better ability to ignore the object’s global aspect when focusing on its parts, but not to a more accurate perception of either local or global visual properties of the object.

Taken together, these studies suggest that drawing skill is not related to changes in sensory processing, which would result in a different final percept (Ahissar & Hochstein, 2004; Hochstein & Ahissar, 2002). Bottom-up perceptual mechanisms may be too well established, physiologically and functionally, by our lifelong experience in vision. Instead, drawing skill may depend on the way the information is subsequently parsed and accessed from the final percept. The visual information available to both novices and experts may be identical, but the way it is used may differ.

3.5 VISUAL ANALYSIS AND MEMORY OF OBJECTS STRUCTURES

We have seen that skill in observational drawing may relate to a particular visual analysis of objects. However, most of the studies that investigated this ability used perceptual tasks without drawing involved. The present section aims at identifying what particular visual analyses and memory representations would be required by the drawing process.

3.5.1 PERCEPTUAL CONSTRAINTS IN DRAWING Observational drawing requires identifying and selecting relevant features to render that will trigger recognition of the object despite the limitations imposed by the chosen medium. Drawing experts have a better knowledge of diagnostic features, which results in a more efficient selection of the to-be-drawn features even when the amount of information that can be represented is controlled (Biederman & Kim, 2008; Kozbelt et al., 2010; Ostrofsky et al., 2012). However, even if selecting relevant features is indeed critical, this would not lead to an accurate depiction if the spatial relationships between the selected features within the original figure are not respected in the copy. Encoding and representing these spatial relations may therefore be one of the most critical aspects of drawing skill.

However, observational drawing is a sequential visuomotor task involving many eye and hand movements, to visually encode the object, guide the hand movement when tracing

INTRODUCTION | Visual analysis and memory of objects structures the encoded feature, and to compare both the drawing and the original to detect mismatches. This requires encoding the spatial relationships between features within both the original and the drawing, to track these relationships despite the many changes in the retinal input due to eye-movements and to be able to alternatively transpose the original’s reference frame to that of the drawing, and vice et versa. These aspects are difficult to capture in psychophysical tasks that ask observers only to make perceptual judgments of target stimuli with unconstrained eye-movements. In this thesis, we will use eye-tracking techniques in order to investigate how drawing experts are able to build a representation of the spatial relationships between features from information gathered at each fixation.

3.5.2 ENCODING EFFICIENCY AND THE CHUNKING HYPOTHESIS One hypothesis for the internal representation used for drawing is that it stores all the features and their relative spatial relationships in memory. Temporarily holding spatially integrated features while producing a particular segment could only be achieved with the help of visual working memory (Baddeley, 1992; Fig 5). However, trying to store all the features and their spatial relations would likely overload visual working memory (Alvarez & Cavanagh, 2004; Schneider, 1999; Wheeler & Treisman, 2002).

An alternative hypothesis would be that features themselves are not encoded in visual working memory, but only their spatial organization. This would only require encoding the structure, or spatial schemata, of the original. Similar perceptual and memory mechanisms have been observed in experts chess players and were used to explain their memory performances in recalling numerous pieces positions from a briefly presented chess game (e.g. Chase & Simon, 1973; Gobet & Simon, 1996a). It has been particularly shown that expert chess players encode the spatial configuration of pieces rather than the pieces themselves. This allows identifying and encoding local spatial patterns of pieces, or local structures, that can be stored as perceptual chunks in working memory. Perceptual chunks refer to meaningful sets of related information that can be represented as a unit in visual working memory. For instance, if one is presented with a series of numbers like 149215151789, he or she would be more likely to encode this information as 3 different dates (1492, 1515, and 1789) rather than as 12 individual items. Such encoding results in a greater amount of information held in visual working memory despite the capacity limits. Interestingly, it has been showed that skill in chess is related to the ability to process larger chunks (more items per individual perceptual chunk) and to spatially integrate them according to their spatial relationships into more complex chunks (memory templates) stored in long-term memory (Gobet & Simon, 1998). When needed, particular perceptual chunks can be retrieved from the memory template and held in visual working memory. Such

34 memory mechanisms would therefore allow drawing experts to sequentially integrate local structures encoded from each fixation into a single representation which would therefore be more robust to changes in retinal inputs and changes in reference frames.

Figure 5. Hierarchy of visual memory systems and the chunking process. Once processed from sensory inputs, visual information are briefly held in a sensory storage (“iconic” memory), which is thought to reflect the sustained activity of visual areas. This stage is pre-attentive, so that stored features are not integrated yet. Attended information during the initial encoding would be transferred in visuospatial working memory and held as long as there are manipulated (rehearsed). Only a limited number of memory chunks (meaningful sets of related features) can be simultaneously held and manipulated in visual working memory, which would also depend on their complexity (e.g. high-resolution images take more place than low-resolution images, so that less of them can be stored on a hard drive). Perceptual chunks stored in visual working memory can further be integrated in more complex and abstract representation in memory (e.g. letters can be chunked as words, which in their turn can be chunked as abstract conceptual categories or as sentences).

Evidence for such a mechanism in drawing experts is indirect. First, eye-movements pattern of drawing experts shows a segmentation process absent in novices: experts cut the original in sequences of lines that can be drawn at once (segment), whereas novices draw this segment by the use of multiple lines and hand movements (Tchalenko, 2009). This would suggest a perceptual chunking process. In the case of drawing experts, they seem to cut the original object into chunks of related lines with the rule that every chunk could be drawn by a single hand movement, which may depend on both motor constraints (e.g. stroke direction preferences; Van Sommers, 1984) and geometrical cues (e.g. colinearity of lines or contour curvature; De Winter & Wagemans, 2006; Tchalenko, 2009). Another indirect piece of evidence is Glazek’s finding (Glazek, 2012). He found that drawing experts produce more strokes than novices while covering a smaller spatial extent of the original with their eyes and fixating for shorter durations. Both studies indicate a more efficient encoding during fixation.

INTRODUCTION | Visual analysis and memory of objects structures

However, these studies could not determine what information is encoded during the fixation, particularly in peripheral vision, or how drawing experts could process more information with shorter encoding duration.

In this thesis, we will bring additional evidence for such perceptual chunking mechanism by using eye-tracking studies in order to determine whether drawing experts are better able to integrate structural information across eye-movements, and whether they would have a more efficient (faster processing) of object structure during a fixation. Related to this latter point, it will be important to determine whether drawing experts are able to encode larger chunks (more items in a single unit), particularly by processing larger extent of space at a single glance (larger visual span; Rayner, 1998) as experts in chess do (Reingold, Charness, Pomplun, & Stampe, 2001b).

3.5.3 MEMORY OF OBJECT STRUCTURE We have suggested that experience with drawing may rely on a chunking mechanism, such as that reported for expert chess players but for visual object structure rather than chess board structure This would enable skilled artists to build a more complex memory representation in visual working memory that would be robust to changes in retinal inputs and help them in guiding the production. However, if such representation would be indeed needed, it is not clear what its content would be.

Several studies discussed the role of visual working memory during the drawing process, although these studies did not directly test visual working memory during a drawing task (Cohen, 2005; Glazek, 2012; McManus et al., 2010; Tchalenko & Miall, 2009; Tchalenko, 2009). In particular, drawing has been compared to other visuomotor tasks involving high load on visual working memory (Cohen, 2005; Tchalenko, 2009). These tasks are associated with more frequent eye-movements made toward the source of information to reduce the amount of information needed to be held in visual working memory (e.g. Ballard, Hayhoe, & Pelz, 1995). Indeed, drawing experts make more gaze shifts between the original and the copy than novices, with fixation focused around the features being drawn (Cohen, 2005; Miall & Tchalenko, 2001). What exactly is held in memory during drawing? Recall that if the participant is prevented from seeing his or her hand and the copy while drawing, the drawing still shows good shape accuracy but poor spatial positioning (Tchalenko & Miall, 2009). So, one conclusion is that fixations made on the copy after the tracing has begun only guide the correct positioning of the hand while drawing is programmed directly by visuomotor mapping directly from the original and not from memory (Gowan & Miall, 2006; Tchalenko & Miall, 2009). However, this rather radical claim was made from studies that never directly tested memory during the drawing process (e.g. Tchalenko, 2009, p. 795). Furthermore, if drawing

36 skill would relate to encoding strategies aiming at reducing the amount of stored information to local information only, then it is unclear why drawing experts outperform novices in delayed recognition tasks using complex figures or global spatial arrangements of features (Cohen & Jones, 2008; McManus et al., 2010; Rosenblatt & Winner, 1988).

A final aspect of this thesis will be to directly test visual working memory during a copying task in order to precisely determine what information would survive once the action programmed and with which accuracy this would be stored.

INTRODUCTION | Visual analysis and memory of objects structures

4 EXPERIMENTS

EXPERIMENTS | Visual analysis and memory of objects structures

4.1 COULD THE “INNOCENT EYE” HYPOTHESIS ACCOUNT FOR DRAWING SKILL?

4.1.1 OBJECTIVES AND SUMMARY OF RESULTS In this first study we demonstrated that drawing ability was not related to a better (more veridical) representation of the stimulus, uncorrected by automatic perceptual mechanisms, such as visual constancies. Instead, artists’ success in overcoming visual constancies when producing a drawing arises from the ability to compare the drawing with the original where both are subject to unavoidable effects of constancies. These visual constancies are perceptual mechanisms enabling our stable perception of the world despite constant changes in viewing conditions (Rock, 1983; Todorović, 2002). Our perception is biased toward the real world (distal) properties of objects (size in meters, objective shape) rather than toward the way their properties are projected onto our retinas (proximal). However, both aspects of object property (distal and proximal) can be accessed if necessary (Carlson, 1966; Rock, 1983), as one can either see the building as a 200 meter tall object or as an object that can be covered by the thumb when pointing at it. Artists have to convert the 3-D world into a 2-D description on the paper, hence making depictions more closely matching the proximal, retinal image. Practicing this ability over years may confer on them an advantage of more easily switching their perception toward the proximal percept rather than the distal percept (Carlson, 1966). Previous studies have already investigated the impact of visual constancies (in particular, shape constancy) on drawing accuracy (Carlson, 1966; Cohen & Jones, 2008; Mitchell et al., 2005; Thouless, 1932). However, it is still unclear whether such ability could transfer to other constancies than that of shape (e.g. size and brightness), nor how artists could better access their proximal percept, if they do.

In the two first experiments, we designed perceptual matching tasks involving size and brightness constancies. Participants were told to adjust the property (size or brightness) of a test stimulus (either a cylinder for size, or a luminance patch for brightness) to match it with the perceived property of a target stimulus. We manipulated the presence or absence of visual context (perspective grid for size, cast shadow for brightness) and we measured the ratio between those two conditions as the amount of visual context effect. We particularly insisted on the instructions given to our participants regarding the target visual property to match (proximal rather than distal), as it has been indeed previously shown that artists are more able to switch between these two sorts of percept than non-artists (Carlson, 1966). In a third experiment, we examined whether artists would be better able to overcome visual constancies by measuring our participants access speed to early, uncorrected visual representations in the context of amodal completion. To do so, we designed a visual search

EXPERIMENTS | Could the “Innocent Eye” hypothesis account for drawing skill? task where participants were asked to find a target (notched square) among a variable number of distractors. The notched square could either be in contact with a companion circle, which results in the perception of a square partially occluded by a disc, or disconnected from its companion disc (absence of context). Again, we measured the ratio between the search slopes each participant obtained in both condition as a measure of the impact of visual context.

We found no evidence of a reduced impact of visual constancies in our artists in the first two experiments, nor were artists faster at accessing representations prior to amodal completion in order to find the notched-square in the third experiment. Our findings call into question the ability of artists to directly access earlier, uncorrected visual representations. In particular, in the two first experiments, art students and professional artists actually took about twice as long to make their perceptual judgments, suggesting that the tasks were not easier for them. Secondly, artists were not faster at finding the target within the context of amodal completion in the third experiment. Taken together, our results strongly suggest that artists arrive at “proximal” depictions by applying post-perceptual corrections on the basis of their knowledge about visual constancies and by using comparisons between the drawing and the original, both of which undergo the same corrections for visual constancies.

4.1.2 DO ARTISTS SEE THEIR RETINAS?

This chapter is based on: Perdreau, F., & Cavanagh, P. (2011). Do artists see their retinas? Frontiers in Human Neuroscience, 5(171), 1–10. doi:10.3389/fnhum.2011.00171

Abstract. Our perception starts with the image that falls on our retina and on this retinal image, distant objects are small and shadowed surfaces are dark. But this is not what we see. Visual constancies correct for distance so that, for example, a person approaching us does not appear to become a larger person. Interestingly, an artist, when rendering a scene realistically, must undo all these corrections, making distant objects again small. To determine whether years of art training and practice have conferred any specialized visual expertise, we compared the perceptual abilities of artists to those of non-artists in three tasks. We first asked them to adjust either the size or the brightness of a target to match it to a standard that was presented on a perspective grid or within a cast shadow. We instructed them to ignore the context, judging size, for example, by imagining the separation between their fingers if they were to pick up the test object from the display screen. In the third task, we tested the speed with which artists access visual representations. Subjects searched for an L-shape in contact with a circle; the target was an L-shape, but because of visual completion, it appeared to be a square occluded behind a circle, camouflaging the L- shape that is explicit on the retinal image. Surprisingly, artists were as affected by context as non-artists in all three tests. Moreover, artists took, on average, significantly more time to make their judgments, implying that they were doing their best to demonstrate the special skills that we, and they, believed they had acquired. Our data therefore support the proposal from Gombrich that artists do not have special perceptual expertise to undo the effects of constancies. Instead, once the context is present in their drawing, they need only compare the drawing to the scene to match the effect of constancies in both.

Keywords: art, vision, visual constancy, visual search, scene perception

4.1.2.1 INTRODUCTION Visual perception is our main access to the outside, “distal”, world which we experience consciously at the end of a long chain of processes. The image projected on our retina is the proximal stimulus, the original data on which these processes operate. If we should see the world as it is represented on the retina, objects would change size as they moved toward or away from us, change color as they moved into different lights, be cut into pieces as they moved behind other objects, and jump to and fro every time we moved our eyes. But instead of perceiving this ever-changing world, we have a coherent, invariant visual representation of objects: we experience visual constancy, that is, our conscious percept is to a large extent in accordance with the distal object’s properties whatever the proximal stimulus projected on our retina.

EXPERIMENTS | Could the “Innocent Eye” hypothesis account for drawing skill?

However, visual artists when rendering an object or a scene on a canvas return to a representation that is closer to the proximal image, depicting distant objects as smaller and nearby objects as larger. Clearly, compared to non-artists, artists are able to depict scenes and objects much more accurately. What is the basis of their expertise? One aspect is of course motor skill but the other of interest to us is the ability to see the proximal pattern of light and dark – to ignore the corrections, the visual constancies, underlying our everyday perception. The artist can pick the right dark pigment for depicting an object in a shadow, a pigment much darker than our subjective impression of the object; can make the distant object the correct size even though it is experienced as not very small. A number of studies have addressed these issues (Cohen and Bennett, 1997; Kozbelt, 2001; Cohen, 2005; Mitchell et al., 2005; Kozbelt and Seeley, 2007; Cohen and Jones, 2008; Matthews and Adams, 2008) showing indeed that drawing accuracy is correlated to perceptual performances: subjects who made more accurate drawings also showed less effect of context and visual constancies. According to Kozbelt (2001), artists are “experts in visual cognition.” The present study addresses whether the expertise of visual artists lies in their ability to access their proximal representation better than non-artists. Have years of experience changed their visual processing and their ability to access early levels of representation? Such plasticity in visual processing as a result of visual experience is seen in many contexts (Hubel and Wiesel, 1970; Goldstone, 1998; Ostrovsky et al., 2006; Green and Bavelier, 2008).

The idea that artists have direct access to early representations has been strongly criticized by the art historian Gombrich (1987). Gombrich agreed with Ruskin (1912) that artists do use special techniques to depict the proximal stimulus but he felt that their training could not lead them to get an “innocent eye”: the “innocent eye is a myth” (Gombrich, 1987, p. 251). Instead, “making comes before matching” (Gombrich, 1987, p. 99), and artists have to deal with their biased perception by drawing sketches according to it, and then make corrections in order to match it with the objective model they wish to represent. In this view, image-making is a hypothesis-testing process, a continuous back and forth between production and correction. This “copyist” approach is an alternative explanation for the representational skills of artists. That is, artists may experience the same visual constancies as non-specialists but learn to make corrections in the context of the drawing itself as it progresses. Speciﬁcally, once sufﬁcient context is present in the drawing, they only have to match the sizes and colors they see in their artwork to the perceived sizes and colors they see in the scene being depicted; the similarity of context in both will impose the same constancies.

To examine whether artists have developed visual expertise or copyist expertise, we tested three different constancies: size, lightness, and shape, all of which must be undone

44 or bypassed for figurative artists to create an accurate copy of a scene. Two of the experiments use matching-to-standard tasks while the third is a visual search task. In all of these tasks, we will use context, perspective grids, shadows, and occlusion to trigger the application of visual constancies (Day, 1972; Todorovic, 2002, 2010), and see whether the artists are less influenced by the context than non-artists. If artists are indeed able to access, or recover their initial retinal image (closer to the proximal stimulus), they would be less affected by context than non-artists. However, this finding would not tell us whether the greater accuracy was due to perception that was uncorrected by visual constancies (Ruskin, 1912) or to skill in undoing the corrections (Gombrich, 1987). The critical factor to distinguish these two possibilities is speed: access to the uncorrected proximal image ought to allow for rapid response whereas the reversal of the corrections should require extra time. To test the speed of access, we use a visual search task for partial shapes in occluded or unoccluded presentation (He and Nakayama, 1992; Rensink and Enns, 1998). If artists are able to access the initial uncorrected image then their processing rates for the occluded versions will be more rapid than those of non-artists.

In these experiments, context is introduced in order to trigger the corrections of visual constancies and we assume that, without any instruction, both artists and non-artists would probably experience these context effects to the same degree. However, the subjects were not asked to judge the perceived size, or lightness, or shape, they were asked to ignore the context, to bypass constancy, and report the “real” size or luminance, or shape of the test. This is a critical point in the procedure: subjects are asked explicitly to report what corresponds to their retinal image. Can artists do this better than non-artists?

4.1.2.2 EXPERIMENT 1: SIZE CONSTANCY Size constancy refers to the accurate perception of an object’s size despite the fact that a distant object will have a smaller size on our retina than a near object. In order to provide such a “veridical perception” (Todorovic, 2002), the visual system needs to infer the object’s size by correcting its size on the retina (in visual angle) for the perceived distance (Figure 1). Because size constancy is related to distance perception, it must be directly dependent on the various cues to depth (Leibowitz and Harvey, 1967; Day, 1972). For example, the inﬂuence of monocular cues (perspective grids) on size constancy has been shown in several experiments (Stuart et al., 1993; Aks and Enns, 1996; Bennett and Warren, 2002).

EXPERIMENTS | Could the “Innocent Eye” hypothesis account for drawing skill?

Figure 1 . In the left panel, the man in the background appears to be about the same height as the woman in the foreground. This perception corresponds to visual constancy. However, in the right panel, the man’s image is moved so that he appears to be adjacent to the woman, and the now appears much smaller than he does on the left. This is the correction for distance that underlies size constancy and we, non-artists, are unable to ignore it even though we know that the two images of the man have identical size on the picture plane (measure them to check). Can artists register that the two images of the man have identical size in the picture plane?

Nevertheless, our perception is not limited strictly to corrected distal image; for example, Rock (1983) suggested that we are aware of both retinal size and actual size of the object, even if we generally do not pay attention to retinal size. However, even when asked to judge an object’s retinal size (say, compare a distant building to our thumb held out beside it), there are residual effects of the actual size in the world (Carlson, 1960, 1962). This suggests that artists may be able to access the uncorrected retinal size of objects, ignoring to some extent the real world sizes of the objects; perhaps, they may do this more effectively than non-artists.

In this first experiment, perceived depth was induced by linear perspective cues of a receding hallway in the context condition. Here, size constancy should make the test stimulus look larger in the hallway than when it is seen against the flat grid (Figure 2), and we assume that, without any instruction, both artists and non-artists would probably experience this effect to the same degree. However, the subjects were asked to adjust the size to match the physical size of a standard (presented on a blank field below) as if they were using their

ﬁngers to measure the size directly on the screen. In other words, subjects were encouraged to ignore the context and report the “real” size of the test.

4.1.2.2.1 MATERIALS AND METHODS

4.1.2.2.1.1 SUBJECTS For the three experiments, the subjects were subdivided in three groups: art students, professional artists, and non-artists. The first were recruited from high-ranked Major Art School [n = 9, six females and three males, age = 22 ± 1.7]. Professional artists were recruited from galleries, workshops, and international artists associations [n = 14, nine females and five males, age = 39 ± 12.9]. Non-artists subjects were recruited from the internal network of Cognitive Science (RISC), a database of voluntary subjects, except for two subjects from our laboratory [n = 14, nine females and five males, age = 23 ± 2.8]. The non- artists reported having no particular drawing skills or specific training in visual arts. All subjects had normal or corrected-to-normal vision and those from outside our laboratory were paid 10€ for their participation. They were informed about the purpose of the experiment and were naïve about our hypotheses. They all gave their informed consent before passing the experiment.

Figure 2. Size task conditions. Subjects were asked to adjust the size of the test cylinder so that it matched the actual size of the standard cylinder, imagining that they were using their ﬁngers to judge the size of both cylinders on the screen. There were two randomly counterbalanced conditions: “normal” condition where the cylinder was displayed on a simple 16 × 16 grid, and “context” condition where the cylinder appears in linear perspective represented by a hallway.

EXPERIMENTS | Could the “Innocent Eye” hypothesis account for drawing skill?

4.1.2.2.1.2 MATERIALS All the experiments took place in a dark room and used the same materials. Also, the subject’s head was always held by a chinrest so that his or her eyes were approximately 52 cm from the center of the screen. The stimuli were projected on a 22’’ CRT screen (LACIE, Electron22blue IV), with a resolution of 1024x768 pixels and with a frequency of 100 Hz. The monitor’s luminance was linearized with a gamma correction. The experiments were programmed with MATLAB Psychtoolbox (version 3.0.8), and were run on an Apple computer.

The screen was divided in two equal vertical halves (21°x16°). In the top half (‘standard’), two possible texture gradients could be displayed: a simple 16x16 black line- drawn grid simulating a vertical wall, or a black line-drawn perspective grid representing a hallway with a central perspective (with a unique vanishing point in the center). The targets were two green cylinders, one in each half (see figure 2). The cylinders were drawn with Adobe Photoshop CS4, and their color saturation was set at 10% in order to avoid any distracting salience. All the visual elements (texture gradients and cylinders) were presented against a white background.

4.1.2.2.2 PROCEDURE Participants were told, “Adjust the size of the cylinder, at the bottom of the screen, so that it matches the size of the standard cylinder at the top. Make your adjustment as if you were using your ﬁngers to measure the size directly on the screen.” They pressed the right arrow on the keyboard to increase the lower cylinder’s size, or the left arrow to decrease it, and then pressed the space button to register the setting. There was no time pressure but the time they took to make their setting was recorded.

The standard cylinder displayed in the top half of the screen could be presented either on a simple grid or on a texture gradient representing a hallway. The former corresponded to the normal condition, while the latter corresponded to the context condition. The two conditions were presented equally often with the order randomized across trials. The standard cylinder could have six possible heights (1.5˚, 1.6˚, 1.7˚, 1.8˚, 1.9˚, and 2˚ of visual angle), which were randomized across trials, and the test cylinder could begin randomly either 50% smaller or bigger than the standard.

Each participant started the experiment with a block of 10 practice trials. The conditions in the test block were the texture gradient (normal/context) and the possible heights of the referential cylinder. There were 5 trials per condition for a total amount of 60 trials for the test bloc (5 × 6 × 2).

4.1.2.2.3 RESULTS Subjects’ settings increased proportionally with the standard size and we summarized each subject’s settings by their means across the six standard sizes. We then computed a ratio between the context mean response and the normal mean response for each subject (group mean ratios are plotted in Figure 3). These ratios are a measure of the context effect on the subject’s judgment. Ratios close to 1 mean that there was no effect of the context, while ratios signiﬁcantly greater than 1 would suggest such an effect, that is, that subjects have overestimated the standard size when presented in the hallway context.

Figure 3 . Group mean ratios. Ratios were computed by dividing the subject’s mean response in the context condition by that obtained in the normal condition. The art students showed a numerically smaller ratio, but this difference was not significant. Nevertheless, all ratios were significantly greater than 1, demonstrating the presence of significant constancy effects in all subjects.

We ran a one-way ANOVA on those ratios with Groups (non-artist, art students, professional artists) as factor. This test showed no significant difference in the effect of context vs. normal conditions across groups [F (2,34) = 0.37, p = 0.69]. Nevertheless, all ratios were significantly greater than 1 [t (36) = 6.36, p < 0.000]. The average ratio was 1.08, where a ratio of 1 would indicate no effect of context. There was therefore no evidence in our results suggesting that artists are better than non-artists at ignoring context in accessing stimulus size. One of our other questions was whether artists’ performances would vary with experience. To address that point, we analyzed the correlation between the context effect expressed as the ratio described above and subjects’ years of art experience. We fixed non-artists’ experience to 0, since they were not supposed to have

EXPERIMENTS | Could the “Innocent Eye” hypothesis account for drawing skill? followed an art training, and used the self-reported years of art training as the other variable in the correlation. The correlation was not signiﬁcant (Pearson’s r = 0.08, ns).

Finally, we analyzed the response time for each subject to evaluate the effort the subjects put into making their settings in each settings in each condition. A longer time would suggest more effort. We found a signiﬁcant main effect of Groups [F (2,219)= 22.59, p < 0.001, η2 = 0.17], as well as a main effect of the condition [F (1,219) = 5.89, p < 0.016, η2 = 0.03], but no interaction between Condition and Groups [F (2,219) = 0.357]. A Post hoc analysis showed that surprisingly, art students, like professional artists, spent more time on each trial, 15.37 and 15.95 s, respectively, almost twice as much as non- artists 8.60 s (both p < 0.001). There was no difference between art students and professional artists. This result is the opposite of our expectation that artists would ﬁnd this task easier.

In summary, size perception was influenced by visual context for all subjects, showing an increase in the estimated size of the standard by in an average of 8% in the context condition compared to the normal condition. We also found no correlation between the degree of context’s influence and the subject’s experience, suggesting that experience and training do not play a crucial role in artists’ performance. In sum, we find no evidence of an advantage for artists in ignoring context when judging object size.

The instructions were of critical importance in this task: if we had asked subjects to match the apparent size, we would expect that size constancy would apply equally to all, independently of their art training. But instead, we were encouraging subjects to ignore the context and evaluate the size of the standard and comparison as if they were measuring them on the screen with their ﬁngers. Our adjustment procedure also allowed subjects time to engage various strategies; this is of particular interest to us as it should bring into play explicit strategies that artists have learned in drawing class as well as the implicit ones acquired through long practice.

Despite these aspects of the experiment that should have favored the artists if they did have special perceptual expertise, we found that the artists were as bound to the context effects as non-artists. Moreover, response time analysis showed that both art students and professional artists spent much more time on each trial than non-artists. We had expected artists to take less time, given their expertise. This opposite result suggests that the artists felt some pressure, as experts in visual perception, to perform well on these tasks, to engage the strategies that they had been taught to correct size perception and to overcome context effect. But despite the instructions to ignore context and despite the longer duration the artists spent on the task, they showed the same extent of constancy as non-artists.

4.1.2.3 EXPERIMENT 2: LIGHTNESS CONSTANCY We perceive objects via the light they reflect back to our retina. The received light is determined by two components: the object’s surface reflectance and the illumination falling on it. The reflectance corresponds to the proportion of the incident light that is reflected at different wavelengths of the spectrum and fully depends on the surface material. It is a property of the object and remains constant whatever the intensity or wavelength distribution of the illumination falling on the object. The amount of light arriving at the retina (the proximal property) is the product of the object’s reflectance (its “color,” the distal property) and the illumination. Here we will focus on achromatic property of the object’s surface – whether it is light or dark, and in the case of the achromatic test patches we use, white, gray, or black. We will use “lightness” as the perceived reflectance (white vs. black surface) and “brightness” or luminance as the perceived luminance (the product of illumination and reflectance). According to those definitions, lightness constancy designates the invariance of the surface’s perceived reflectance despites changes in illumination (Gilchrist, 1988; Moore and Brown, 2001).

Figure 4. Lightness constancy and shadows (Adelson, 1993). Squares A and B have identical luminance as shown by the vertical gray stripes that contact both in the right hand panel. However, B appears to lie in a shadowed region indicating a reduction in illumination. Once the visual system compensates for the illumination difference, B appears to be a lighter (whiter) surface than A.

To recover the surface reﬂectance of an object, most authors assume a process that can discount the illumination falling on it. To do so, the visual system must estimate the illumination. A number of proposals have been made for this process (Gilchrist, 1988, 2006; Adelson, 1993, 2000; Arend and Spehar, 1993a,b; Agostini and Galmonte, 2002). Although lightness constancy has often been explained in terms of low-level mechanisms (simultaneous contrast effect caused by lateral inhibition in retina’s ganglion cells), it now appears that in some cases, a high-level computation of spatial relationships of surfaces and light is required. For example, a cast shadow on a surface can be recognized by the visual

EXPERIMENTS | Could the “Innocent Eye” hypothesis account for drawing skill? system because it is darker, its borders are unrelated to object borders, the surrounding texture continues into the shadow area with a reduction of luminance but not contrast, and it appears to have no volume of its own (Cavanagh and Leclerc, 1989). Thus the visual system would attribute change of luminance within the shadow limits to a change in illumination, not reﬂectance (Gilchrist, 1988).

However, a painter can only vary the reflectance of the paint used to depict the object and so this one pigment must correspond to the luminance coming from the real object where the luminance is the product of the object’s reflectance and the illumination falling on it. Can normal observers make these luminance judgments with any accuracy (brightness) – how well could they pick a paint to match it? For instance, when a cast shadow falls on a test surface it leads the observer to perceive the object’s surface as lighter (Figure 4). Can artists ignore the perceived reflectance and “see” the actual luminance any better than normal observers?

To examine this we introduce a cast shadow into a simple scene (Figure 5) where lightness constancy should make the test stimulus look lighter, more white, when the shadow falls on it even though its luminance remains the same. We assume that, without any instruction, both artists and non-artists would probably experience this effect to the same degree. However, the subjects were not asked to judge the perceived surface lightness (light or dark) but to judge the amount of light as if the shadow were not present or they could look at the gray patch through a tube. In other words, subjects were encouraged to ignore the context, to bypass lightness constancy and report the “real” luminance of the test.

4.1.2.3.1 MATERIALS AND METHODS

4.1.2.3.1.1 STIMULI For this experiment, the screen was divided in two vertical halves having the same height and width (21˚ × 16˚). In those two halfscreens were displayed two identical boards textured with a wood surface and on which a piece of wood shaped as a cylinder was lying, each of them was made with Adobe Photoshop CS4. The wood surface’s average luminance was 9.60 ± 0.12 cd/m2 (mean and SD), while the white background’s luminance was 68.4 cd/m2.

On the top board that served as standard, a cast shadow was rendered to correspond to the effect of a light source on the right. Within the shadow, the wood surface’s luminance was 3.04 ± 0.10 cd/m2 and then rose gradually to the adjacent value to simulate a shadow penumbra. Also, the shadow could have two possible locations covering

52 or not the ellipse position. The target stimuli were two ellipses (2˚ × 1.5˚) colored with middle gray and were presented with the same luminance whether or they fell in the shadow region.

Figure 5 . Brightness task conditions. The task was to adjust the brightness (luminance) of the test ellipse (B and D) so that it corresponded to the actual brightness of the standard ellipse (A and C). Two conditions were randomly presented to the subject: the “normal” condition where the standard was outside the shadow, and the “context” condition where the standard was within the shadow.

4.1.2.3.1.2 PROCEDURE The subjects were asked to adjust the luminance of the test ellipse, so that it corresponded to the actual luminance of the standard ellipse. More particularly, subjects were told “adjust the luminance of the test ellipse, at the bottom of the screen, so that it matches the luminance of the standard ellipse, that at the top. Focus on the standard ellipse’s inside, as if there was no cast shadow, and ignore the context of the scene”. They pressed the right arrow for increasing the luminance and the left arrow for decreasing it. Once the subject was satisﬁed with the adjustment he or she pressed the space key to register the choice. The subject had all the time he or she wanted to give a response.

The standard ellipse could have six possible luminance levels randomized across the trials (14, 16.5, 19, 21.8, 24.6, 27.6; values given in cd/m2), while the test ellipse’s luminance could be initially and randomly (before the subject’s adjustment) either 25% smaller or bigger than the standard luminance. On the half of the trials, the standard ellipse was outside the shadow, and in the other half, it was inside. The ellipse’s position was randomized across the trials.

EXPERIMENTS | Could the “Innocent Eye” hypothesis account for drawing skill?

Each subject started the task with a block of 10 practice trials to ensure that he or she had understood instructions. Conditions that composed the test block were the six possible luminance levels of the standard ellipse and the two positions of the shadow. There were 5 trials per condition, and so 60 trials in the test block (5 × 6 × 2).

4.1.2.3.2 RESULTS As in the first experiment’s analyses, we averaged the mean response for each subject over the stimulus conditions and computed a ratio between the mean in the context and normal conditions (Figure 6). A one-way ANOVA was run on the individual ratios with Groups (non-artists, art students, professional artists) as factor. There was no significant main effect of group [F (2,34) = 1.65, p = 0.21]. Nevertheless, all ratios were significantly greater than 1 [t (36) = 14.48, p < 0.000]. The average ratio was 1.35, where a ratio of 1 would indicate no effect of context and a ratio of 3.16 would indicate complete lightness constancy. As in the first experiment, we asked whether context’s effect on perceptual performance, quantified by ratios, varies with experience. The correlation between ratios, expressing the context effect, and individual self-reported years of art training was not significant (r = 0.02, ns).

Finally we analyzed the subjects’ response time to evaluate the effort the subject made to perform the task. We found a signiﬁcant main effect of Groups [F (2,219) = 18.91, p < 0.001, η2 = 0.15], and a signiﬁcant main effect of condition [F (1,219) = 25.53, p < 0.001, η2 = 0.10]. There was no interaction between both factors. Post hoc comparisons showed that non-artists spent less time (9.67 s) than art students (11.44 s, p < 0.003), and the professional artists were slower still (14.31 s, p < 0.03).

Our present ﬁndings revealed that non-artists, art students, and professional artists were all strongly affected in their brightness (luminance) judgment when a cast shadow was overlapping the position of the standard ellipse. Subjects perceived the standard about 30% brighter than it was, and thus showed a strong effect of lightness constancy (all ratios were signiﬁcantly greater than 1) despite being asked to ignore the shadow context.

As was the case for the size task, art students and professional artists again took signiﬁcantly longer than non-artists to make their setting on each trial, suggesting that they put more efforts into doing the task. Nevertheless, this extra effort, and their substantial expertise did not allow them to overcome lightness constancy. Finally and consistently with our ﬁrst experiment’s results, we found no correlation between the effect of context and the subject’s art experience.

Figure 6 . Group mean ratios for brightness. Both art students and professional artists had numerically smaller ratios than non-artists, but these differences were not signiﬁcant. Nevertheless, their ratios were about 1.32, showing a strong inﬂuence of visual context in all cases.

4.1.2.4 EXPERIMENT 3: AMODAL COMPLETION Amodal completion, another instance of perceptual constancy (e.g., Rock, 1983), is a phenomenal completion of an object’s shape even though some of its parts are occluded by another, intermediate object (e.g., Kanizsa, 1979). Despite the lack of information concerning the occluded parts of the far object, our perception of this object seems to remain complete so that, even if the object is separated into two visible parts by the occluder, we know that the different parts belong to the same object (Kanizsa, 1985). These completion phenomena have been explained in terms of Gestalt conﬁguration laws, such as collinearity (good continuation, e.g., Kellman and Shipley, 1991), similarity, and so forth. Such laws are largely implemented by low-level mechanisms (e.g., edge detection, line orientation, and size discrimination in V1; problem-solving of “border-ownership” in V2 complex cells, e.g., Bruno et al., 1997; Rensink and Enns, 1998; Tse, 1999; Wolfe and Horowitz, 2004).

The processing of visual shape proceeds principally from an analysis of the parts (mosaic stage) to that of the whole (completion stage) where independence from vantage point and completion of missing details emerge. Surprisingly, our conscious access to the object does not seem to follow the same sequence, but rather the reverse (Hochstein and Ahissar, 2002). Several visual search studies have demonstrated that the individual parts of an object are accessed after the percept of the whole object, even when the whole object is not

EXPERIMENTS | Could the “Innocent Eye” hypothesis account for drawing skill? presented (it is partially hidden, He and Nakayama, 1992; Rensink and Enns, 1998; Wolfe and Horowitz, 2004). For example, He and Nakayama (1992) reported that searching for an L- shape is more difﬁcult when it appeared touching an adjacent square. In this case, subjects seem to see not an L-shape but a square completed behind the occluder, thus camouﬂaging the L (Figure 7). Similar results have been found by Rensink and Enns (1998), where searching for a notched square touching a circle led to greater reaction times and to search slopes that were steeper than when it was isolated.

Figure 7. Amodal completion. When the L-shape touches the square, it is harder to ﬁnd as it is no longer seen as an L but as the visible part of an occluded square. The T-junctions (here, circled) ordinarily suggest the presence of occlusion, and thus presence of depth. On the basis of these cues, the visual system extends the contours until they meet together (showed with the dashed lines) to form a square.

If object-level descriptions are the first representations available to conscious perception (Tse, 1999; Lee and Vecera, 2005), any task that requires access to an object’s parts requires that the object be “unbundled,” a step that requires extra time (Hochstein and Ahissar, 2002). Can visual artists better ignore the completed form of the object’s representation and then access the “mosaic” image that would be present on our retina? In our task, subjects were instructed to locate the notched square (Figure 8) so, if the notch contacted the adjacent circle, it would normally be completed and appear as a partially hidden square, camouflaging the notched square shape. If artists have any special expertise in accessing early representations, prior to the completion step, they should find these targets faster than non-artists.

4.1.2.4.1 MATERIAL AND METHODS

4.1.2.4.1.1 STIMULI

Figure 8. Shape task, stimuli, and conditions. The target was a notched square that could be either green or red. In “normal” condition, the target was free, while distractors were Pacman- like circles with a square as companion. In “context” condition, the target was bounded with an “occluding” circle, whereas distractors were squares overlapping a circle. In both conditions, target, and distractors had the same overall size, and there were six isolated circles as supplementary distractors to prevent subjects from searching for a circle overlapping a square.

We designed a visual search task based on Rensink and Enns’ (1998) experiment using amodal completion. The target was a notched square generated by subtracting a circle shape overlapping a square (see Figure 8) and that could possibly be either red or green. For both colors we decreased the saturation by 90% so that neither seemed more salient than the other while remaining discriminable. The distractors were circles with a missing quarter sector (which was generated by subtracting a square shape overlapping a circle), which could also be green or red of the same saturation as the target. An item could accompany the target as the distractors. This added item was a green/red circle for

EXPERIMENTS | Could the “Innocent Eye” hypothesis account for drawing skill? the target, whereas it was a green/red square for the distractors. Depending on the condition, those paired items had a speciﬁc spatial relationship, either adjacent (mosaic condition) or touching (occlusion condition). The overall size spanned by the pair was 1.5˚ in the mosaic condition (notched square separated from the accompanying circle), and of 1.13˚ in the occlusion condition (notched square touching with the circle).

All the items were projected in a 12˚ × 8˚ visual array centered on the screen. Their position was randomly distributed within a 6 × 4 invisible grid. The set number was randomly chosen between 2, 8, or 12 items, and all the displayed elements were jittered by ±0.5˚ to avoid the item collinearity that could help the subject to ﬁnd the target. To avoid alternative cues to the target pair, we added six isolated circles that were either green of red. The circles’ size was approximately 0.77˚.

4.1.2.4.1.2 PROCEDURE The subject had to ﬁnd a speciﬁc target presented among a set of distractors. A target was present on all trials, but could have one of two colors: red or green. The subject had to report the color of the target by pressing the “Z” key on the keyboard if the target was red, or the “N” key if it was green. Subjects were asked to use their two hands, one per key.

The target could have two different orientations: either upright or upside-down. In the mosaic condition, the target was isolated, that is not bounded to another item, whereas in the occlusion condition, the target was attached to a circle so that it appeared as a square occluded by a circle. In the former condition, the distractors were a Pacman-like shape accompanied by a square, while in the latter condition they were a circle occluded by a square. Distractors were designed so that the shapes of those of the mosaic condition corresponded to those of the occlusion condition.

The task was divided in a practice block and a test block. The practice block consisted of 30 trials to ensure that the subjects had well understood the instructions and that they were able to discriminate the colors (green/red). The test block was designed as follows: at the beginning of each trial a black ﬁxation cross was displayed at the center of the stimulus array for 1000 ms and the subject had to look at it. After its disappearance, the items were displayed for a maximum of 12 s, the time interval within which the subject had to respond. If the subject took too much time to respond, the message “too long” appeared and the experiment moved to the next trial.

The subjects had to respond as quickly as possible but keep the error rate below 10%. Each time they made an error, feedback including the current error rate was shown

(computed on the basis of the total number of the errors they made over the total number of trials). Their reaction times were the dependent variable we measured.

The conditions were the spatial relationship between the target and its companion- item (mosaic/occlusion), the number of items (2, 8, or 12), the target’s color (green/red), and the target’s orientation (upright, up-down). There were 15 trials per condition, and hence 360 trials per subject (2 × 3 × 2 × 2 × 15). Those 360 trials were divided into two equal parts of 180 trials, and a short break between them was proposed to the subject.

4.1.2.4.2 RESULTS We first analyzed the reaction times of the subject as a function of the number of displayed items (target and distractors) and we then computed linear regression slopes (Figure 9). The linear regression slopes show that subjects’ reaction times linearly increased with the number of items (R2 = 0.67 ± 0.01) and that the slopes in both context (occlusion) and no-context (mosaic) conditions were steep, with an average of 178 ms/item for the context case, and 89 ms/item for the no-context case. This difference between conditions was significant [F (1,34) = 173.47, p < 0.001, η2 = 0.84], showing a strong effect of context; however, there was no effect of Groups [F (2,34) = 1.74, p = 0.19] or interaction between Groups and Conditions [F (2,34) = 1.31]. A similar pattern of results held for the intercepts of these linear regressions. Because of the absence of the group effects and of interactions in the regression analysis, we could proceed to an analysis of the mean response times, as we had in the two previous experiments, calculating a ratio between mean in the context conditions and in the normal condition for each group (Figure 10). We ran a one-way ANOVA on the individual ratios with Groups (non-artists, art students, professional artists). This analysis revealed no main effect of Groups [F (2,34) = 0.30, p = 0.74]. This result is consistent with the absence of interaction between Groups and Conditions in the slope and intercept analyses. It suggests once again that artists (students and professionals) were not better than non-artists at accessing the raw image data of the target’s L-shape. Nevertheless, again all ratios were significantly greater than 1 [t (36) = 17.88, p < 0.000] indicating a strong effect of context. The average ratio was 1.50, where a ratio of 1 would indicate no effect of context.

EXPERIMENTS | Could the “Innocent Eye” hypothesis account for drawing skill?

Figure 9. Reaction times as a function of number of items in display with group regression slopes and intercepts for normal and context conditions. While non-artists and art students did not show differences between their slopes, professional artists were numerically slower in both conditions. But this was not signiﬁcant. No main effect of groups was found for either the slopes or the intercepts.

In order to address the question whether artists’ ability to overcome the effect of context can be explained by their years of art training, we analyzed the correlation between the individual ratios and the individual self-reported experience. As in the two ﬁrst experiments, we found no correlation between ratios and subjects’ experience (r = −0.13, ns).

Visual search tasks allow us to quantify approximately the time that attention spends on every visual object (e.g., Treisman and Gormican, 1988; Wolfe, 1998; Wolfe and Horowitz, 2004; Nakayama and Martini, 2011). Previous articles have shown that accessing the visible part of an occluded object takes more time than when the partial shape is isolated (He and Nakayama, 1992; Rensink and Enns, 1998). Consistent with these earlier results, we also ﬁnd that visual search for a notched square was slower when the notch was contacting a circle than when it was isolated indicating a strong effect of context even though subjects were instructed to ignore it and look for L-shapes. Finally, as in the two ﬁrst experiments, ratios between the mean response times in the context and normal conditions did not correlate with subject’s experience.

Figure 10. Group mean ratios for shape. Neither professional artists nor art students showed different ratios from those of non-artists. All the subjects were similarly affected by the “occlusion” condition where the target appeared to be a visible part of an occluded square.

4.1.2.5 GENERAL DISCUSSION Visual constancies, such as those of size, lightness, or shape, are known to depend on both low-level, automatic mechanisms and high-level, attentive processing. Our conscious perception emerges with appropriate corrections for the context in the scene (Hochstein and Ahissar, 2002; Ahissar and Hochstein, 2004). This makes sense since we need to recognize objects for what they are, bypassing the particular details of how they arrived on our retina. Although this top-ﬁrst strategy may be useful for our action in everyday life, visual artists have different goals. They must capture exactly those low-level details that broadly match what lands on our retina. Our present study asked whether visual artists like painters and draftsmen can really access this proximal representation or if they are as much affected by visual context and visual constancies as non-artists, even when asked explicitly to ignore context. One could expect that the intensive training of artists might modify the functional organization of the visual brain to allow artists faster access to the early visual information that they need to reproduce in their artwork.

Indeed, several previous studies have reported that visual artists outperformed non- artists in many visual tasks: mental imagery (e.g., Calabrese and Marucci, 2006), object recognition, visual search for embedded shape and Gestalt completion (Kozbelt, 2001). Other studies have shown that artists were also less inﬂuenced by shape constancy in a

EXPERIMENTS | Could the “Innocent Eye” hypothesis account for drawing skill? drawing task, as well as in a perceptual task (Mitchell et al., 2005; Cohen and Jones, 2008). Both Mitchell et al. (2005) and Cohen and Jones (2008) have related reduced effects of shape constancy to drawing accuracy. All those ﬁndings would suggest that, because artists are more accurate in depicting objects, they should be less inﬂuenced by their conceptual knowledge, and perhaps they would rely more on their present raw, early level representation than on their past knowledge.

However, the results of our three experiments, two matching-to-standard tasks and one visual search task, showed that art students and professionals do not differ from non- artists in their ability to ignore perceptual context. Indeed, in all of the three tasks, all the groups’ ratios were significantly greater than 1, showing a significant effect of visual context on their settings. In the first two cases, we found that judgments for size were shifted an average 8 and 35% from veridical by the context (perspective and cast shadow). In the third, the amodal completion context slowed visual search by 50%.

These results argue against theories that suggested that artists’ drawing accuracy is solely due to perceptual expertise. Moreover, all three experiments showed similar, significant effects of context for all groups even though the subjects were instructed to ignore context. There is no evidence here for plasticity in the visual systems of artists. It is possible that we might find some significant differences between artists and non-artists if we had more than the 23 artist and 14 non-artist subjects we tested here; or if we changed our tasks and insisted even more strongly that the subjects ignore the context and report what was on the screen. However, even so, there would not be much joy for those who would want to see artists with an access to early representations. Our data did show significant large effects of context and the best the artists did at reducing this was a non-significant decrease in context effect of about 10% compared to the non-artists’ ratio in the second experiment (lightness). This is far from the 100% reduction that would be required to be able to paint based on “seeing the proximal image.”

Although there is little evidence of any visual system plasticity from all those years of training, there is evidence that their training did affect their performance in the matching tasks but in a different way: they took a very long time to make their settings compared to non-artists. The tasks were not easier for them as we would expect if they had special perceptual expertise. It suggests instead that artists may have found the tasks a personal challenge to their self-image as artists and so they spent more time, perhaps trying to apply speciﬁc strategies that they had learned to deal with depicting size and lightness. But to no avail. The visual search task also showed no advantage for artists, again giving no support to the possibility of a direct, more rapid access to a low-level visual representation.

According to the Gombrich (1987) model of schemata, artists act as copyists, starting with a rough approximation to the scene they are painting. They then compare the depiction with the original and make corrections so that they look the same. This interpretation of the skills of artists does not require them to “see” their retina, the proximal stimulus. Yes, they may make initial errors in selecting a paint, having chosen a value that is more in line with what they “see,” affected as it is by visual constancies. They can quickly correct it once it is in play on the canvas and subject to the same constancies from the context surround it on the canvas, just as the original object is surrounded by its context in the world.

Nevertheless, our results have only examined perceptual factors. In contrast, the visual arts are not only visual but also motor as they involve the drawing task itself. Isolating the perceptual factor allowed us to argue against perceptual expertise as a contributing factor to the difference in drawing skills between artists and non-artists. However, the expertise of visual artists may only emerge in tasks that call on artists’ to actually produce works of art. Further research should assess the role of visual factors in tasks where artists produce artworks.

4.1.2.6 ACKNOWLEDGMENTS This research was supported by a Chaire d’Excellence grant to Patrick Cavanagh.

4.1.2.7 REFERENCES

Adelson, E. H. (1993). Perceptual organization and the judgment of brightness. Science 262, 2042–2044.

Adelson, E. H. (2000). “Lightness perception and lightness illusions,” in The New Cognitive Neurosciences, Vol. 3, ed. M. Gazzaniga (Cambridge, MA: MIT Press), 339–351.

Agostini, T., and Galmonte, A. (2002). Perceptual organization overcomes the effects of local surround in determining simultaneous lightness contrast. Psychol. Sci. 13, 89–93.

Ahissar, M., and Hochstein, S. (2004). The reverse hierarchy theory of visual perceptual learning. Trends Cogn. Sci. (Regul. Ed.) 8, 457–464.

Aks, D. J., and Enns, J. T. (1996). Visual search for size is influenced by a background texture gradient. J. Exp. Psychol. Hum. Percept. Perform. 22, 1467–1481.

Arend, L. E., and Spehar, B. (1993a). Lightness, brightness, and brightness contrast: 1. Illuminance variation. Percept. Psychophys. 54, 446–456.

EXPERIMENTS | Could the “Innocent Eye” hypothesis account for drawing skill?

Arend, L. E., and Spehar, B. (1993b). Lightness, brightness, and brightness contrast: 2. Reflectance variation. Percept. Psychophys. 54, 457–468.

Bennett, D. J., and Warren, W. (2002). Size scaling: retinal or environmental frame of reference? Percept. Psychophys. 64, 462–477.

Bruno, N., Bertamini, M., and Domini, F. (1997). Amodal completion of partly occluded surfaces: is there a “mosaic” stage? J. Exp. Psychol. Hum. Percept. Perform. 23, 1412.

Calabrese, L., and Marucci, F. S. (2006). The influence of expertise level on the visuo-spatial ability: differences between experts and novices in imagery and drawing abilities. Cogn. Process. 7, 118–120.

Carlson, V. R. (1960). Overestimation in size-constancy judgments. Am. J. Psychol. 73, 199.

Carlson, V. R. (1962). Size-constancy judgments and perceptual compromise. J. Exp. Psychol. 63, 68–73.

Cavanagh, P., and Leclerc, Y. G. (1989). Shape from shadows. J. Exp. Psychol. Hum. Percept. Perform. 15, 3–27.

Cohen, D. J. (2005). Look little, look often: the influence of gaze frequency on drawing accuracy. Percept. Psychophys. 67, 997–1009.

Cohen, D. J., and Bennett, S. (1997). Why can’t most people draw what they see? J. Exp. Psychol. Hum. Percept. Perform. 23, 609–621.

Cohen, D. J., and Jones, H. E. (2008). How shape constancy relates to drawing accuracy. Psychol. Aesthet. Creat. Arts 2, 8–19.

Day, R. H. (1972). The basis of perceptual constancy and perceptual illusion. Invest. Ophthalmol. 11, 525–532.

Gilchrist, A. (2006). Seeing Black and White. New York: Oxford University Press.

Gilchrist, A. L. (1988). Lightness contrast and failures of constancy: a common explanation. Percept. Psychophys. 43, 415–424.

Goldstone, R. L. (1998). Perceptual learning. Annu. Rev. Psychol. 49, 585–612.

Gombrich, E. H. (1987). Art and Illusion: A Study in the Psychology of Pictorial Representation. Oxford: Phaidon.

Green, C. S., and Bavelier, D. (2008). Exercising your brain: a review of human brain plasticity and training-induced learning. Psychol. Aging 23, 692–701.

He, Z. J., and Nakayama, K. (1992). Surfaces versus features in visual search. Nature 359, 231–233.

Hochstein, S., and Ahissar, M. (2002). View from the top: hierarchies and reverse hierarchies in the visual system. Neuron 36, 791–804.

Hubel, D. H., and Wiesel, T. N. (1970). The period of susceptibility to the physiological effects of unilateral eye closure in kittens. J. Physiol. (Lond.) 206, 419–436.

Kanizsa, G. (1979). Organization in Vision: Essays on Gestalt Perception. New York: Praeger.

Kanizsa, G. (1985). Seeing and thinking. Acta Psychol. (Amst.) 59, 23–33.

Kellman, P. J., and Shipley, T. F. (1991). A theory of visual interpolation in object perception. Cogn. Psychol. 23, 141–221.

Kozbelt, A. (2001). Artists as experts in visual cognition. Vis. Cogn. 8, 705–723.

Kozbelt, A., and Seeley, W. P. (2007). Integrating art historical, psychological, and neuroscientific explanations of artists’ advantages in drawing and perception. Psychol. Aesthetics Creativity Arts 1, 80–90.

Lee, H., and Vecera, S. P. (2005). Visual cognition influences early vision: the role of visual short-term memory in amodal completion. Psychol. Sci. 16, 763–768.

Leibowitz, H., and Harvey, L. O. Jr. (1967). Size matching as a function of instructions in a naturalistic environment. J. Exp. Psychol. 74, 378.

Matthews, W. J., and Adams, A. (2008). Another reason why adults find it hard to draw accurately. Perception 37, 628–630.

Mitchell, P., Ropar, D., Ackroyd, K., and Rajendran, G. (2005). How perception impacts on drawings. J. Exp. Psychol. Hum. Percept. Perform. 31, 996–1003.

EXPERIMENTS | Could the “Innocent Eye” hypothesis account for drawing skill?

Moore, C. M., and Brown, L. E. (2001). Preconstancy information can influence visual search: the case of lightness constancy. J. Exp. Psychol. Hum. Percept. Perform. 27, 178–194.

Nakayama, K., and Martini, P. (2011). Situating visual search. Vision Res. 51, 1526–1537.

Ostrovsky, Y., Andalman, A., and Sinha, P. (2006). Vision following extended congenital blindness. Psychol. Sci. 17, 1009–1014.

Rensink, R. A., and Enns, J. T. (1998). Early completion of occluded objects. Vision Res. 38, 2489–2505.

Rock, I. (1983). The Logic of Perception. Cambridge, MA: MIT Press.

Ruskin, J. (1912). Elements of Drawing and Perspective. London: J.M. Dent & Sons.

Stuart, G. W., Bossomaier, T. R., and Johnson, S. (1993). Preattentive processing of object size: implications for theories of size perception. Perception 22, 1175–1193.

Todorovic, D. (2002). Constancies and illusions in visual perception. Psihologija 35, 125– 207.

Todorovic, D. (2010). Context effects in visual perception and their explanations. Rev. Psychol. 17, 17–32.

Treisman, A., and Gormican, S. (1988). Feature analysis in early vision: evidence from search asymmetries. Psychol. Rev. 95, 15–48.

Tse, P. U. (1999). Volume completion. Cogn. Psychol. 39, 37–68.

Wolfe, J. M. (1998). “Visual search,” in Attention 1, Vol. 15, ed. H. Pashler (London: University College London Press), 77–84.

Wolfe, J. M., and Horowitz, T. S. (2004). What attributes guide the deployment of visual attention and how do they do it? Nat. Rev. Neurosci. 5, 495–501.

4.2 DRAWING EXPERTS ARE BETTER ABLE TO ACCUMULATE OBJECT STRUCTURE ACROSS EYE-MOVEMENTS

4.2.1 OBJECTIVES AND SUMMARY OF RESULTS Drawing accuracy requires attending to viewpoint-dependent visual properties, which would be similar to accessing a retinal-like percept. Our first study showed that training of professional artists and art students does not modify bottom-up visual processes to provide direct access to such a percept. Instead, we suggested that a long training in observational drawing may affect memory encoding strategies. A critical aspect of drawing is the correct positioning of the depicted features according to the global spatial organization of the original object (Locher, 2010). This would require a good understanding of the part-whole relationships defining object structures – namely, a good constructional ability. However, because drawing is a visuomotor task characterized by many eye-movements, it may be very difficult to hold a stable representation of these spatial relationships across the frequent changes in retinal input and reference frames.

This second study investigated whether people more proficient in drawing could better build a representation of an object’s structure from limited samples of information seen at each fixation.

To examine this issue, we designed two gaze-contingent experiments using either a moving window (experiment 1) or a moving scotoma (experiment 2). This enabled us to control the amount of visible information seen at each gaze location in central vision (experiment 1) or in peripheral vision (experiment 2) while participants were scanning an object in order to categorize its structure as possible or impossible. We used structurally possible and impossible line-drawings of objects (Penrose & Penrose, 1958; Schacter, Cooper, Delaney, Peterson, & Tharan, 1991; Soldan, Hilton, & Stern, 2009) as these required our participants to build a representation of the whole object to determine whether it was possible or not. We varied the size of the window (experiment 1) and of the scotoma (experiment 2), and found that people more skilled and more experienced in drawing could discriminate possible and impossible object structures based on smaller samples of central information (experiment 1), but showed no advantage when sampling from peripheral information (experiment 2).

Our results suggest that drawing skill relies on the ability to sequentially build an internal representation of the global structure of objects across eye movements, a representation that is robust as new information is accumulated in subsequent fixations.

EXPERIMENTS | drawing experts are better able to accumulate object structure across eye- movements

4.2.2 THE ARTIST’S ADVANTAGE: BETTER INTEGRATION OF OBJECT

INFORMATION ACROSS EYE-MOVEMENTS

This chapter is based on: Perdreau, F., & Cavanagh, P. (2013). The artist’s advantage: Better integration of object information across eye movements. I-Perception, 4(6), 380–395. doi:10.1068/i0574.

Abstract. Over their careers, figurative artists spend thousands of hours analyzing objects and scene layout. We examined what impact this extensive training has on the ability to encode complex scenes, comparing participants with a wide range of training and drawing skills on a possible versus impossible objects task. We used a gaze-contingent display to control the amount of information the participants could sample on each fixation either from central or peripheral visual field. Test objects were displayed and participants reported, as quickly as possible, whether the object was structurally possible or not. Our results show that when viewing the image through a small central window, performance improved with the years of training, and to a lesser extent with the level of skill. This suggests that the extensive training itself confers an advantage for integrating object structure into more robust object descriptions.

Keywords: artists, drawing, gaze-contingent, object recognition, eye movements, visual integration.

4.2.2.1 INTRODUCTION Why are some people better at drawing than others? This question is not as trivial as it may seem at first. Drawing is one of the earliest visuomotor tasks that humans mastered, arriving well before writing (at least 40,000 BCE against 3200 BCE) and perhaps emerging around the time at which language evolved. This order of appearance is also seen in child development, where young children start to learn handwriting by drawing the letter’s shape, but with practice the drawing and the writing processes become dissociated, with writing becoming more automated (e.g. Adi-Japha & Freeman, 2001; Feder & Majnemer, 2007). Drawing and writing both require a fine motor control, so why do many people master writing but fail at accurate drawing? One key difference is the amount of time spent learning to write versus learning to draw. While most children in developed countries spend many hours learning to write, only a subset extensively practice their drawing skills. What are the consequences of this extensive training? Does it alter the way a trained draftsperson sees the world, creating a more photorealistic perception? Or does training leave perception unaffected and instead change the robustness of the representations of objects and scenes, creating more stable codes for complex object structure just as master chess players create stable codes or chunks, for complex chess configurations (Gobet & Simon, 1996).

The motivation for our study derives from Cohen and Bennett’s (1997) description of three factors that contribute to better drawing accuracy: motor coordination, the perception of the model, and the selection of the relevant, to-be-drawn object’s features. These authors ruled out coordination as a factor because both artists and non-artists had equally accurate hand movements in a tracing task. The second factor—more veridical perception by artists of the to-be-drawn object (e.g. “the innocent eye” of Ruskin, 1912)—can also be ruled out. Initially, it was reported that individuals with better drawing skills saw objects more veridically, that is, they were less affected by size or shape constancies (Cohen & Jones, 2008; Mitchell, Ropar, Ackroyd, & Rajendran, 2005; Ostrofsky, Kozbelt, & Seidel, 2012, for their size constancy task). However, many failed to show that artists’ perceptual judgments are veridical (McManus, Loo, Chamberlain, Riley, & Brunswick, 2011; Ostrofsky et al., 2012, for their shape constancy task; Perdreau & Cavanagh, 2011, 2013). In particular, we (Perdreau & Cavanagh, 2011) found no differences between artists and non-artists in size, lightness, and shape constancies. Indeed, even across studies reporting a link between drawing skill and ability to discount visual constancy, the reduction in visual constancy (ranging from 0%, Perdreau & Cavanagh, 2011, to 5% to 10%, Ostrofsky et al., 2012) is far from the 100% reduction required if visual artists had indeed a veridical perception of the object as suggested by the innocent-eye hypothesis (Ruskin, 1912; see also Perdreau & Cavanagh, 2013, for a discussion of this point).

If neither motor coordination (Cohen & Bennett, 1997; Kozbelt, Seidel, ElBassiouny, Mark, & Owen, 2010) nor more photorealistic perception (McManus et al., 2011; Ostrofsky et al., 2012; Perdreau & Cavanagh, 2011) explains the artists’ advantage, which leaves the third factor (Cohen & Bennett, 1997): the selection and representation of the relevant object structure. Indeed, recent studies report that more skilled artists reproduced significantly more junctions in a drawing task than novices (Kozbelt et al., 2010; Ostrofsky et al., 2012), suggesting better knowledge of what must be selected from the object to produce an accurate drawing. However, although the selection of the relevant features is a necessary step, it is not the only requirement. What may be more important is the representation of the spatial relations between these features, as these relations underlie the object’s structure and proportions (Tchalenko, 2009).

Here again, studies have provided evidence for this advantage for spatial relations in artists. For instance, artists are better and faster at encoding complex sets of lines than novices (Glazek, 2012) and also better at recalling complex Rey-Osterrieth figures after relatively short delays (>30 s; McManus et al., 2010). In these tasks, the object’s structure must be encoded and also maintained in memory. These abilities repeatedly come into play during a drawing task as the object is processed sequentially over numerous gaze shifts

EXPERIMENTS | drawing experts are better able to accumulate object structure across eye- movements between the object and the drawing (Cohen, 2005; McManus et al., 2010). Interestingly, Glazek’s study (2012) found that more skilled subjects tend to have more motor output – draw more – than novices during the same fixation duration, whatever the object’s complexity. This suggests that drawing accuracy might be related to the ability to encode more structural information from a single fixation. However, in Glazek’s study, participants were free to scan the entire object while drawing it and could thus encode information present in both central and peripheral vision. These studies show the importance of selection and representation of spatial relations but they did not compare the relative contribution of these two components.

The present study addresses the role of these two processes to examine whether an artist’s advantage lies in (1) the ability to construct a robust, global representation in visual memory (visual chunks) across fixations or (2) the encoding of larger spatial extent of the object’s structure on each fixation (visual span). To do so, we designed two experimental tasks using a gaze-contingent moving window and a gaze-contingent moving mask that controlled the amount of visual information available from central and peripheral vision (e.g. Geisler, Perry, & Najemnik, 2006; Rayner, McConkie, & Zola, 1980). In the first experiment, the moving window only allowed the participants to see the center of their visual field (window radius varying from 1° to 5°), the surround being masked. Participants had to identify, within a limited time, whether the stimuli were structurally possible or impossible objects (see Figure 1 for examples). However, making such a decision requires either seeing the entire object or building up a mental representation across individual glimpses, for both possible and impossible objects share the same local features (vertices and junctions; e.g. Biederman, 1987; Soldan, Hilton, & Stern, 2009). If the artist’s advantage is explained by the ability to build and maintain a more robust description of an object’s structure, then they should be able to perform better with sparser inputs, constructing an internal model of the object from smaller samples (i.e. smaller window sizes). In other words, they may have access to more complex and robust codes for objects – bigger chunks (e.g. Chase & Simon, 1973) – that allow them to hold more complex structures in memory as they build up a representation from small samples. In contrast, our second experiment was identical except that it used a gaze-contingent central mask (from 5° to 10° of radius) that left the surround visible. The motivation for this task was to evaluate whether training and skill in drawing lead to better integration of structural information in peripheral vision. This increase in integration area on each fixation might be developed in order to process the global structure of an object while focusing on the local features currently being drawn, allowing them to be more appropriately placed relative to the overall structure. If this is the case, both experienced and skilled participants should be able to integrate information across larger

70 visual spans (e.g. Rayner, 1998; Reingold, Charness, Pomplun, & Stampe, 2001) and be more tolerant of this central scotoma.

To foreshadow our results, our data support the idea that the artists’ advantage lies, at least in part, in a better integration of the visual information picked up from different locations as the eyes move over a scene (experiment 1) but not in a better integration of that information across space (experiment 2). Moreover, we show that this advantage increases monotonically with the experience in drawing, and to a limited extent with drawing skill when foveal information is available.

Figure 1. Examples of possible and impossible objects used in the two experiments.

4.2.2.2 EXPERIMENTS In two experiments, participants determined whether a test object was structurally possible or impossible while viewing the figure through a gaze-contingent window of various sizes (experiment 1) or with a central scotoma with various sizes (experiment 2). To evaluate the ability to integrate object structure from across fixations and across space, we compute a critical window size that led to 75% correct responses. By correlating the estimated critical window size and scotoma size with drawing skill and experience, we examined whether artists can better construct an accurate memory representation of the test object from smaller samples.

EXPERIMENTS | drawing experts are better able to accumulate object structure across eye- movements

4.2.2.2.1 METHOD

4.2.2.2.1.1 MATERIAL All the experiments used the same apparatus. The participant’s head was always held by a chinrest so that his or her eyes were approximately 57 cm from the center of the screen. The stimuli were projected on a 22-inch CRT screen, with a resolution of 1024×768 pixels and with a frame rate of 120 Hz. The experiments were programmed in MATLAB using the Psychophysics and Eyelink Toolbox extensions (Brainard, 1997; Cornelissen, Peters, & Palmer, 2002; Pelli, 1997), and were run on an Apple computer.

Participants’ eye movements were recorded with an eye-tracking system (Eyelink 1000 monocular, 35-mm lens) at a 1000-Hz sampling rate. The eye tracker was always calibrated for the participant’s dominant eye. Finally, eye movement events and saccades were parsed using the Eyelink 1000 algorithm (saccade acceleration threshold: 9500 deg/s², saccade velocity threshold: 35 deg/s).

4.2.2.2.1.2 STIMULI The majority of the possible and impossible objects were taken from the studies of Schacter, Cooper, Delaney, Peterson, and Tharan (1991) and Soldan et al. (2009), used with the authors’ permission (Figure 1). The rest of the objects were designed by one of the authors (FP). In order to avoid any ambiguities about the objects’ nature, we first asked 20 independent observers, not participating in the experiments, to judge all the line drawings as structurally possible or impossible. We specifically instructed them, according to Schacter et al.’s procedure (1991), that each object’s edges were necessarily represented with lines, that surfaces were flat and could only face a single orientation, and that the drawings represented solid and 3D objects. The 20 independent observers saw all the objects in a random order. For the main experiments, we only kept the objects that had an inter-observer agreement of 95% or better, leaving 85 possible objects and 66 impossible objects. However, we used two random subsets that were different across the experiments. These subsets had 30 objects of each category. Objects were distributed among the trials so that they were repeated three times within the same experiment, but never in the same orientation and never with the same window size. The objects’ pictures were presented at 21°×21° of the visual angle and were centered on the screen.

4.2.2.2.1.3 GAZE-CONTINGENT DISPLAY In the “moving window” experiment, the gaze-contingent display was a circular, fully transparent area with a diameter that was the independent variable, whereas it was a circular, fully opaque area in the “moving scotoma” experiment. Their position was

72 continuously updated and centered on the subject’s fixation location provided by the eye- tracking system. We measured the effect of window and mask diameter on task performance with eight different sizes ranging from 1° to 5° in radius (1°, 1.5°, 2.1°, 2.6°, 3.1°, 3.7°, 4.2° and 5°) in the “moving window” experiment, while the scotoma radius ranged from 5° to 9.8° (5°, 5.7°, 6.4°, 7.1°, 7.8°, 8.4°, 9.1° and 9.8°).

4.2.2.2.1.4 PARTICIPANTS Twenty-six participants were tested (mean age of 28.3±1.9 years, 11 males, 15 females). Twenty-four of the participants ran in both experiments. One participant ran only in the first experiment and one only in the second. All the participants were informed of the experiment’s purpose and risks, and all gave their informed consent before starting the experiment. Finally, all participants were paid 20 euros for the entire session. The characteristics of participants are described in Table 1.

For convenience in this article, we have referred to our more skilled and trained participants as artists even though there is no real definition for “artist” or “visual art,” and clearly drawing accuracy alone does not make someone an artist. Nevertheless, using this very general label is not unreasonable, for drawing is a common task in many different artistic activities (e.g. illustration, movie or dance; for a discussion of this point, see Seeley & Kozbelt, 2008).

4.2.2.2.1.5 PROCEDURE Each experiment included 9 blocks of 20 trials each, and began with a practice block of 8 trials. Each block was preceded by a calibration of the eye movement monitor. The first block served as a baseline, where the objects were fully visible, while the eight other blocks used the gaze-contingent window. Except for the presence of the window, the procedure was identical for all the blocks.

Each trial started with a fixation dot at a randomly chosen location within the image. The subject had 1 s to fixate it for 250 ms. After this fixation test, the gaze-contingent display (Figure 2) appeared centered on the fixation dot’s coordinates. The participants had to report whether the object was structurally possible or impossible as fast as possible by pressing the appropriate key (“L” if possible, “S” if impossible). The display only remained for 10 s, and participants had to give their response within this time. In the cases where the participants did not, a new screen appeared asking them to give a response.

We tested eight window sizes in addition to the full-view condition tested during the first block (“baseline”). There were 20 trials per window size, 20 trials for the baseline

EXPERIMENTS | drawing experts are better able to accumulate object structure across eye- movements condition, and thus 90 trials per object category (possible and impossible), for a total of 180 trials.

Figure 2. Trial procedure. (a) The trials started with a fixation dot appearing at a random position. Once the participant fixated it for 250 ms, the gaze-contingent window appeared, allowing the participant to only see the center of his or her vision (except in the baseline block, where the stimuli were fully visible). The participant had to respond, as fast as possible, whether the partially hidden object was structurally possible or impossible. If the participant did not respond within 10 s, the object and the window disappeared and were replaced by a screen asking him or her to give a response. (b) In experiment 2, the procedure was similar to that of the first experiment, except that we manipulated a gaze-contingent moving mask (scotoma) that blocked central vision.

4.2.2.2.1.6 EVALUATION OF DRAWING SKILL Participants’ drawing skills were evaluated, first with a short set of questions about their drawing experience and then with a drawing task where they had 15 min to make a pencil drawing of a 31°±21° gray-scale photograph of an octopus (the same as used by Kozbelt et al., 2010). The participants were instructed to copy the original as realistically and accurately

74 as possible, that is, without emphasizing aestheticism and creativity. They were allowed to use all the drawing techniques they knew and to erase and correct their drawing as many times as they needed. Once the 15 min were over, the model disappeared from the screen.

Figure 3. Examples of drawings made by the participants during the drawing task. Drawings presented in the lower row are those judged as less accurate, while the upper drawings were judged more accurate.

To score the subject’s drawing skill, we asked eight independent judges, blind to the participants’ identities, to evaluate the accuracy and the realism of the productions. Accurate drawings respected the model’s proportions, the position and shape of the shadows, and details of the texture. During the judging, the drawings were presented on a computer screen simultaneously with the model. In a first step, the raters could look at all of the drawings, without making a decision, to get an idea of the range of skills (Figure 3). Then, in a second pass, the drawings were presented in a random order and the judges rated each on an 8- point scale (1 for a very low accuracy and 8 for the best accuracy). This ratings cycle was repeated five times with the drawings in a new random order each time. All the judgments were consistent within and between the raters (Cronbach’s α; intra-α: 0.94±0.01; inter-α: 0.97). The mean of the judges’ ratings was taken as each participant’s drawing score (Table 1).

EXPERIMENTS | drawing experts are better able to accumulate object structure across eye- movements

Name Age Handedness Gender School Experience Rating

AC 28 R F 0 0 1,69 AF 20 R F 0 0 2,31 CC 38 L M 0 0 4,72 DG 22 R F 0 0 2,59 JZ 52 R M 0 0 1,50 LM 22 R M 0 0 1,47 LMO 18 L M 0 0 1,56 MDM 23 L M 0 0 3,25 PS 19 R F 0 0 1,47 LC 24 R M 0 1 1,09 JS 40 R M 0 2 2,81 IV 44 L F 0 3 5,16 ML 27 R F 0 3 6,16 TC 31 R F 0 3 3,28 CR 22 L M 1 4 4,00 CL 44 R F 1 5 2,72 GC 21 R F 1 7 5,00 AG 28 R F 1 8 6,59 BE 25 L M 1 10 3,81 AL 24 R F 1 15 4,94 LRF 25 R F 0 15 4,34 TH 22 L M 0 15 5,44 VP 23 L M 1 15 6,00 LD 31 R F 1 20 5,28 SD 33 R F 1 20 3,19 YM 54 L F 0 28 2,66 Table 1. Participants. Some of our participants were or are following a formal training in visual arts (reported in the “School” column. Years of experience corresponds to full years of weekly training. “Years of experience” was set to 0 for participants with no training in drawing. The “Rating” column reports mean rating that participants obtained in our drawing task. Handedness is reported as follows: “R” for right-handed, “L” for left-handed. Finally, in the “Gender” column, “F” and “M” correspond to female and male, respectively.

4.2.2.3 RESULTS

4.2.2.3.1 ACCURACY The data of the two baseline blocks (unobstructed, full view of the line-drawn objects) were pooled together (from both experiment 1 and 2), giving 40 baseline trials. Participants who did not reach 75% correct responses in the baseline condition were excluded from the analysis (n=3). However, in the first experiment, one subject’s behaviour strongly deviated from others and could reasonably be considered as an outlier (mean Cook’s distance = 0.44. Bollen & Jackman, 1985; Cook, 1979). Although this subject reached the 75% criterion in the baseline condition, she never exceeded 50% correct in the test conditions, seriously skewing

76 her threshold value. In the following section, we will present the results without this subject but, overall, this had little effect on the results. Consequently, 21 participants were included in the first experiment’s analyses, and 22 participants in those of the second experiment.

We used two approaches to evaluate the influence of drawing skill and years of experience: 1) we took their drawing skill and self-reported years of experience in drawing as indexes to correlate with their performance; 2) we split our participants into two groups, separately accordingly to their drawing rating (skilled vs unskilled) and their years of experience (trained vs untrained). For drawing accuracy, participants were considered as unskilled if their rating was smaller than the median of all ratings; whereas were taken as untrained participants who had no experience in drawing.

To measure the critical window and scotoma size in the possible vs. impossible task, we plotted each participant’s performance against the size of the viewable area. Psychometric curves were then fitted (Fig. 4) with Weibull functions using a maximum likelihood method (Prins, 2012; Wichmann & Hill, 2001a, 2001b). This resulted in a mean deviance (quality of fit) of 4.90 (SE: 0.58) for the first experiment’s fits, and of 8.43 (SE: 1.11) for the second experiment’s (all ps >.05). The participant’s critical window and scotoma size were both defined as the viewable area leading to 75% correct responses. In both experiments, reaction times decreased as performance increased with window size, ruling out a speed-accuracy trade-off [Exp1: r(7)=-0.98,p<.0001; Exp2: r(7)=-0.96, p<.0001].

We then plotted each participant’s critical window size and critical scotoma size against his or her drawing score and separately against his or her years of drawing experience. Because the distribution of our data violated assumptions of bivariate normality, we used a non-parametric, rank-based Spearman’s correlation to evaluate the relationship among these variables (p-values are adjusted for multiple comparisons with a Holm- Bonferroni sequential procedure; Holm, 1979). The confidence intervals of the correlation coefficients were computed with non-parametric bootstraps (10000 runs) and we used the percentile method for the confidence interval computation (DiCiccio et al., 1996).

In the “central window” experiment, we found (Fig. 4) a significant negative correlation between the critical window sizes and the drawing scores [r(19)=-0.56(-0.83,-0.12), p<.02] as well as a significant, negative correlation between the critical window sizes and self-reported years of drawing experience [r(19)=-0.72(-0.82,-0.62), p<.001]. The finding that both are significant is not very surprising since drawing scores and experience were also strongly correlated [r(19)=0.75(0.52,0.88) , p<0.001] and shared about 56% (R²) of their variance. In contrast, in the “moving scotoma” experiment, no significant correlation was found between

EXPERIMENTS | drawing experts are better able to accumulate object structure across eye- movements critical scotoma size and either drawing skills [r(20)=-0.44(-0.72,-0.16), p=0.13] or drawing experience [r(20)=-0.40(-0.70,-0.09), p=0.13].

Similar results were found when comparing the participants in groups for both drawing accuracy and experience. The critical window sizes of the more skilled participants were significantly smaller than those of novices [Skilled: 24.85(5.74), Unskilled: 68.50(11.23), t(19.0)=-3.56, CI(95%)=[-69.32,-17.99], d=-1.56, p<0.004], as well as when considering experience [Trained: 26.94(6.00), Untrained: 70.56(12.03), t(19.0)=-3.50, CI(95%)=[-69.69,- 17.54], d=-1.54, p<0.004]. However, despite insignificant correlations between critical scotoma sizes and both experience and drawing accuracy, the group comparisons show a significant advantage for more skilled participants in the second experiment [Skilled: 262.73(29.12), Unskilled: 375.09(14.31), t(20.0)=-3.46, CI(95%)=[-180.03,-44.68], d=-1.48, p<0.005], but not for more experienced participants [Trained: 295.74(28.15), Untrained:359.46(19.07), t(20.0)=-1.58, CI(95%)=[-147.57,20.14], d=-0.70, p=0.13].

These results suggest that increased drawing practice and drawing skill are related to increased ability to integrate object structure from smaller samples of foveal information (experiment 1). Is there any difference between the contribution of drawing practice and drawing skill in this task? To compare the two correlation levels – for years of experience and for drawing score, which are themselves correlated – we used Fisher’s Z transformed correlation coefficients (Meng, Rosenthal, & Rubin, 1992). We found no significant differences between these correlation coefficients in the first experiment [rratings-ryears=-0.16(- 0.67,0.14), Z(18)=-1.30, p=0.17].

One simple explanation of the link between years of practice and better performance with smaller window sizes in the first experiment is that participants with more drawing practice were just better overall at the possible vs. impossible task. However, when the drawings were in full view in the baseline condition, there was no correlation between performance and years of experience [Exp. 1: r(19)=0.32(-0.03,0.67), p=.47; Exp. 2: r(20)=0.30(-0.05,0.54), p=.56]. Another explanation is that years of drawing experience simply reflect the participant’s age irrespective of drawing experience. However, this was not the case either. We found no correlation between critical window sizes and the participants’ age [Exp. 1: r(19)=0.03, p=.90; Exp. 2: r(20)=0.13, p=.56], where if anything, the (non- significant) effect is the opposite of that of experience: participants got worse (although not significantly worse), requiring larger window sizes, as they got older.

In contrast to Experiment 1, Experiment 2 showed no link between years of practice and better performance and central scotoma size. Nevertheless, we again checked the possible contributions of secondary variables and, as in Experiment 1, we found no

78 significant correlations between age and critical scotoma size or between baseline performance and years of experience [Age: r(20)=0.13(-0.28,0.54), p=.56; Baseline performance: r(20)=0.30(-0.05,0.65), p=0.50].

We next analyzed possible learning effects. Our test objects were presented three times with different rotation angles in each experiment, and some of them already appeared fully visible in the baseline condition. Since it is known that visual artists show an advantage in recalling complex figures after a delay of 30 seconds (McManus et al., 2010), it is possible that our results could be due to a learning effect that would be greater in more skilled subjects. To test this hypothesis, we computed the participants’ mean performance for each block of trials and we ran a linear regression for each individual to determine if there were any improvements across blocks in either experiment. Participants’ slopes were on average significantly higher than 0 [Exp1: t(20)=10.75, CI(95%)=[0.03,0.05], p<.001; Exp2: t(21)=4.76, CI(95%)=[0.01,0.03], p<.01], suggesting indeed the presence of a learning effect. However, when correlating individual regression slopes against participants’ experience and drawing accuracy separately, we found no evidence of a greater learning rate in more skilled participants with the critical central window size in experiment 1 [ryears(19) =0.35(0.01,0.68), p=0.24; rratings(19)=0.30(-0.05,0.66), p=0.24] or with the critical central scotoma size in experiment 2 [ryears(20) =0.37(-0.04,0.69), p=0.19; rratings(20)=0.48(0.22,0.73), p=0.17].

Finally, we examined whether there was any performance advantage for the 20 of 60 objects that were seen in the baseline condition and then seen again in the subsequent central window and scotoma test conditions. To do so, we computed the mean performance for the objects seen in both the baseline and test conditions and compared that to the mean performance for the objects only seen in the test conditions (using an arcsine square root transformation). There was no significant difference between these two categories in either experiments [Exp1: t(42)=-0.44; Exp2: t(42)=0.69].

EXPERIMENTS | drawing experts are better able to accumulate object structure across eye- movements

Figure 4. (a) Scatter plots of the subjects’ critical window size against subjects’ years of experience and drawing accuracy. We first fit Weibull function to each participant’s proportion of correct response and we took the critical window size as the size leading to 75% correct response (the left panel shows an individual fit). We then correlated participants’ critical sizes to their experience and drawing rating (right panel, top row). We found a significant negative correlation between the subjects’ critical window size and their amount of training as well as with their drawing skills (dotted lines are 95% confidence interval of the regression line). Similar results were also found when comparing groups of participants made accordingly to levels of skill and of experience (box plots in the right panel, bottom row; error bars are 95% confidence intervals of the means). This effect cannot be attributed to age, which was not correlated with critical window size, or to a better knowledge of what is an impossible or a possible object, represented by how subjects performed in the baseline condition, which was not correlated with years of experience. (b) Scatter plots of the subjects’ critical scotoma size. We did not find a significant correlation between the participants’ critical scotoma size and either their amount of training or their drawing skills. However, we found a significant advantage for more skilled

participants when splitting the participants into two groups according to drawing scores although not when they were divided according to years of experience.

4.2.2.3.2 REACTION TIMES Not surprisingly, participants classified the objects more quickly when more of the image was visible (Fig. 5). Specifically, linear contrasts showed that participants’ reaction times significantly decreased as the viewable area increased [Exp1: F(3.42,20)=100.1, p<.000, η²=.83; Exp2: F(4.59,21)=8.80, p<.000, η²=.29].

To measure whether this effect of window size on participants’ reaction times changed with their experience or accuracy in drawing, we computed linear regressions with window sizes (baseline excluded) as predictors and reaction times as independent variable for each participant. We found no significant correlation between either experience or accuracy and participants’ reaction time regression slopes in the first experiment [Experience: r(19)=-0.45(-0.728,-0.169), p=0.12; Accuracy: r(19)=-0.41(-0.71,-0.10), p=0.13], whereas these variables were significantly related in the second experiment [Experience: r(20)=-0.62(-0.78,-0.46), p<0.006; Accuracy: r(20)=-0.59(-0.77,-0.41), p<0.008]. This suggests that in the presence of a moving scotoma, more experienced and more skilled subjects took more time when less peripheral information was visible. However, this had no effect on these participants’ performances, since we found no correlation between critical scotoma sizes and either experience or drawing accuracy.

Finally, mean reaction times were not correlated to participants’ drawing experience in either the test phase [Exp1: r(19)=-0.003(-0.44,0.44), p=.99; Exp2: r(20)=0.01(−0.41,0.44), p=.99] or the baseline condition [Exp1: r(19)=-0.42(-0.71,0.12), p=.12; Exp2: r(20)=-0.42(- 0.71,0.13), p=.10]. In contrast, we found that reaction times in the baseline conditions were related to drawing accuracy in the first experiment [Exp1: r(19)=-0.56(-0.76,-0.35), p<0.03; Exp2: r(20)=-0.48(-0.73,-0.23), p=.07], whereas these variables were not correlated in the test conditions with partial visibility [Exp1: r(19)=0.006; Exp2: r(20)=0.02; ps>>.05]. The correlation between drawing score and reaction time in the first experiment baseline condition is only marginally significant but it suggests that more skilled, (but not more trained) participants took less time to categorizing possible vs. impossible objects when they were fully visible. Perhaps this advantage did not appear in the baseline accuracy results because of a ceiling effect under full visibility. This might suggest that more skilled participants are indeed faster at encoding complex set of lines (Glazek, 2012). Whatever the case, this reaction time advantage for skilled participants was no longer seen when the stimuli were partially visible.

EXPERIMENTS | drawing experts are better able to accumulate object structure across eye- movements

Figure 5. (a) Mean reaction times (RTs) against the viewable area sizes (baseline excluded) for the first experiment (central window). RTs significantly decreased as the viewable area increased. This effect of window size did not significantly differ across levels of experience or of drawing accuracy. (b) Mean reaction times for the second experiment. As in the first experiment, RTs decreased as the viewable area increased.

4.2.2.3.3 EYE MOVEMENTS We next analysed whether artists used different strategies to explore the test images and that could explain the better performances of our skilled and experienced participants in our tasks. We began by characterizing the effects of the window and scotoma size on the number and duration of fixations, independently of drawing skill or experience. We computed the average number of fixations per second made during each trial (Fig. 6). A repeated- measure two-way ANOVA with linear contrasts ran on subjects’ fixation rates and with object

82 categories (possible, impossible) and window sizes as factors shows that fixation rates increased with the window’s size in the “moving window” experiment [F(1,20) = 29.67, p<.000, η²=.54], as well as in the “moving scotoma” experiment [F(1,21) = 27.22, p<.000, η²=.57].

Figure 6. (a) Mean fixation rate as a function of window size. In experiment 1, the fixation rate (number of fixations per second) significantly increased with window sizes. (b) Fixation rate also linearly increased with viewable area size in the second experiment. However, in both experiments, this effect of window size did not differ with skills or experience in drawing.

Are any of these eye movement properties affected by the participants’ experience or skill? We first analyzed whether fixation rates were related to the better performances found in our more trained and skilled subjects. However, none of the correlations between either years of experience and drawing ratings with mean fixation rates was significant in either experiment [all ps>>.05].

EXPERIMENTS | drawing experts are better able to accumulate object structure across eye- movements

Figure 7. (a) Summary box plots of the most frequently fixated features in the baseline conditions with participants grouped according to their drawing accuracy (left panel) and experience (right panel). Error bars correspond to 95% confidence interval of the means. We first extracted junctions from every object and we then binned every image into 1°x 1° areas of interest. All areas without any information were taken as “empty blocs,” while areas with segments information but not junctions were labeled “lines blocs.” Finally, areas with a junction present in was considered as a “junction bloc.” We then computed the total number of fixations made by the subjects for objects seen in baseline (full view) trials. We found no correlation between these results and participants’ training or skill. (b) Summary box plots of fixation coverage. Subjects’ fixation coverage was computed using Wooding’s procedure (Wooding, 2002) on fixations made in the test conditions (baseline excluded). We found no significant link between subjects’ fixation coverage and their years of experience or the ratings of their drawing.

Next, we analysed the fixation patterns from the full-view conditions, where participants could make saccades toward visible features, to see if they revealed any strategy for classifying the objects that might differ with experience and drawing accuracy. We constructed fixation maps for every object and classified each fixation as falling on (within a 1° x 1° region. Gaze position accuracy: 0.15°) empty space, a line segment (but not a junction) or a junction (using Harry corner detector, Harris & Stephens, 1988, and Canny

84 filter, Canny, 1986, to define these features). The frequency of fixating these features is shown in Figure 7a for participants split into Skilled vs Unskilled groups (left panels) and Trained vs Untrained groups (right panels). The results show that most fixations in the baseline conditions fell on empty space (nearest line or junction at least 0.5° distant), and that participants fixated significantly more junctions than line segments [respectively, t(40)=5.84, p<.0001; t(40)=9.83, p<.0001, and t(40)=6.68, p<.0001]. However, neither more skilled or more trained subjects showed more frequent fixations on empty space [Drawing skill: t(19)=-1.82(-4.18,0.29), p=.08; Experience: t(19)=-0.93(-3.45,1.32), p=.36], on junctions [Drawing skill: t(19)=-0.81(-1.61,0.71), p=.43; Experience: t(19)=-0.62(-0.83,1.53), p=.54] or on line segments [Drawing skill: t(19)=-1.91(-0.89,0.04), p=.07; Experience: t(19)=-1.18 (- 0.78,0.22), p=.25].

Finally, we examined whether more trained and skilled participants were more efficient in placing their eye movements in order to sample a larger extent of the image in each trial, To do so, we computed how much of each image landed on participants’ foveas as they scanned the image in the moving window conditions (Experiment 1, baseline excluded), using Wooding’s procedure (Wooding, 2002). We counted only the 2° diameter foveal area for each fixation (dcrit=50%, σ=2°). Participants covered on average 4.1 % (SE: 0.1) of the test images (Figure 7b) and the extent of coverage did not differ between novices and either the more skilled or more trained participants [Drawing skill: t(19)=1.32(-0.19,0.85), p=.20; Experience: t(19)=1.44(-0.16,.88), p=.16].

Altogether, these results provide no evidence for different fixation patterns between skilled or experienced subjects and novices. This suggests that the better performance found in more trained and skilled participants may be due to a more robust internal representation rather than a different pattern of visual exploration.

4.2.2.4 GENERAL DISCUSSION This study investigated whether drawing skill and years of practice in drawing lead to a more efficient analysis of objects and scene organization. Making a drawing requires an accurate integration of the to-be-drawn object’s features in order to reproduce it in proper proportion. Our hypothesis was that training would (1) increase the size of the integrated object structure (larger visual chunks) and (2) improve the ability to encoding information from a larger extent of peripheral vision (larger visual span). To test these hypotheses, we designed two experiments using a gaze-contingent moving window and moving scotoma that controlled the amount of information available in central and peripheral vision, respectively. Our participants had to categorize line drawings of objects, seen through these central or peripheral samples, as either possible or impossible.

EXPERIMENTS | drawing experts are better able to accumulate object structure across eye- movements

The results of experiment 1 demonstrated that both the number of years of drawing experience and drawing accuracy reliably predict better performance in the identification of an object’s structure from small foveal samples. This result was not explained by the participants’ age or by overall skill in discriminating possible from impossible objects. Our first hypothesis was confirmed.

In addition to improving the integration of sequentially acquired samples from central vision, training might also improve integration over larger spatial areas in each fixation. This larger visual span should allow accurate performance even with larger amounts of central vision blocked out. However, when we varied the size of a moving central mask in our second experiment, we found no significant correlation between the participant’s experience in drawing or drawing scores and their critical scotoma size. Our second hypothesis was not confirmed: we found no evidence that the artist’s training affects the efficiency in the use of peripheral information.

Moreover, we found no significant difference in the fixation patterns of our participants. The more skilled and trained subjects did not fixate salient features (junctions or vertices) more frequently nor did they scan a larger extent of space during each trial to see more of the object. In the absence of any strategic differences in scanning the images, our results suggest that the artists’ advantage must be in the representation of the information sampled from the partial images.

Although we found a relation between drawing experience and skill and the performance in our tasks, it still remains unclear what is acquired when learning to draw. Our tasks were perceptual, so we are unable to assess the contribution of changes in motor coordination in our findings. Moreover, drawing accuracy and drawing experience shared about half of their variance, so that we are not able to conclude whether this advantage for more trained subjects is the result of training or simply a pre-existing, innate ability that led them to pursue drawing. Despite this ambiguity, we favor the effect of practice as a source of improved performance in our experiments. We have two reasons for this: first, performance was more correlated to years of experience than drawing skill itself; and years of experience is an imperfect measure of the choice of a career in art since many in our population are just starting their training. Moreover, it has been shown that children’s drawing skill is improved by training and by learning drawing rules (Rand, 1973), suggesting that experience – the time spent at practicing – may play a crucial role in developing drawing skills. As hypothesized by Gombrich (1960), learning explicit rules is indeed a first step in the acquisition of drawing skills, but it has to be embodied through practice, thus becoming an implicit visuomotor knowledge or schema. Nevertheless, to

86 show that training causes the improved performance in our tasks, we would need longitudinal training experiments that control for the subjects’ initial drawing skill.

4.2.2.4.1 VISUAL INTEGRATION AND DRAWING If the artists’ advantage arises from a better integration of visual samples into a more robust representation of object structure, how is this related to the demands of drawing? Several studies have examined the contribution of object representation to drawing skills (Cohen & Bennett, 1997; Glazek, 2012; Kozbelt, 2001; Kozbelt et al., 2010; Ostrofsky et al., 2012). A trained draftsperson might have a more coherent representation of an object to make an accurate copy of it and that “coherence” might be specifically tuned to the requirements of reproducing it. For example, Kozbelt et al. (2010) and Ostrofsky et al. (2012) have shown that people more skilled in drawing selected and reproduced more structural information (e.g. vertices and junctions) than those who were less skilled. The relevant structural information must also be combined with accurate spatial relations between elements. While constructing a drawing, these spatial relations are built up across many sequential eye movements, where local samples of the original scene are independently encoded at every fixation location (Coen-Cagli et al., 2009; Locher, 2010). Consequently, a key requirement for constructing an accurate representation of the objects in a scene must be to integrate these local samples, spaced appropriately according to the size of each eye movement and this integration must require a representation that is robust enough to retain its accuracy during this building process. Our results showed that participants with more training and better accuracy in drawing are indeed better able to integrate samples of information across eye movements. Constructions of robust representations have been observed in expert chess players who develop robust representations of the structure of the chess pieces (Curby, Glazek, & Gauthier, 2009; Gobet & Simon, 1996; Reingold et al., 2001). We suggest that drawing accuracy might similarly arise from the ability to represent more complex sets – or chunks – of relevant visual information obtained from small local samples. In particular, while creating drawings, the artist is continually focusing on small portions of a scene in order to reproduce it patch by patch, all the while keeping track of the larger organization of the object and the scene in order to place each element appropriately (Coen-Cagli et al., 2009; Tchalenko & Miall, 2008). Over thousands of hours of training, we suggest that this particular style of attention to the scene leads the artists to develop robust representations that are optimal for the step-by-step production of the drawing and the motor planning of hand movements.

EXPERIMENTS | drawing experts are better able to accumulate object structure across eye- movements

4.2.2.5 REFERENCES

Adi-Japha, E., & Freeman, N. H. (2001). Development of differentiation between writing and drawing systems. Developmental Psychology, 37(1), 101–114.

Biederman, I. (1987). Recognition-by-components: A theory of human image understanding. Psychological Review, 94(2), 115–117.

Bollen, K. A., & Jackman, R. W. (1985). Regression Diagnostics: An Expository Treatment of Outliers and Influential Cases. Sociological Methods & Research, 13(4), 510–542.

Brainard, D. H. (1997). The Psychophysics Toolbox. Spatial Vision, 10(4), 433–436.

Canny, J. (1986). A Computational Approach to Edge Detection. IEEE Transactions on Pattern Analysis and Machine Intelligence, 8(6), 679–698.

Chase, W. G., & Simon, H. A. (1973). Perception in chess. Cognitive Psychology, 4(1), 55– 81.

Coen-Cagli, R., Coraggio, P., Napoletano, P., Schwartz, O., Ferraro, M., & Boccignone, G. (2009). Visuomotor characterization of eye movements in a drawing task. Vision research, 49(8), 810–8.

Cohen, D. J. (2005). Look little, look often: The influence of gaze frequency on drawing accuracy. Perception & Psychophysics, 67(6), 997–1009.

Cohen, D. J., & Bennett, S. (1997). Why can’t most people draw what they see? Journal of Experimental Psychology: Human Perception and Performance, 23(3), 609–621.

Cohen, D. J., & Jones, H. E. (2008). How shape constancy relates to drawing accuracy. Psychology of Aesthetics, Creativity, and the Arts, 2(1), 8–19.

Cook, D. R. (1979). Influential in Linear Observations Regression. Journal of the American Statistical Association, 74(365), 169–174.

Cornelissen, F. W., Peters, E. M., & Palmer, J. (2002). The Eyelink Toolbox: Eye tracking with MATLAB and the Psychophysics Toolbox. Behavior Research Methods, Instruments, & Computers, 34(4), 613–617.

Curby, K. M., Glazek, K., & Gauthier, I. (2009). A Visual Short-Term Memory Advantage for Objects of Expertise. Journal of experimental psychology. Human perception and performance, 35(1), 94 –107.

DiCiccio, T. J., Efron, B., Hall, P., Martin, M. A., Canty, A. J., Davison, A. C., … Young, G. A. (1996). Bootstrap confidence intervals. Statistical Science, 11(3), 189–228.

Feder, K. P., & Majnemer, A. (2007). Handwriting development, competency, and intervention. Developmental medicine and child neurology, 49(4), 312–7.

Geisler, W. S., Perry, J. S., & Najemnik, J. (2006). Visual search: the role of peripheral information measured using gaze-contingent displays. Journal of vision, 6(9), 858–73.

Glazek, K. (2012). Visual and motor processing in visual artists: Implications for cognitive and neural mechanisms. Psychology of Aesthetics, Creativity, and the Arts, 6(2), 155– 167.

Gobet, F., & Simon, H. a. (1996). Templates in chess memory: a mechanism for recalling several boards. Cognitive psychology, 31(1), 1–40.

Gombrich, E. (1960). Art and Illusion : A Study in the Psychology of Pictorial Representation Summary (Fifth edit.). Oxford: Phaidon Press Limited.

Harris, C., & Stephens, M. (1988). A Combined Corner and Edge Detector. In Proceedings of the 4th Alvey Vision Conference (pp. 147–152). Alvey Vision Club.

Holm, S. (1979). A simple sequentially rejective multiple test procedure. Scandinavian journal of statistics, 6(2), 65–70.

Kozbelt, A. (2001). Artists as experts in visual cognition. Visual Cognition, 8(6), 705–723.

Kozbelt, A., Seidel, A., ElBassiouny, A., Mark, Y., & Owen, D. R. (2010). Visual selection contributes to artists’ advantages in realistic drawing. Psychology of Aesthetics, Creativity, and the Arts, 4(2), 93–102.

Locher, P. (2010). How Does a Visual Artist Create an Artwork? In J. C. Kaufman & R. J. Sternberg (Eds.), The Cambridge Handbook of Creativity (pp. 131–144). Cambridge: Cambridge University Press.

EXPERIMENTS | drawing experts are better able to accumulate object structure across eye- movements

McManus, I. C., Chamberlain, R., Loo, P. W., Rankin, Q., Riley, H., & Brunswick, N. (2010). Art students who cannot draw: Exploring the relations between drawing ability, visual memory, accuracy of copying, and dyslexia. Psychology of Aesthetics, Creativity, and the Arts, 4(1), 18–30.

McManus, I. C., Loo, P., Chamberlain, R., Riley, H., & Brunswick, N. (2011). Does Shape Constancy Relate to Drawing Ability? Two Failures to Replicate. Empirical Studies of the Arts, 29(2), 191–208.

Meng, X., Rosenthal, R., & Rubin, D. (1992). Comparing Correlated Correlation Coefficients. Psychological Bulletin, 111(1), 172–175.

Mitchell, P., Ropar, D., Ackroyd, K., & Rajendran, G. (2005). How Perception Impacts on Drawings. Journal of experimental psychology. Human perception and performance, 31(5), 996 –1003.

Ostrofsky, J., Kozbelt, A., & Seidel, A. (2012). Perceptual constancies and visual selection as predictors of realistic drawing skill. Psychology of Aesthetics, Creativity, and the Arts, 6(2), 124–136.

Pelli, D. G. (1997). The VideoToolbox software for visual psychophysics: transforming numbers into movies. Spatial Vision, 10(4), 437–442.

Perdreau, F., & Cavanagh, P. (2011). Do artists see their retinas? Frontiers in human neuroscience, 5(171), 1–10.

Perdreau, F., & Cavanagh, P. (2013). Is Artists’ Perception more Veridical? Frontiers in Neuroscience, 7(6), 1–11.

Prins, N. (2012). The psychometric function: the lapse rate revisited. Journal of vision, 12(6), 1–16.

Rand, C. W. (1973). Copying in Drawing: The Importance of Adequate Visual Analysis versus the Ability to Utilize Drawing Rules. Child Development, 44(1), 47.

Rayner, K. (1998). Eye movements in reading and information processing: 20 years of research. Psychological Bulletin, 124(3), 372–422.

Rayner, K., McConkie, G. W., & Zola, D. (1980). Integrating information across eye movements. Cognitive Psychology, 12(2), 206–226.

Reingold, E. M., Charness, N., Pomplun, M., & Stampe, D. M. (2001). Visual Span in Expert Chess Players: Evidence From Eye Movements. Psychological Science, 12(1), 48–55.

Ruskin, J. (1912). Elements of drawing. New York. London: J.M. Dent & sons.

Schacter, D. L., Cooper, L. a, Delaney, S. M., Peterson, M. a, & Tharan, M. (1991). Implicit memory for possible and impossible objects: Constraints on the construction of structural descriptions. Journal of Experimental Psychology: Learning, Memory, and Cognition, 17(1), 3–19.

Seeley, W. P., & Kozbelt, A. (2008). Art, Artists, and Perception: A Model for Premotor Contributions to Perceptual Analysis and Form Recognition. Philosophical Psychology, 21(2), 149–171.

Soldan, A., Hilton, H. J., & Stern, Y. (2009). Bias effects in the possible/impossible object decision test with matching objects. Memory & cognition, 37(2), 235–47.

Tchalenko, J. (2009). Segmentation and accuracy in copying and drawing: experts and beginners. Vision research, 49(8), 791–800.

Tchalenko, J., & Miall, R. C. (2008). Eye-hand strategies in copying complex lines. Cortex, 45(3), 368–76.

Wichmann, F. A., & Hill, N. J. (2001a). The psychometric function: I. Fitting, sampling, and goodness of fit. Perception & Psychophysics, 63(8), 1293–1313.

Wichmann, F. A., & Hill, N. J. J. (2001b). The psychometric function: II. Bootstrap-based confidence intervals and sampling. Perception & Psychophysics, 63(8), 1314–1329.

Wooding, D. S. (2002). Eye movements of large populations: II. Deriving regions of interest, coverage, and similarity using fixation maps. Behavior Research Methods, Instruments, & Computers, 34(4), 518–528.

EXPERIMENTS | drawing experts are better able to accumulate object structure across eye- movements

4.3 ENCODING OF STRUCTURAL INFORMATION FROM INDIVIDUAL FIXATIONS

4.3.1 OBJECTIVES AND SUMMARY OF RESULTS In the first paper, we demonstrated that drawing skill is not linked to more veridical perception. Training, even for thousands of hours, does not change the basic processes of perception required for drawing. Instead, we suggested in this second paper that drawing skill may involve a more efficient accumulation of information into more robust internal representations. To be efficient, such visual analysis should be more suited to the actual constraints of drawing, which are not solely perceptual but also visuomotor. In particular, it has been previous suggested that skilled artists would encode more information from each individual fixation as they produce more strokes while fixating on the original for shorter duration and over smaller extent of space (Glazek, 2012). However, this study could not disentangle what particular regions of the visual field are actually encoded during each fixation.

The present study was designed to examine whether drawing experts could encode a larger extent of space around the fixation location (larger visual spans; e.g. Rayner, 1998), and maintain or increase this advantage for information presented to more peripheral locations. Information in the periphery is often surrounded by nearby clutter. Such a visual context is known to produce crowding (Whitney & Levi, 2011) whereby individual features or objects can no longer be isolated and identified. It is therefore possible that drawing skills may benefit from an enhanced ability to individualize features in a cluttered periphery – in other words, artists may be less vulnerable to crowding.

In a first experiment, participants had to categorize an object centered on the fixation location (monitored with an eye-tracker) as structurally possible or impossible. We varied the presentation duration of the object from 8 msec to 1500 msec in order to measure our participants’ encoding efficiency and the object size around the fixation point (8° or 28° of diameter) to measure the extent of space that could be processed at a single glance. Categorizing an object as possible or impossible implies a visual search for structural violations. Increasing the size of the object around the fixation location will also increase the distance between individual features and possibly the time needed to visit each location until the target feature is found. However, this effect of object size may be smaller in more skilled participants if drawing skill is related to the encoding of a larger visual span during each fixation. In a second main experiment, the possible or impossible object was displayed in the right visual periphery of our participants, while their fixation at the center of the screen was monitored to ensure they did not move their eyes toward the object. We manipulated the

EXPERIMENTS | Encoding of structural information from individual fixations object’s eccentricity in the periphery (3° or 8°) as well as its size (from 1° to 12°). For the same spacing between the object’s features (same object size), increasing the eccentricity will reinforce the effect of crowding. In contrast, increasing the object size will diminish such effect, helping participants to access and analyze the object’s individual features. If drawing skill relates to a better individuation of object’s features in periphery, experts should better perform in this task with smaller object sizes.

Here, we show that more skilled participants could overall discriminate object structure with shorter presentation durations, although showing the same advantage independently of the spatial extent to be processed at a single glance. This indicates a more efficient visual analysis of object structure. This advantage was also observed in our second experiment where more skilled participants could categorize smaller objects in periphery. Taken together, these findings demonstrate that drawing skill relates to a better encoding of structural information, regardless of whether the information has to be integrated across eye- movements (as in our previous study) or across information available within a single glance. In addition, our second experiment suggests that drawing skill may also relate to a better access to individual features present within a crowded visual environment – to a reduction in crowding.

4.3.2 DRAWING SKILL RELATES TO THE EFFICIENCY OF ENCODING

OBJECT STRUCTURE

This chapter is based on: Perdreau, F., & Cavanagh, P. (2014). Drawing skill is related to the efficiency of encoding object structure. I-Perception, 5(2), 101–119. doi:10.1068/i0635

Abstract. Accurate drawing calls on many skills beyond simple motor coordination. A good internal representation of the target object’s structure is necessary to capture its proportion and shape in the drawing. Here we assess two aspects of the perception of object structure and relate them to participants’ drawing accuracy. First, we assessed drawing accuracy by computing the geometrical dissimilarity of their drawing to the target object. We then used two tasks to evaluate the efficiency of encoding object structure. First, to examine the rate of temporal encoding, we varied presentation duration of a possible vs. impossible test object in the fovea using two different test sizes (8° and 28°). More skilled participants were faster at encoding an object’s structure, but this difference was not affected by image size. A control experiment showed that participants skilled in drawing did not have a general advantage that might have explained their faster processing for object structure. Second, to measure the critical image size for accurate classification in the periphery, we varied image size with possible vs. impossible object tests centered at two different eccentricities (3° and 8°). More skilled participants were able to categorize object structure at smaller sizes, and this advantage did not change with eccentricity. A control experiment showed that the result could not be attributed to differences in visual acuity, leaving attentional resolution as a possible explanation. Overall we conclude that drawing accuracy is related to faster encoding of object structure and better access to crowded details.

Keywords: artists, drawing accuracy, expertise, object structure, object perception, impossible objects

4.3.2.1 INTRODUCTION Drawing skills vary enormously across individuals, particularly when participants are asked to simply draw an object present in front of them. Here we will examine if this skill is related to the ability to process object structure when tested in a non-drawing context. The common excuse many of us give for our poor drawing performance is a lack of the motor coordination required to draw the object outlines. However, Tchalenko ( 2007) showed that novices were as accurate as experts in tracing simple straight and curved lines and experts were only more accurate when copying more complex arrays of lines (Tchalenko, 2009). If it is not simply an advantage in motor control, perhaps experts are able to perceive the target patterns more accurately (Cohen & Bennett, 1997). For example, studies using Navon stimuli (Chamberlain et al., 2013; Drake & Winner, 2011), have shown that observers who are skilled in drawing are better able to disregard the global shape of the stimulus when required

EXPERIMENTS | Encoding of structural information from individual fixations to report local shape. In line with this perceptual advantage, Mitchell, Ropar, Ackroyd, and Rajendran (2005) reported that the magnitude of perceptual errors made in a Shepard illusion task was inversely correlated with the participants’ drawing skills. Several studies extended this result to perceptual tasks involving shape (Cohen & Jones, 2008) and size constancy (Ostrofsky et al., 2012). However, these findings have been challenged by several failures to replicate (McManus, Loo, Chamberlain, Riley, & Brunswick, 2011; Ostrofsky et al., 2012 in their shape constancy task; Perdreau & Cavanagh, 2011, 2013a) calling into question the idea that skill in drawing is related to more accurate (veridical) perception. If accuracy in drawing is not a result of better perception (i.e. more veridical) or motor coordination, where does it come from? In a previous study (Perdreau & Cavanagh, 2013b) we suggested that drawing skill relies on the ability to construct and maintain a robust internal representation of object structure in visual memory (i.e. the relative spatial position of the object’s segments).

This robust internal representation is critical because observational drawing is characterized by many sequential eye-movements between the original object and the drawing, both to encode the to-be-drawn information and to visually control the hand position on the drawing (Coen-Cagli et al., 2009; Land, 2006; Locher, 2010; Tchalenko & Miall, 2009; Tchalenko, 2007, 2009). In particular, Tchalenko (2009) showed that experts in drawing used a segmentation process, encoding and copying the object line by line, while novices tended to process larger areas of the object. This is consistent with previous findings (Cohen, 2005) showing that artists shift their gaze more often between the original object and the drawing than beginners. According to these studies, a segmentation process selects salient portions of the object, reducing the amount of information to be encoded as well as the load on visual memory. Consistent with this hypothesis, Glazek (2012) found that experts in drawing drew more information than novices even when both groups saw the same extent of the target for the same duration. This suggests that more skilled participants may be able to encode more information (bigger chunks, e.g. Gobet & Simon, 1996) at each fixation. It remains unclear, though, how the various segments and chunks can be successfully integrated despite the many changes in retinal inputs and reference frames (Cai, Pouget, Schlag-Rey, & Schlag, 1997; Crawford et al., 2004).

In a previous study (Perdreau & Cavanagh, 2013b) relating drawing skill to object encoding, we examined the ability to discriminate possible from impossible objects (Schacter et al., 1991; Schacter, Cooper, & Delaney, 1990) while viewing only a portion of the object viewed through a window, centered on the fovea, that moved with the gaze. We found that more skilled and experienced participants could discriminate an objects’ structure based on a smaller visible portion of the object. These results suggested that the more skilled and

96 trained participants had a more robust mental representation of the object’s global organization that allowed them to add new portions from small samples acquired across several eye-movements. Many of the requirements of this contingent window task are shared with those of a drawing task – for example, sampling small segments of the object over many fixations and building an internal representation of its global structure.

However, our previous study could not rule out the possibility that the advantage of skilled participants was due to more efficient coding of the object’s structure “at a single glance” in addition to a more robust internal representation sequentially built across eye movements. Our present study will address this question of perceptual encoding (how much can be encoded at a glance) versus construction across eye movements. We compared how quickly object processing can operate in a single fixation and related it to participants’ drawing skills. We presented possible and impossible objects and varied the stimulus-mask SOA (from 8 ms to 1500 ms) as well as the extent of space covered by the object around the central fixation dot (8° or 28° of width, fixation controlled with an eye-tracker). This allowed us to measure the time required by our participants to encode object structure for a single fixation.

A second question that we addressed was the importance of peripheral information. Tchalenko (2007) has suggested that during drawing, even with gaze locked within a 5° range either around the pencil tip or the currently selected target segment, peripheral information may help to locate the current region of interest within the object’s global organization (e.g. using the paper’s edge that is visible in the periphery to check the line’s orientation). We assessed the role of peripheral information in a separate experiment in a previous study (Perdreau & Cavanagh, 2013b) where central vision was masked and participants could only see the periphery as they explored the test object. We found no advantage for the skilled participants but in that task participants could see only the periphery (masked central vision). In contrast, during a drawing task, the entire structure of the target object is visible so it is possible that peripheral information might help build the representation if central information is also present. Therefore, in addition to varying the stimulus duration in this first experiment, we also varied the size of the object, centered at fixation point (8° or 28°) to see how large an expanse of spatial information could be analyzed at a single glance (visual span, Rayner, 1998). If drawing accuracy is related to larger visual spans, then the more skilled participants may be better able to make use of peripheral information (i.e. beyond about 3° of eccentricity) giving them more of an advantage at the larger image size than the smaller.

EXPERIMENTS | Encoding of structural information from individual fixations

Finally, classifying an object as possible or impossible requires access to the individual features of the object’s structure, its lines and junctions, in order to construct its global representation. During free viewing, the different local features can be scanned with eye movements to access them, but in our experiments, participants had to fixate a central dot throughout the trial (monitored by an eye-tracker) so that stimuli were inspected without eye movements. This means that when adjacent features are closely spaced and presented in the periphery, they may be hard to access or individuate as they may be crowded. In a second experiment we examined the role of crowding by varying object size (1° to 12°) at two locations in the periphery (3° or 8°. crowding experiments are typically run at 3° or more in the periphery) to see if participants with better drawing skills were also better able to access closely spaced details in the periphery. It is well known that the presence of flankers surrounding a target can impair the identification, recognition and position encoding of a target (Greenwood, Bex, & Dakin, 2009; Martelli, Majaj, & Pelli, 2005; Toet & Levi, 1992; Whitney & Levi, 2011; Intriligator & Cavanagh, 2001). This phenomenon is commonly referred to as visual crowding and it has been described as a limit of the spatial resolution of visual attention (He, Cavanagh, & Intriligator, 1996, 1997).This effect is dependent on the spacing between items and the critical spacing to produce crowding is a constant proportion of the target’s eccentricity (Bouma, 1970). The crowding effect is seen not only for discrete items – targets and flankers – but also for inner features of complex stimuli like faces, where the resulting difficulty in recognizing the whole stimulus has been called self-crowding (Martelli et al., 2005). This self-crowding will impose a limitation on the access to local features required to build the object structure in our possible vs impossible task and this limitation should be particularly critical for peripheral presentation where crowding is stronger. Interestingly, it has been shown that attentional resolution, and the ability to individuate features within a cluttered visual environment in periphery, can be enhanced by an intensive training (e.g. in video game players; Green & Bavelier, 2007). For this reason, we suggest that one consequence of the years of drawing experience of participants with better drawing skills is to reduce the effects of crowding. They should be able access individual features of the objects at smaller image sizes in our impossible vs possible object task, when their close spacing will produce significant crowding and performance loss for the less skilled participants.

Although object size was manipulated in both experiments, we expected that size would have different effects in the two tasks. In the first experiment, the size was manipulated with the test object always centered at the fovea (Experiment 1a). This should not affect the degree of crowding because foveally centered scaling maintains the critical ratios of feature spacing to eccentricity (Bouma, 1970). However, the larger size will increase

98 the absolute inter-feature distance, which may require extra time to covertly scan the figure (Chakravarthi & VanRullen, 2011; Donnelly, Found, & Müller, 1999; Kosslyn, Ball, & Reiser, 1978; Posner, Petersen, Fox, & Raichle, 1988) especially if the image size exceeds the participant’s perceptual span (Rayner, 1998). In contrast, varying the size of an object that is centered in the periphery (Experiment 2a) will interact with crowding. Increasing its size in this case will increase feature spacing relative to eccentricity for the more peripheral features, diminishing the effect of crowding. We expected that increasing the size, and therefore the spacing between the object’s features, leads to better classification performances in this second experiment and that the critical size needed for accurate performance also increases with eccentricity (we tested 3° vs 8°). However, we hypothesized that more skilled participants have better attentional resolution – they can individuate features with closer spacing – and so they should be able to accurately classify the stimuli at smaller sizes. Again, if the skilled participants use peripheral information more efficiently (have larger visual spans), their performance may not decrease as much at larger eccentricities as that of the novices.

4.3.2.2 GENERAL METHOD

4.3.2.2.1 PARTICIPANTS Novice participants were all recruited from a pool of voluntary participants (RISC), whereas professional artists and art students were recruited in art schools and workshops (n=10; age=27.3±1.9, 6 females). Although the present study focused on drawing skill and not artistic ability, however that might be defined, we included artists and artists in training to increase the range of tested drawing skills. All the participants were naïve about the purpose of this study and our hypotheses. They all had a normal or corrected-to-normal vision. Before taking part in the experiments, they all gave their explicit written consent according to the principles defined by the University Paris Descartes ethics committee.

4.3.2.2.2 MATERIAL All the experiments, except for experiment 2B, used the same apparatus. Participant’s head was held by a chinrest so that his or her eyes were approximately 55 cm from the screen’s center. The stimuli were projected on a 22’’ CRT screen, with a resolution of 1024x768 pixels and with a frame rate of 120 Hz. The experiments were programmed in MATLAB using the Psychophysics and Eyelink Toolbox extensions (Brainard, 1997; Cornelissen, Peters, & Palmer, 2002; Pelli, 1997), and were run on an Apple computer.

Participants’ eye movements were recorded with an eye-tracking system (SR research Eyelink 1000 monocular, 35 mm lens) at 1000 Hz sampling rate. The eye-tracker was always calibrated for the participant’s dominant eye (9-points calibration. Eye dominance

EXPERIMENTS | Encoding of structural information from individual fixations was assessed with an aiming task). Finally, eye-movements events and saccades were parsed using the Eyelink 1000 algorithm (saccade acceleration threshold= 9500 deg/s², saccade velocity threshold=35 deg/s).

4.3.2.2.3 OBJECT SELECTION The object selection procedure we used in this study was similar to that presented in Perdreau and Cavanagh, (2013b). Part of the set of possible and impossible objects used in the present study was taken from previous collections (Schacter, Cooper, Delaney, Peterson, & Tharan, 1991; Soldan, Hilton, & Stern, 2009; used with authors' permission). The others objects were outlined versions of impossible objects provided by an Internet database (“Impossible world” website), and one of the authors (FP) designed corresponding possible versions using Adobe Illustrator CS4. At the end, our collection included a possible and an impossible version of 260 objects (total: 520) with an original size of 1667x1667 pixels.

In order to reduce the ambiguity about the structural possibility or the impossibility of the objects used in our experiments, we asked 20 independent observers, not participating in this study, to judge all our line-drawn objects as being structurally possible or not. They were particularly instructed that: objects were drawn with lines, every line represented a visible edge of the object, every visible edge was necessarily represented, an object’s surfaces could only face one direction and were opaque, objects were volumes standing in a 3D space. All objects were seen in a random order. Only objects that had an inter-observers agreement equal to or larger than 95% were kept and used in the experiments presented in this paper (a total of 317 objects of which 170 were possible objects).

In addition, a previous study reported that perceived impossibility may be related to the object’s complexity (Carrasco & Seamon, 1996). Thus, we computed complexity of every object according to the number of line segments and junctions. Objects complexity was normalized and objects were categorized as below (“Low complexity”) or above (“High complexity”) the median object complexity. Next, for each main experiment, we randomly selected 160 objects to create Low- and High-complexity levels for both possible and impossible object categories (40 objects in each category) such that the average complexity of Low-possible vs. Low-Impossible and High-possible vs. High-Impossible were not statistically different (p>.05). These categories were subsequently used to match object complexity across our experimental conditions.

100

Figure 1. Examples of structurally possible and impossible line-drawn objects used in the present study.

4.3.2.2.4 DRAWING ACCURACY To assess our participants’ drawing accuracy, we asked them to perform a drawing task where they had 15 minutes to copy a grayscale picture of an inverted house (Fig. 2A) as accurately as possible, that is, without emphasizing aestheticism or style. The use of an inverted house may avoid any canonical perception of the stimulus that could decrease drawing accuracy, even if the existence of such effect has not been demonstrated (Cohen & Earls, 2010). Moreover, without well-practiced elements to represent (e.g. an ordinary house), the artists, like the non-artists, would be showing their basic copying skills. The original picture was displayed on a computer screen and our participants had to copy it on a white A4 sheet of paper using a pencil. They were allowed to erase and correct their drawing as many times as they wished. Because the picture had many elements to draw, we asked participants to start by tracing the house’s structure, then to depict its details, and if they had enough time, to copy the house’s environment (trees, etc.). This made sure that all of our participants went through the same stages and had completed at least the first of them (house’s structure, which we used in our subsequent measurements).

Drawing accuracy is usually measured by asking independent observers to subjectively rate the drawings made by participants on a scale according to specific instructions (Cohen & Bennett, 1997; Cohen & Jones, 2008; Ostrofsky et al., 2012). Although this procedure can result in high inter-observers agreement, it does not allow the experimenter to know what criteria have been used by the observers, nor if these criteria vary across observers and across studies (Perdreau & Cavanagh, 2013a). Several recent

101

EXPERIMENTS | Encoding of structural information from individual fixations studies have presented novel geometrically-based methods to measure drawing accuracy, either using angles and proportion (Carson, Millard, Quehl, & Danckert, 2012; Chamberlain & McManus, 2013) or more sophisticated shape-matching algorithms (Hayes & Milne, 2011). In this study, we present a simple procedure based on the object’s structure (i.e. the relative spatial position of the object’s junctions). This single measure can account for angle accuracy, proportion and shape. First, we manually selected sixteen junctions present in all drawings that defined the house’s shape (Fig. 2A). Next, we centered both the original picture and the drawings on their leftmost selected junction. Moreover, because we hypothesized that object’s structure might be independent from the object’s overall size, as it is defined as the relative spatial position of the object’s parts, we normalized the junctions coordinates to the maximum horizontal and vertical coordinates in order to match the house size (Fig. 2B). Finally, we computed the drawing’s accuracy as the mean of the percentage root-mean-square error (%RMSE) computed on each axis (x and y) between the 16 points in the original and in the drawing:

(1)

Where Cd are the transformed coordinates of the copied house, Co the transformed coordinates of the original house, and n the number of selected points (n=16). Smaller root- mean-square errors mean better drawing accuracy.

Figure 2. Geometrically measured drawing error. (a) First, we manually selected 16 junctions present in all the drawings (colored dots) and the original picture, which defined the house’s shape and structure. (b) Next, we normalized their coordinates to the maximum

102

horizontal and vertical coordinates, and computed the mean RMS error of each drawing on the basis on these transformed coordinates.

We next evaluated this geometrical measure by comparing it to human observers’ subjective judgments of accuracy. To do so, we designed an online experiment where randomly selected pairs of drawings were presented to participants (n=245) as well as the original picture. Online experiments are known to yield to a similar data quality compared to experiments run in laboratories (Germine et al., 2012). The task was to choose, by clicking on it with the mouse, which of the two drawings more precisely matched the original house. Participants were instructed to make their judgment on the basis of the house’s proportion and structure, and not on the amount of detail, the size, style or aestheticism of the drawing. Each participant saw a maximum of two hundred pairs of drawings and each possible pair was compared by at least 40 independent participants. To rank each drawing, we used the ELO ranking algorithm developed to rank players in two-competitor games like chess (Elo, 1978). The algorithm computes a probability of win for each compared item (a vs. b) according to its past results and to the score of the other item in the comparison:

(2)

where diff is the difference between item a and b’s previous ELO scores:

(3) This probability is then used to compute the item a’s new score:

(4)

where W is the item’s outcome for the current comparison, a vs b, (W = 1 for a win, 0.5 for a draw, and 0 for a loss) and n is the number of previous comparisons for the item. These scores for each item reach a final value across all the comparisons made for that item across all the participants.

These subjective rankings were highly consistent with the original geometrical measures [r(35) = -0.68 (CI: -0.82, -0.49), p<.0001]. Because of the objective nature of the geometrical scores, we used them in our evaluations of the participants’ performance in the experiments that follow.

103

EXPERIMENTS | Encoding of structural information from individual fixations

Figure 3. Correlation between objectively (x axis) and subjectively (y axis) measured drawing accuracy. Each dot represents a participant’s drawing. The filled area is the 95% bootstrapped confidence interval of the regression line.

4.3.2.3 EXPERIMENT 1A: VISUAL MASKING The purpose of this experiment was to test two hypotheses: 1) whether drawing accuracy is related to the efficiency of encoding object structure as measured by the threshold presentation duration of the stimulus prior a mask, and 2) whether drawing accuracy is related to the extent of space that can be encoded at a single glance.

4.3.2.3.1 PARTICIPANTS Thirty participants took part in this experiment [average age 24.4±0.8, 15 females, 10 artists, of which 6 were female]. Three of these participants were dropped from the analysis as their performance in the main experiment never reached threshold in the tested range and no psychometric function could be fit to their results.

4.3.2.3.2 STIMULI We generated a list of 160 randomly selected objects from our original collection (40 objects per category and complexity level) that was identical for every participant. All objects were repeated three times during the experiment and they were randomly distributed so that they could not appear within the same experimental condition and with the same orientation (θ = 0°, 90° or 180°).

104

4.3.2.3.3 PROCEDURE Prior to the main experiments, participants performed in a practice session to familiarize themselves with the task and the difference between structurally possible and impossible objects. We first instructed our participants, like those participating in the object selection procedure, that: objects are drawn with lines, every line represents an edge of the object, every edge is necessarily represented, object’s surfaces can only face one direction and are opaque, and objects are volumes standing in a 3D space. All objects used in this practice session were different from those presented in the other experiments. Next we showed some examples of possible and impossible objects and asked the participants whether this categorization made sense to her or to him. Finally, participants ran in 80 practice trials. On each trial, a 21°x21° line-drawn object was displayed centered on the screen. Participants had to answer, as fast and accurately as possible, whether this object was structurally possible or not. They had a maximum of 5 seconds to give their answer. Otherwise, the screen was blanked out and the participant was told to give a response.

In the main experiment, eye-movements were monitored using an eye-tracking system. Each trial started with a central fixation dot that the participant had to fixate for 200 ms to start the trial (Fig. 3). After a random delay of 600-900 ms, a line-drawn object was displayed centered on the screen. We varied the object’s presentation duration from 8 to 1500 ms (eight SOAs: 8, 25, 67, 183, 525, 1500 ms), as well as its size (8° or 28° of visual angle). After this duration, the object was immediately replaced by a dynamic mask with a duration of 16 frames (Bacon-Macé, Macé, Fabre-Thorpe, & Thorpe, 2005). Finally, a red central dot was displayed, indicating to the participants that he or she had a maximum of 2 seconds to respond. Participants had to fixate the central dot during the entire trial. If participant’s gaze deviated by more than 1° from this dot, the trial was ended and was replayed later in the experiment. Specifically, participants were told to report whether the object was structurally possible or impossible, as fast and accurately as possible, by pressing the appropriate key (“left control” or “right control”, respectively). Feedback was shown to participants every 20 trials to indicate their overall performance as well as their progression. This was designed to keep the participants motivated.

We manipulated the object’s presentation duration (6 timings, from 8 to 1500 msec) as well as its size (8° or 28°) and its complexity level (low or high) as within-subject factors, although the latter factor was only treated as a control variable for an effect of complexity on perceived impossibility.

The experiment started with a practice block of 24 trials where we presented every possible condition to the participants. This was followed by 480 trials divided into 6 blocks of

105

EXPERIMENTS | Encoding of structural information from individual fixations

80 trials each. We calibrated the eye-tracking system and made a drift correction at the beginning of each block.

Figure 4. Experiment 1’s procedure. Participants had to fixate a central dot for 200 ms to start the trial. After a random delay (600-900 ms), the stimulus was displayed, centered on the screen. We manipulated stimulus presentation duration (from 8 ms to 1500 ms) and size (8° or 28°). Participants had to maintain their fixation on the central dot throughout the trial (within ±1°).

4.3.2.3.4 RESULTS We first compared the participants’ pretest performance (percent correct categorization of the 80 possible and impossible objects without eye movement constraints) to their drawing accuracy. The two measures showed a correlation of 0.45 [p<.005], suggesting that more skilled participants could better discriminate possible from impossible objects in a free viewing condition. This preliminary correlation should be viewed with some caution as a similar baseline condition in our previous study (Perdreau & Cavanagh, 2013b) did not show a significant correlation, again between drawing accuracy and performance. One difference between the two studies was that the time available for the response was shorter (5 seconds) here than in the previous study (10 seconds). It is possible that this extra time in the previous experiment was sufficient to allow the less skilled participants to catch up to the performance of the skilled participants.

In the analysis of the main experiment, we first fitted a logistic function on proportions of correct responses plotted against SOAs (Fig. 4A) for each participant and object size (8 and 28°), using a maximum likelihood procedure with a fixed gamma parameter of 50% (2- AFC. Prins, 2012). The overall quality of individual fits was good and reliable [mean deviance: 3.12(SE: 0.29), Ps>>.05]. Next, we computed the critical SOA for each participant

106 and object size, defined as the SOA value leading to 75% correct responses. However, three of our participants had very poor performance in this experiment despite good performance in the pretest, which led to undefined (infinite) thresholds. We excluded these three participants from the following analyses.

To measure whether better accuracy in drawing is related to faster encoding of object structure, and to an ability to make such encoding over a larger extent of space, we computed the regression between participants’ drawing error score and their critical SOA using a linear mixed-effects1 model on log-transformed2 SOA and error values with drawing error and object size as fixed factors and with participants as a random factor to account for our repeated-measures design. We found a significant linear relation between log-error and log-critical SOA [β=1.39 (SE: 0.44), χ²(1)=8.59, p<.004] and a significant main effect of object size [β=1.08 (SE: 0.51), χ²(1)=14.26, p<.0002]. However, we found no significant interaction between object size and drawing accuracy [β=-0.44 (SE: 0.37), χ²(1)=1.38, p=.24]. These results show that more skilled participants needed overall shorter presentation durations to make accurate decision about the object’s structure. Although noisy, the effect was quite large with the regression showing a 3 to 5-fold increase in the critical duration required to classify the objects for least skilled versus most skilled participants. This relation was not affected by object size.

1 Linear mixed-effects models are extensions of linear regression models, including both fixed effects (linear regression part) and random effects that allow random individual variations across a grouping factor’s levels (as in repeated-measures ANOVA). Linear mixed-effects models presented in this study were computed using the lme4 R package with a ML procedure (Bates, Maechler, & Bolker, 2012). Main effects and interaction effects were tested by using a likelihood ratio-test, which compares the residual deviance of both the full model and the reduced model (full model without the factor of interest) and approaches a χ² distribution. In addition, we report the log-log regression slope (β) to characterize the effect direction. 2 A diagnostic of the linear model’s residuals showed an heavy heteroscedasticity when we used raw values. This can be explained by the presence of a lower bound in both participant’s critical SOAs and drawing accuracy (theoretical minimum at 0). Using log-transformed values solved this issue. We applied this data transformation in all of the following analysis.

107

EXPERIMENTS | Encoding of structural information from individual fixations

Figure 5. Experiment 1A and 1B results. (a, c) Examples of individual fits for the object and word tasks. We fitted logistic functions on participant’s proportion of correct responses plotted against SOA and we measured a critical SOA for each object size (8° and 28°) as the SOA leading to 75% correct response. Horizontal error bars are the 95% bootstrapped confidence intervals of the estimated critical SOA. (b, d) Next, we computed regressions between drawing error and critical SOA using a linear-mixed effects model for both tasks. Filled areas are the 95% bootstrapped confidence intervals of the regression lines.

4.3.2.4 EXPERIMENT 1B: LEXICAL-DECISION The more skilled participants needed less time to make accurate judgments of possible vs impossible objects. However, this might reflect some general advantage of the skilled participants for any task. As a control for a general processing advantage, we ran a lexical-decision task that required integration of letters across space to determine if the letters made a possible or impossible word. Other than the nature of the stimuli, the experiment had the exact same conditions as the possible-impossible object task. If participants with better drawing skills had a general advantage, it should also be apparent better in this lexical decision task.

108

4.3.2.4.1 PARTICIPANTS Participants were identical to those of the previously described experiment, with the exception that only the 28 French native-speakers of the original 30 participants ran in this control experiment. Moreover, to allow a more straightforward comparison between this experiment’s results and the previous findings, we also excluded the two outliers excluded from Experiment 1a because of indeterminate thresholds (leaving n=26) although this did not affect the results.

4.3.2.4.2 STIMULI Experimental conditions were identical to the visual masking task with possible/impossible objects. However, instead of using these line-drawn objects we used French words and non-words. These words and non-words were collected from a French database (LEXIQUE 3.55; New, Pallier, Brysbaert, & Ferrand, 2004). More particularly and to avoid any ambiguities of an effect of words frequency, and possibly of participants’ education, on participants’ performances, we selected words with a frequency greater than 100 times per million words, based on appearances in books, subtitles, webpages, etc. Likewise, for non-words we selected actual French trigrams (sequences of letters making a pronounceable non-word) that had the same mean frequency of occurrence in French as that of the words we used in this task. Finally, to match the first experiment’s conditions, we manipulated the number of letters of words and non-words (4 or 8) to simulate words visual complexity as well as their horizontal size (8° or 28°). As in the previous experiment, words complexity was matched across presentation durations. We had a total of 90 words and 90 non-words, each of them having been seen 3 times during the experiment.

4.3.2.4.3 PROCEDURE The procedure was identical to that used in the first experiment, with the exception that now participants were asked to categorize the briefly displayed word as a word present in French language or not.

4.3.2.4.4 RESULTS As previously, we fitted logistic functions on participants’ proportions of correct responses, plotted against words presentation duration, and we computed a critical SOA for each participant and each stimulus size (Fig 4C). The overall quality of fit was very high [mean deviance: 2.45(SE: 0.29), Ps >> .05].

Next, we computed the regression between drawing accuracy and critical SOA using a linear mixed-effects model on participants’ critical SOAs with participants’ drawing error and word size as fixed factors, participants as random factor. We found a significant main

109

EXPERIMENTS | Encoding of structural information from individual fixations effect of word size on participants’ critical SOAs [β=0.99 (SE: 0.33), χ²(1)=132.62, p<.00001], suggesting that participants overall need longer presentation duration to encode larger words accurately. In contrast to the first main experiment, we found no main effect of drawing error [β=-0.11 (SE: 0.29), χ²(1)=0.28], as well as no interaction between drawing error and word size [β=-0.06 (SE: 0.24), χ²(1)=0.07].

Taken together, these results show that the more efficient processing of object structure by skilled participants that we found in the first main experiment could not be attributed to a more efficient encoding of visual stimuli in general.

4.3.2.5 EXPERIMENT 2A The first experiment’s results suggest that drawing accuracy might be related to a more efficient encoding of object structure seen in a single fixation but not to the ability to encode this structural information over larger visual spans without eye movements. This latter finding might not be surprising, for drawing is characterized as a sequential, segmentation process (Tchalenko, 2009). For instance, an artist might need to encode the segment’s position only in the context of elements located in a particular area of the segment’s immediate surroundings. The present experiment aims at testing whether drawing accuracy would be related to a more efficient encoding of structural information located in participants’ visual periphery.

4.3.2.5.1 PARTICIPANTS A total of thirty-four participants ran in this experiment [average age 25.9±1.1, 12 females, 10 artists of which 6 were female]. Four of these participants were dropped from the analysis (leaving 30) as their performance in one of the two main conditions never reached threshold in the tested range and no psychometric function could be fit to their results.

4.3.2.5.2 STIMULI In this experiment, we only used the subset of “low-complexity” line-drawn objects from our collection, as a pilot experiment revealed that the task was nearly impossible to perform in the periphery with the most complex objects. We used a total of 60 objects per category (possible and impossible) that were repeated 4 times during the experiment, but never within the same condition or with the same orientation.

4.3.2.5.3 PROCEDURE Participants had to fixate a central dot for 200 ms to automatically start the trial. After a random delay varying between 600 and 900 ms, a line-drawn object was displayed in the right visual periphery of the observer. Participants were told to fixate a central dot all along

110 the trial. Eye movements were continuously recorded with an eye-tracker, and every time participants’ gaze deviated by more than 1° from this central dot, the object disappeared. Participants were asked to categorize the object as structurally possible or impossible, and they had 5 seconds maximum to give their response. They were asked to respond as fast and accurately as possible by pressing the appropriate key (“left control” or “right control”). If no response was given within this time, the screen was blanked out and the participant was told to give a response. We varied the object size (8 sizes from 1° to 12°) as well as its eccentricity from the central dot (3° or 8°).

Figure 6. Experiment 2’s procedure. Participants had to fixate a central dot for 200 ms to start the trial. After a random delay of 600-900 ms, a line-drawn object was displayed in the participant’s right visual periphery. It remained on the screen until the participant gave his or her response. They were asked to categorize the object as structurally possible or impossible, and they had 5 seconds maximum to respond. If no response was given within this time, the screen was blanked out and the participant was asked to answer. In this experiment, we particularly varied the object’s size (from 1° to 12°) and its eccentricity (3° or 8°).

Each participant started the experiment with a training block of 32 trials, which crossed all experimental conditions. Then, participants ran in 480 trials (8 sizes x 2 eccentricity, with 30 trials per condition), which were divided into 4 blocks. As in the first experiment, participants received feedback about their overall performance and progression every 20 trials. They were free to take a break during these feedback pauses and between every block of trials.

111

EXPERIMENTS | Encoding of structural information from individual fixations

4.3.2.5.4 RESULTS Logistic functions were fitted on participants’ proportion of correct responses plotted against object’s size for each tested eccentricity [mean deviance: 6.24(SE: 0.49), Ps>.05]. We then measured the critical size as the object size leading to 75% correct responses.

To measure how the efficiency of discriminating possible from impossible figures varies with eccentricity and drawing accuracy, we computed regressions between drawing error and critical object size using a linear mixed-effects model on the log-transformed participants’ critical size, with log drawing error and object eccentricity as fixed factors, participants as random factor. Not surprisingly, we found a significant main effect of eccentricity on participants’ critical object size [β=1.16 (SE: 0.62), χ²(1)=23.97, p<.0001], suggesting that participants needed in average larger object sizes to accurately discriminate between possible/impossible figures at farther eccentricities. As in the first experiment, there was a significant linear relation between drawing error and critical object size [β=1.04 (SE: 0.40), χ²(1)=7.43, p<.007], suggesting that more skilled participants needed on average smaller sizes to categorize an object as structurally possible or impossible when presented in visual periphery. Again, although noisy, the effect of drawing skill was large, with the regression showing a 2 to 3-fold increase in the required object size required for accurate classification between the least skilled and most skilled participants. Finally, we found no significant interaction between object eccentricity and the relation between drawing error and participants’ critical sizes [β=-0.30 (SE: 0.49), χ²(1)=0.39, p=0.53].

One explanation for the ability of more skilled participants to accurately process smaller objects is that they have smaller critical zone for crowding – they can access local features when they are more closely spaced. However, this effect of drawing skill did not vary with eccentricity. This advantage for skilled participants in processing smaller stimuli is surprising at a first glance, since we found no difference in the effect of object size for skilled vs unskilled participants in our first experiment. However, as we stated in our introduction, object size has different effects for an object centered at the fovea, as was the case in Experiment 1, compared to one centered in periphery, as here in Experiment 2. In the first case, size changes should not affect the strength of inter-feature crowding because the ratios of feature spacing to eccentricity that determine crowding are maintained when size changes are centered at the fovea. In contrast, with objects centered in the periphery, size changes do affect the feature-spacing to eccentricity ratios. Therefore, in line with our hypothesis of a better attentional resolution for skilled participants, one explanation of their advantage is that they were better with smaller stimuli at finding and individuating features required to construct the object’s structure and detect violations.

112

Figure 7. Experiment 2A’s and 2B’s results. (a) Example of individual fits. As in the first experiments, we fitted logistic functions on participant’s proportion of correct responses plotted against SOA and we measured a critical size for each eccentricity (3° and 8°) as the object size leading to 75% correct response. Horizontal error bars are the 95% bootstrapped confidence intervals of the estimated critical SOA. (b) Next, we computed the regressions between drawing error and critical size. Filled areas are the non-parametric 95% bootstrapped confidence intervals of the regression slopes (10,000 runs). (c) To measure participants’ visual acuity, we used a Landolt’s C test and we measured the critical gap size for each eccentricity with an adaptive staircase procedure. (d) The regressions between drawing error and critical size for the Landolt’s C task. Filled areas are the non-parametric 95% bootstrapped confidence intervals of the regression slopes (10,000 runs).

4.3.2.6 EXPERIMENT 2B The relationship we found between drawing accuracy and the participants’ critical object size could easily be due to better peripheral visual acuity in more skilled participants. To evaluate this alternative explanation, we ran a Landolt’s C test at the same eccentricities.

4.3.2.6.1 PARTICIPANTS Participants were identical to those who participated in the previous experiment (2a), and we again excluded the four who had been removed from the analysis in that experiment (n=30).

113

EXPERIMENTS | Encoding of structural information from individual fixations

4.3.2.6.2 MATERIAL The participant’s chin was held by a chinrest so that his or her eyes were centered on the screen at a distance of 105 cm. In this experiment, we used a 30-inch flat screen (30-inch Apple Cinema HD Display), with a spatial resolution of 2560x1600 pixels and a refresh rate of 60 Hz. This setting allowed us to display the very small visual angles required by our task.

4.3.2.6.3 PROCEDURE Each trial started with a fixation dot centered on the screen for a random time of 300- 600 ms. Next, a Landolt’s C was briefly flashed for 100 ms in the right visual periphery. Once the central dot disappeared, the participants could give his or her response. Participants were asked to report the position of the gap in the C (left, right, top or bottom) by pressing the corresponding arrow key. In this experiment, we manipulated the C’s eccentricity (3° or 8°) as well as the C’s gap size. The latter was determined by using 2 interleaved and independent adaptive staircases (adaptive stochastic approximation procedure; Kesten, 1958) for each eccentricity, one starting at 0.013° and the other at 0.20°. The aim of these staircases was to determine the threshold C’s gap size for which participants’ performance on this 4-AFC task was 62.5% correct (chance level: 25%).

Figure 8. Experiment 2B’s procedure. Each trial start with a fixation dot displayed for a randomly chosen between 300 and 600 ms. Then a Landolt’s C was briefly flashed for a duration of 100 ms. Once it disappeared, participants could give their response. We manipulated the position of the gap in the C (top, right, bottom or left) and participants had to report this position by pressing the corresponding arrow key. We varied the C’s eccentricity as well as its gap size (proportional to the overall size). For each tested eccentricity, two independent and intermixed staircases were used to measure the gap size leading to 62.5% correct responses (4-AFC).

114

4.3.2.6.4 RESULTS In order to assess the relationship between our participant’s drawing accuracy and their visual acuity, we computed the regression between drawing error and threshold gap size using a linear mixed-effects model on log-transformed thresholds, with eccentricity and drawing error as fixed factors, participants as random factors. As expected, we found a significant main effect of eccentricity on participants’ thresholds [β=0.25 (SE: 0.22), χ²(1)=32.51, p<.00001]. However, we found no significant effect of drawing error [β=-0.11 (SE: 0.15), χ²(1)=0.43] and no interaction between drawing error and eccentricity [β=0.06 (SE: 0.17), χ²(1)=0.12].

These results show no evidence of a relationship between drawing error score and participants’ visual acuity in visual periphery. Therefore, it is unlikely that the advantage of more skilled participants in discriminating possible from impossible objects in the periphery could be due to a more accurate perception of these objects.

4.3.2.7 GENERAL DISCUSSION Do artists and others skilled at drawing have any general advantages in processing visual information? Several previous experiments have suggested that they do (e.g. Glazek, 2012; Perdreau & Cavanagh, 2013b; Tchalenko, 2009). In many of these experiments, however, participants were allowed to scan a test image and we cannot know if the performance differences arose from advantages in integrating information across eye movements or from advantages in encoding information from individual fixations. In our experiments here we asked whether participants with better drawing skills would have any advantages when their access to the stimulus was limited to a single glance (no scanning eye movements allowed). Participants of varying levels of drawing skill classified line drawings as possible or impossible objects: in the first experiment, presentation time was varied to measure the speed of processing object structure as well as the spatial extent over which this processing could be accomplished at a single glance; in the second experiment, object size was varied at different locations in the visual periphery to see whether participants with better drawing skills could correctly classify stimuli at smaller sizes or further eccentricity.

First, we measured our participants’ drawing errors both objectively and subjectively. We extended an objective method that geometrically compared the participants’ drawings to the original picture and quantified the drawing’s matching error (e.g. Carson et al., 2012; Chamberlain & McManus, 2013; Hayes & Milne, 2011). In addition, we conducted a large- scale, on-line experiment where drawings were pair-wise compared by participants and then ranked using an ELO ranking algorithm (Elo, 1978). Both measurements were consistent and

115

EXPERIMENTS | Encoding of structural information from individual fixations in the main experiments we used the objective measure where the criteria for accuracy are explicit.

In the first experiment (Experiment 1a), we found a significant relationship between objectively measured drawing error and the presentation duration needed by our participants to accurately categorize an object as structurally possible or impossible. Specifically, the regression result indicated that the most skilled participants reached threshold performance with presentation durations 3 to 5 times shorter than the threshold durations for the least skilled. However, the size of the stimulus (8° vs 28°) did not affect this difference between skilled and unskilled participants. We also ran an additional control experiment using the exact same experimental conditions but using a lexical decision task instead of an object decision task. This control assessed whether the performance advantage of the more skilled participants was specific to the object structure discrimination or whether they were related to more general cognitive abilities that affected all tasks including drawing and the object structure task. However, no significant relationships between drawing skill and performance were found in the lexical-decision task (Experiment 1b). So drawing accuracy appears be related to a more efficient processing of object structure on a single glance. In contrast, more skilled participants were not any better at processing larger stimuli than those who were less skilled suggesting that they had no advantage in terms of visual span (Rayner, 1998).

In the second experiment, we found a significant relationship between drawing error and the object size needed to correctly identify its structure as possible or impossible. The regression results showed that the most skilled participants reached threshold performance for stimuli that were 2 to 3 times smaller than the threshold size for the least skilled. The test objects were presented at different locations in the near periphery (3° and 8°) so one explanation of the advantage for the participants with better drawing skills was that they had better visual acuity – they might simply be able to see the stimuli more clearly. To test this, we conducted a Landolt’s C task at the same tested eccentricities and found no differences related to drawing skill, ruling out acuity as an explanation. In the absence of differences in visual acuity, a possible explanation of the better performances of our skilled participants with smaller images may be that they were able to individuate features present within small local structures. This finding suggests a relationship between drawing skill and attentional resolution (He et al., 1996, 1997).

In summary, our results showed that the more skilled participants were notably faster at processing object structure at a single glance and were able to accurately classify test stimuli at much smaller sizes. The advantage in speed was independent of the size of the object and the advantage in size was independent of the eccentricity. This strongly suggests

116 that drawing accuracy may be related to a more efficient encoding and representation of structural information that can be captured during a single fixation, regardless of the viewing condition (either in central or peripheral vision).

4.3.2.7.1 DRAWING ACCURACY AND POSSIBLE/IMPOSSIBLE OBJECT

DISCRIMINATION Before considering how the processing of object structure differs between participants who are skilled at drawing and those who are not, we first consider what the possible vs impossible task tells us about the processing of object structure. Most of the studies using the impossible/possible object-decision test investigated object representation in long-term memory (e.g. Schacter, Cooper, & Delaney, 1990; Schacter et al., 1991; Soldan et al., 2009), and only a few of them have considered the perceptual aspects underlying this discrimination (Carrasco & Seamon, 1996; Donnelly et al., 1999; Freud, Avidan, & Ganel, 2013; Freud, Ganel, & Avidan, 2013; Seamon & Carrasco, 1999; Williams & Tarr, 1997). Impossible objects have valid object properties (e.g. closed contours, volume, edges) despite their inherent structural violations. This may explain why possible and impossible objects share the same early perceptual processes with possible objects (Freud, Ganel, et al., 2013), as well as why impossible objects can be processed holistically to the same extent as structurally valid objects (Freud, Avidan, et al., 2013), and can be perceived as valid objects (Cowie & Perrott, 1993; Williams & Tarr, 1997). However, discriminating possible from impossible structures takes much more time than discriminating object categories (e.g., animal vs vehicle) or identities (e.g., spoon vs fork). For example, our participants needed almost 1000 ms on average to accurately discriminate an impossible from a possible object, whereas objects or scenes can sometimes be classified in less than 100 ms (Bacon-Macé et al., 2005; Greene & Oliva, 2009; Thorpe, Fize, & Marlot, 1996). Such a difference may mean that possible/impossible object discrimination is not based on the entry–level object “gist”, but rather on a scrutiny of its parts and on a serial verification of the consistency between these parts and the object’s global structure (Donnelly et al., 1999; Soldan et al., 2009). These verification processes undoubted require more extensive and higher level processing (Hochstein & Ahissar, 2002). There are undoubtedly several factors that would contribute to the duration of these serial verifications: for example, the number of locations covertly sampled, the time spent to analyze the features at each location, and the time needed to move from one location to the next.

In line with this hypothesis of serial inspection and verification and consistent with previous findings (Donnelly et al., 1999), we found that all participants needed longer durations for an object that covered a larger extent of space (Experiment 1a). Since fixation

117

EXPERIMENTS | Encoding of structural information from individual fixations was maintained throughout, we might expect that object size would not affect the number of locations visited and scrutinized by attention (e.g. Treisman & Gelade, 1980) nor the analysis of the features at each location. The effect of size might arise from the longer time required to move attention from one location in the object to the next in a larger object (Chakravarthi & VanRullen, 2011; Donnelly et al., 1999; Kosslyn et al., 1978; Posner et al., 1988). The speed advantage of the participants who are skilled at drawing (Experiment 1a) may arise from better performance on any or all of the three factors we mentioned: they may be able to construct the object structure based on covertly sampling fewer locations; they may spend less time processing the information from each location; and they may be able to move attention more quickly from one location to the next. However, our experiment was not designed to distinguish the separate effects of these three factors and further research will be needed to address this issue. Nevertheless, we were able to show in the second experiment, that more skilled participants could access and analyze more closely spaced details than could the less skilled participants.

4.3.2.7.2 RELATING DRAWING ACCURACY TO THE OBSERVED PERFORMANCES Our results here are specific to object analyses performed at a single glance, so we must ask, among the visual processes that would become more developed in artists and those skilled at drawing, which of these would be specific to analysis completed without eye movements.

Let us start by roughly outlining the stages of making a drawing. First an internal representation of the original object must be constructed, shifting back and forth between local and global levels, to select and analyze local features in the context of their location within the object (Chamberlain et al., 2013; Kozbelt et al., 2010). On first glance, it is this stage that appears most relevant to our possible vs impossible objects task although typically, this stage would involve many eye movements across the original image to build its representation. The drawing then requires a selection of local elements to draw (Kozbelt et al., 2010; Tchalenko, 2009), adding them to the already drawn depiction, maintaining spatial relations. This selection and verification of local elements depends on understanding of the global structure and the relation of the selected element to that structure. If the original is in view, frequent eye movements are made back and forth between the original and the drawing in progress (Tchalenko & Miall, 2009; Tchalenko, 2009) to check details in the original and guide the new additions. If the original is not in view, then the back and forth inspection and drawing is based on the memory representation of the original. In either case, there is again a dependence on a robust internal representation of the object’s structure and a requirement to access individual local elements to see if they correspond properly to the expected spatial relations.

118

Much of this effort involves back and forth eye movements with sampling of information from local regions and integration and comparison of this information with the internal model of the object’s global structure. Several previous articles have looked at the link between drawing and visual tasks where eye movements are allowed (e.g. Glazek, 2012; Mitchell et al., 2005; Perdreau & Cavanagh, 2013b) and integration across fixations plays a role. For example, our previous study (Perdreau & Cavanagh, 2013b) focused on the same possible vs impossible visual task with free eye movements but restricting the amount of the test object that could be seen around fixation. The results showed that participants who were skilled in drawing could perform accurately with smaller samples of foveal information. This suggests that drawing accuracy may relate to a more robust internal representation with local features better integrated into more complex chunks.

What is different when eye movements are not allowed, when the analysis must take place in a single fixation? Clearly, rather than integrating samples acquired foveally from one fixation to the next, participants must integrate the information encoded at different eccentricities by moving attention over different locations of interest. We again find that participants skilled in drawing are better at this. The same factors should contribute to good performance here: more robust internal representation, more efficient coding and better integration of local features. The continued superiority of skilled participants with or without eye movements suggests that the management of eye movements and retention of information across fixations is not the critical factor in their advantage. Nor is the difference between foveally acquired local features and peripherally acquired features. The skilled participants appear to be better wherever they acquire object features. They also have an additional advantage for closely spaced features, possibly based on better attentional resolution (He et al., 1996, 1997) that allows them to access closely spaced details that are crowded for less skilled.

We conclude that those skilled at drawing are able to construct better internal representations of objects in less time and are better able to access details despite local clutter from adjacent features. These processes are critical in the rapid analysis of images when producing drawings and the long experience of relying on these processes for making drawings may have altered these core visual and memory processes. Of course, it is also possible that those who are skilled at drawing have those skills because these processes were already more developed. We cannot resolve this question of causality without a longitudinal study following, say, art students across their years of training. Our previous study of the impossible vs possible objects task did find a relation between years of drawing experience and performance that suggested that experience was a causal factor, but it was not a longitudinal study of the same participants across their training so it was not conclusive.

119

EXPERIMENTS | Encoding of structural information from individual fixations

In contrast to these factors that do seem to link drawing to performance in this possible vs impossible object task, we found no evidence for additional advantages for skilled participants in processing more peripheral information. Drawing accuracy was not related to any increase in performance over larger visual spans (8° vs 28°, Experiment 1a) nor to improved processing at larger eccentricities (3° vs 8°, Experiment 2a). This result is consistent with the absence of a relationship between drawing accuracy and the ability to make structural discrimination based only on peripheral information in our previous study (Perdreau & Cavanagh, 2013b). It appears that whatever the relation between drawing skill and visual processes, it does not include advantages in processing information from the periphery (the advantage that we saw in dealing with crowded details was similar over the two eccentricities tested). However, it is possible that a difference would have emerged if we had tested a larger range of eccentricities.

To conclude, our present results suggest that drawing skill is related to improved visual processes active in a single glance, including a combination of more a robust internal representation, more efficient coding, better integration of local features, and better access to closely spaced details. These results indicate that the advantages seen previously in tasks where eye movements were allowed (e.g. Glazek, 2012; Mitchell et al., 2005; Perdreau & Cavanagh, 2013b) were not critically dependent on differences in eye movement related processes, but demonstrate general advantages in processing object structure no matter what the context. Our results here are only correlational, however, so there is no evidence of causality. We did not find any evidence for a link between drawing skill and a larger visual span (Rayner, 1998) or a more efficient use of peripheral information. We did find evidence suggesting better attentional resolution (less crowding) for skilled participants and although this is advantage only becomes measurable in the periphery (there is no crowding in the fovea), this advantage did not increase over the range eccentricities that we tested.

4.3.2.8 REFERENCES

Bacon-Macé, N., Macé, M. J.-M., Fabre-Thorpe, M., & Thorpe, S. J. (2005). The time course of visual processing: backward masking and natural scene categorisation. Vision Research, 45(11), 1459–69.

Bates, D., Maechler, M., & Bolker, B. (2012). lme4: Linear mixed-effects models using S4 classes.

Bouma, H. (1970). Interaction effects in parafoveal letter recognition. Nature, 226, 177–178.

Brainard, D. H. (1997). The Psychophysics Toolbox. Spatial Vision, 10(4), 433–436.

120

Cai, R. H., Pouget, A., Schlag-Rey, M., & Schlag, J. (1997). Perceived geometrical relationships affected by eye-movement signals. Nature, 386(6625), 601–604.

Carrasco, M., & Seamon, J. G. (1996). Priming impossible figures in the object decision test : The critical importance of perceived stimulus complexity. Psychonomic Bulletin & Review, 3(3), 344–351.

Carson, L., Millard, M., Quehl, N., & Danckert, J. (2012). Drawing expertise predicts not just quality but also accuracy. Perception, 41(ECVP Abstract Supplement), 231.

Chakravarthi, R., & VanRullen, R. (2011). Bullet trains and steam engines : Exogenous attention zips but endogenous attention chugs along. Journal of Vision, 11(4), 1–12.

Chamberlain, R., & McManus, I. C. (2013). Subjective and Objective Measures of Drawing Accuracy and their Relationship to Perceptual Abilities. Perception, 42(ECVP Abstract Supplement), 106.

Chamberlain, R., McManus, I. C., Riley, H., Rankin, Q., & Brunswick, N. (2013). Local processing enhancements associated with superior observational drawing are due to enhanced perceptual functioning, not weak central coherence. The Quarterly Journal of Experimental Psychology, 66(November), 1448–66.

Coen-Cagli, R., Coraggio, P., Napoletano, P., Schwartz, O., Ferraro, M., & Boccignone, G. (2009). Visuomotor characterization of eye movements in a drawing task. Vision Research, 49(8), 810–8.

Cohen, D. J. (2005). Look little, look often: The influence of gaze frequency on drawing accuracy. Perception & Psychophysics, 67(6), 997–1009.

Cohen, D. J., & Bennett, S. (1997). Why can’t most people draw what they see? Journal of Experimental Psychology: Human Perception and Performance, 23(3), 609–621.

Cohen, D. J., & Earls, H. (2010). Inverting an image does not improve drawing accuracy. Psychology of Aesthetics, Creativity, and the Arts, 4(3), 168–172.

Cohen, D. J., & Jones, H. E. (2008). How shape constancy relates to drawing accuracy. Psychology of Aesthetics, Creativity, and the Arts, 2(1), 8–19.

121

EXPERIMENTS | Encoding of structural information from individual fixations

Cowie, R., & Perrott, R. (1993). From line drawings to impressions of 3D objects: developing a model to account for the shapes that people see. Image and Vision Computing, 11(6), 342–352.

Crawford, J. D., Medendorp, W. P., & Marotta, J. J. (2004). Spatial transformations for eye- hand coordination. Journal of Neurophysiology, 92(1), 10–9.

Donnelly, N., Found, a, & Müller, H. J. (1999). Searching for impossible objects: processing form and attributes in early vision. Perception & Psychophysics, 61(4), 675–90.

Drake, J. E., & Winner, E. (2011). Realistic drawing talent in typical adults is associated with the same kind of local processing bias found in individuals with ASD. Journal of Autism and Developmental Disorders, 41(9), 1192–201.

Elo, A. E. (1978). The rating of chessplayers, past and present. London: Batsford.

Freud, E., Avidan, G., & Ganel, T. (2013). Holistic Processing of Impossible Objects: Evidence from Garner’s speeded-classification task. Vision Research.

Freud, E., Ganel, T., & Avidan, G. (2013). Representation of possible and impossible objects in the human visual cortex: evidence from fMRI adaptation. NeuroImage, 64, 685–92.

Germine, L., Nakayama, K., Duchaine, B. C., Chabris, C. F., Chatterjee, G., & Wilmer, J. B. (2012). Is the Web as good as the lab? Comparable performance from Web and lab in cognitive/perceptual experiments. Psychonomic Bulletin & Review, 19(5), 847–57.

Glazek, K. (2012). Visual and motor processing in visual artists: Implications for cognitive and neural mechanisms. Psychology of Aesthetics, Creativity, and the Arts, 6(2), 155– 167. 4

Gobet, F., & Simon, H. a. (1996). Templates in chess memory: a mechanism for recalling several boards. Cognitive Psychology, 31(1), 1–40.

Green, C. S., & Bavelier, D. (2007). Action-video-game experience alters the spatial resolution of vision. Psychological Science, 18(1), 88–94.

122

Greene, M. R., & Oliva, A. (2009). Recognition of natural scenes from global properties. Cognitive Psychology, 58(2), 137–176.

Greenwood, J. a, Bex, P. J., & Dakin, S. C. (2009). Positional averaging explains crowding with letter-like stimuli. Proceedings of the National Academy of Sciences of the United States of America, 106(31), 13130–5.

Hayes, S., & Milne, N. (2011). What’s wrong with this picture? an experiment in quantifying accuracy in 2D portrait drawing. Visual Communication, 10(2), 149–174.

He, S., Cavanagh, P., & Intriligator, J. (1996). Attentional resolution and the locus of visual awareness. Nature, 383(3), 334–337.

He, S., Cavanagh, P., & Intriligator, J. (1997). Attentional resolution. Trends in Cognitive Sciences, 1(3), 115–21.

Hochstein, S., & Ahissar, M. (2002). View from the Top: Hierarchies and Reverse Hierarchies in the Visual System. Neuron, 36(5), 791–804.

Kesten, H. (1958). Accelerated Stochastic Approximation. The Annals of Mathematical Statistics, 29(1), 41–59.

Kosslyn, S. M., Ball, T. M., & Reiser, B. J. (1978). Visual images preserve metric spatial information: evidence from studies of image scanning. Journal of Experimental Psychology. Human Perception and Performance, 4(1), 47–60.

Land, M. F. (2006). Eye movements and the control of actions in everyday life. Progress in Retinal and Eye Research, 25(3), 296–324.

Locher, P. (2010). How Does a Visual Artist Create an Artwork? In J. C. Kaufman & R. J. Sternberg (Eds.), The Cambridge Handbook of Creativity (pp. 131–144). Cambridge: Cambridge University Press.

Martelli, M., Majaj, N. J., & Pelli, D. G. (2005). Are faces processed like words? A diagnostic test for recognition by parts. Journal of Vision, 5, 58–70.

123

EXPERIMENTS | Encoding of structural information from individual fixations

Mitchell, P., Ropar, D., Ackroyd, K., & Rajendran, G. (2005). How Perception Impacts on Drawings. Journal of Experimental Psychology. Human Perception and Performance, 31(5), 996 –1003.

New, B., Pallier, C., Brysbaert, M., & Ferrand, L. (2004). Lexique 2 : A new French lexical database. Behavior Research Methods, Instruments, & Computers, 36(3), 516–524.

Pelli, D. G. (1997). The VideoToolbox software for visual psychophysics: transforming numbers into movies. Spatial Vision, 10(4), 437–442.

Perdreau, F., & Cavanagh, P. (2011). Do artists see their retinas? Frontiers in Human Neuroscience, 5(171), 1–10.

Perdreau, F., & Cavanagh, P. (2013a). Is Artists’ Perception more Veridical? Frontiers in Neuroscience, 7(6), 1–11.

Perdreau, F., & Cavanagh, P. (2013b). The artist’s advantage: Better integration of object information across eye movements. I-Perception, 4(6), 380–395.

Posner, M., Petersen, S., Fox, P., & Raichle, M. (1988). Localization of cognitive operations in the human brain. Science, 240(4859), 1627–1631.

Prins, N. (2012). The psychometric function: the lapse rate revisited. Journal of Vision, 12(6), 1–16.

Rayner, K. (1998). Eye movements in reading and information processing: 20 years of research. Psychological Bulletin, 124(3), 372–422.

124

Schacter, D. L., Cooper, L. A., & Delaney, S. M. (1990). Implicit memory for unfamiliar objects depends on access to structural descriptions. Journal of Experimental Psychology. General, 119, 5–24.

Seamon, J., & Carrasco, M. (1999). The effect of study time on priming possible and impossible figures in the object decision test. Psicothema, 11(4), 801–813.

Soldan, A., Hilton, H. J., & Stern, Y. (2009). Bias effects in the possible/impossible object decision test with matching objects. Memory & Cognition, 37(2), 685–692.

Tchalenko, J. (2007). Eye movements in drawing simple lines. Perception, 36(8), 1152– 1167.

Tchalenko, J. (2009). Segmentation and accuracy in copying and drawing: experts and beginners. Vision Research, 49(8), 791–800.

Tchalenko, J., & Miall, R. C. (2009). Eye–hand strategies in copying complex lines. Cortex, 45(3), 368–376.

Thorpe, S., Fize, D., & Marlot, C. (1996). Speed of processing in the human visual system. Nature, 381, 520–522.

Toet, A., & Levi, D. (1992). The two-dimensional shape of spatial interaction zones in the parafovea. Vision Research, 32(7), 1349–1357.

Treisman, A. M., & Gelade, G. (1980). A feature-integration theory of attention. Cognitive Psychology, 12(1), 97–136.

Whitney, D., & Levi, D. M. (2011). Visual crowding: a fundamental limit on conscious perception and object recognition. Trends in Cognitive Sciences, 15(4), 160–8.

Williams, P., & Tarr, M. J. (1997). Structural processing and implicit memory for possible and impossible figures. Journal of Experimental Psychology. Learning, Memory, and Cognition, 23(6), 1344–61.

125

EXPERIMENTS | Encoding of structural information from individual fixations

126

4.4 SPECIALIZED VISUAL MEMORY ENGAGED ONLY DURING THE DRAWING PROCESS

4.4.1 OBJECTIVES AND SUMMARY OF RESULTS A critical aspect of observational drawing is the correct positioning of the depicted features according to the original object’s structure. Consistently, our previous studies found that skilled people were more efficient in encoding object structures either across or within gaze fixations made on the object (Perdreau & Cavanagh, 2013b, 2014). This final study investigates whether some memory benefits seen in experts are limited to contexts where they are producing drawings.

Previous studies have investigated the role of visual memory in drawing and reported eye-hand strategies similar to those observed in visuomotor tasks involving a high load on visual working memory (Ballard et al., 1995): only local information relevant for the execution of the current action are fixated and hence encoded into memory (Cohen, 2005; Tchalenko, 2009). However, none of these studies directly tested visual memory during the drawing process. Moreover, if drawing experts develop optimized strategies to encode less in visual memory (rather than having larger memory capacity), then it remains to be explained how drawing experts could outperform novices in delayed recognition task involving complex figures or spatial organizations of features (Cohen & Jones, 2008; McManus et al., 2010; Rosenblatt & Winner, 1988). In particular, observational drawing requires comparing the copy to the original in order to find and correct any deviations. Because both the copy and the original are not easy to compare without eye movements and the necessary delay between fixations, this comparison will require memory. What particular information is encoded into visual working memory to enable this comparison? Is information relative to the original more accurately represented in memory than the copy in order to guide the production? Is the whole object’s structure stored or only information directly relevant for the production of the feature being drawn?

To address these issues, we designed an interactive pen tablet experiment coupled with a change detection task. In a first experiment, participants were asked to copy on a pen tablet a figure displayed on a screen. However, at random moments changes could occur in either the drawing or the original at various locations relative to the last drawn point. Participants were told to correct their drawing so that it continuously matches the original every time they noticed a change. However, the drawing and the original figure were displayed in alternation, so that participants could not compare them simultaneously. We varied the amount of change and we computed the critical amount of change that led to an appropriate correction in 50% of cases. We analyzed each location of change relative to the

127

EXPERIMENTS | Specialized visual memory engaged only during the drawing process last drawn point for changes occurring in the drawing and in the original. Finally, all of these results were analyzed as a function of participants’ drawing skill, which was assessed separately by a traditional drawing task. We found that all participants had better detection performance for changes closest to the segment currently being drawn and also better detection for changes occurring in the original than in the drawing. Interestingly, more skilled participants had an overall 40% advantage in detecting changes. In a second experiment, we determine that this advantage was specific to the drawing context and not to more general abilities in detecting changes. We designed a simple change detection using the same conditions as in the first experiment but without drawing involved. We found no significant advantage for more skilled participants.

Clearly, this study shows that all the feature positions are not equally represented in visual memory. First, spatial positions of the original’s features are much more accurately encoded into visual memory than those of the drawing itself. This suggests that an internal representation of the original could indeed guide the drawing process. Secondly, spatial positions related to the segment currently being reproduced are more accurately represented than those of previous segments. This is consistent with the idea that only information relevant to the ongoing actions are stored and maintained in visual working memory (Ballard et al., 1995, 1997). Thirdly, although our more skilled participants showed the same pattern of results as novices, they had a strong overall advantage in recalling the tested positions in both original and drawing. This suggests that drawing skill may rely on a better resolution of spatial positions in memory but not on a greater capacity of working memory in terms of number of items that can be stored. Critically, this advantage was no longer observed in our second experiment, which did not involve drawing. This result indicates a specialization of memory mechanisms for the context of drawing.

128

4.4.2 DRAWING EXPERTS HAVE BETTER VISUAL MEMORY — BUT ONLY

WHILE DRAWING

This chapter is based on: Perdreau, F., & Cavanagh, P. (in prep). Drawing experts have better visual memory – but only while drawing.

Abstract. An accurate drawing must respect the relative positions of the depicted object’s features, which define its structure. However, accurately representing these relative positions may require an internal representation in visual memory that is robust to the disruptions from the many eye-movements made between the object and the drawing. Moreover, it remains unclear whether such representation should include the entire object’s structure or only the local features relevant to the current drawing position. Therefore, we designed a pen tablet experiment coupled with a change detection task. A polygonal shape was displayed on a screen and participants had to copy it on an interactive pen tablet. At unpredictable moments during the copying process, modifications could be made on the drawing and the original figure. Participants had to correct their drawing every time they perceived a change so that their drawing always matched the current original figure, although the drawing and the original figure were presented in alternation. We found a strong relationship between drawing skill and the ability to detect and correct these changes. Interestingly, this effect of drawing skill was no longer present when our participants were asked to detect changes in task where drawing was not involved.

Keywords: drawing accuracy, visual memory, spatial position, change detection

4.4.2.1 INTRODUCTION Observational drawing is a visuomotor task in which visual information (an object, a photograph, or a figure) has to be translated into marks on the paper so that the copy ideally matches the original. As such, it is a good example of interaction between vision and action. In most of daily situations, these two systems cooperate quite well, enabling us acting on or avoiding objects in our environment. However, when it comes to copy those same objects, it turns out to be a very challenging task for the bold of us. At a first glance, this should be surprising as observational drawing may only require one to carefully depict the content of his or her own visual percept. Such difficulty may indicate that observational drawing is not a simple task and that extra practice is required to develop the required processes. Drawing experts, such as professional artists and draftspersons, are able to produce very convincing depictions of these same objects. Acquisition of skill is known to be related to the adaptation of existent mechanisms to the task constraints, and investigating expert performance may reveal processes that are unnoticed in novices (Ericsson & Lehmann, 1996a). For instance, expertise in action video games or chess has been related to long-lasting improvements of

129

EXPERIMENTS | Specialized visual memory engaged only during the drawing process sensory processes (Li et al., 2009), visual attention (Green & Bavelier, 2003, 2007), visual working memory (Chase & Simon, 1973; Gobet & Simon, 1996a) as well as changes in brain structures (Draganski & May, 2008). In the present study, we will show that drawing skill is relate to a specialization of visual memory mechanisms that enable skilled artists to build the internal representation required to guide the drawing process.

An accurate drawing must respect the relative spatial positions of the object’s features. However, the encoding of features’ spatial positions may be impaired by the many eye and hand movements made throughout the drawing process (Coen-Cagli et al., 2009; Crawford et al., 2004; Glazek, 2012; Land, 2006; Ogawa et al., 2010; Tchalenko & Miall, 2009; Tchalenko, 2007, 2009). One possibility is that all the features and their spatial relationships are stored in visual working memory. However, this would likely overload visual working memory (e.g. Alvarez & Cavanagh, 2004; Wheeler & Treisman, 2002). In two previous studies we hypothesized that features themselves would not be encoded to visual working memory but only their spatial organization – the object’s structure. Similar memory mechanisms have been observed in experts in chess who encode local spatial structures of chess pieces (perceptual chunks) that are integrated according to their spatial relationships into more complex chunks (memory chunks or templates) stored in long-term memory (Gobet & Simon, 1996a).

A similar chunking mechanism would help draftspersons during drawing. Consistent with this idea, Tchalenko and Miall (Tchalenko, 2009) found that drawing experts segment the original into complex sets of related lines (segment) that can be drawn in a single movement (De Winter & Wagemans, 2004; Van Sommers, 1984). In addition, drawing experts produce more strokes while making shorter fixations covering smaller spatial extents of the objects (Glazek, 2012), which suggest a more efficient encoding of visual information. In addition, we have previously found that drawing accuracy was related to the ability to better integrate object structural information across eye-movements (Perdreau & Cavanagh, 2013b) as well as to a more efficient, faster encoding of object structure during a single fixation (Perdreau & Cavanagh, 2014). Despite this clear advantage in processing structural information, it has not been investigated yet whether artists have any additional advantage in visual memory during an actual drawing task as compared to a non-drawing context.

The role of visual working memory in drawing has been previously discussed in several studies (Cohen, 2005; Glazek, 2012; McManus et al., 2010; Tchalenko & Miall, 2009; Tchalenko, 2009). First, a relationship between memory performances in delayed recognition tasks and drawing skill has been reported in both adults and children (Cohen & Jones, 2008; McManus et al., 2010; Rosenblatt & Winner, 1988). However, these studies did not address

130 how precisely visual information would be encoded to visual working memory during the drawing process. Visuomotor tasks involving high memory load are known to be associated with overt eye-movements strategies aiming at reducing the amount of information to be stored: observers more frequently update visual information by making regular eye- movements toward the source of information rather than relying on a memory representation in order to guide the ongoing action (Ballard et al., 1995). However, it seems unlikely that expertise in drawing is unrelated to visual working memory. In particular, drawing experts outperform novices in visual memory tasks using complex figures and spatial organization of features (Cohen & Jones, 2008; McManus et al., 2010; Rosenblatt & Winner, 1988). What remains to be determined is whether this memory advantage is accentuated in the context of actually making a drawing.

It is worthwhile to note that visual memory may have two different roles during a drawing task: guiding the production while the original is out of sight and comparing the original to the copy in order to detect mismatches. Studies investigating memory-guided reaching movements have claimed that visual memory plays no role in guiding action. Instead of relying on a visual memory representation of the target position, the motor system uses the trace of the visuomotor transformation (visuomotor memory, e.g. Hesse & Franz, 2009). The accuracy of this internal representation decays quickly after 2 seconds (Elliott & Calvert, 1990), which may explain why increasing the time interval during which the original information is not visible impacts on drawing accuracy, as it does for reaching movements, as well as why drawing experts would make more frequent eye-movements toward the source of information in order to constantly update this representation (Cohen, 2005). In contrast, such a representation could not be used in order to compare the depiction to the original. Visually comparing the original to the copy may depend on visual memory representation of either the original or the copy depending on which one is currently fixated. Although previous studies already discussed the implication of visual memory in drawing (Cohen, 2005; Glazek, 2012; McManus et al., 2010; Tchalenko & Miall, 2009; Tchalenko, 2009), none of these studies directly tested visual memory during the drawing process. It remains therefore unknown what particular information would be encoded to memory and how they would be encoded.

The aim of the present study is twofold: 1. identifying what information is stored in visual memory when encoding the original and the drawing separately during the drawing process and whether the efficiency of this memory process varies with drawing skill. It is particularly important to determine whether the representation stored in memory includes the whole object’s structure or only the features related to the being drawn segment, as well as whether the resolution of this memory representation of the drawing is as accurate as that of

131

EXPERIMENTS | Specialized visual memory engaged only during the drawing process the original. 2. If the efficiency of the memory process relates to drawing skill, is it specific to the context of producing a drawing or is it linked to a general advantage in visual memory? To examine these issues, we coupled a drawing task with a change detection task. Participants copied a simple shape (“original”) on an interactive pen tablet (“drawing”). They could only see one at a time, alternating between the original and the drawing. At various times, changes were made to either their drawing or the original figure at different possible locations more or less distant from the last drawn point. In addition to continuing the drawing, participants had to correct any deviations they noticed. All of these were analyzed as a function of the participant’s drawing skill, measured separately. In a second control experiment, we used a similar change detection task but with no active drawing involved to see whether any advantages found for skilled artists in the first experiment were specific to the drawing task.

According to the hypothesis that visual memory would contribute to drawing accuracy, we expected a positive relationship between participants’ drawing skill and their sensitivity to change in our tasks. Moreover, if the memory representation of features’ positions would be more suited to the requirements of motor planning, participants’ sensitivity to change should be higher for locations related to the tracing of the current segment than for other locations. Alternatively, participants’ sensitivity to change should be equal across all the tested locations if this representation would include the whole object’s structure.

4.4.2.2 GENERAL METHOD

4.4.2.2.1 PARTICIPANTS 22 adults participated in our experiment (average age: 27.7±1, 13 females). Although the present study did not aim at investigating artistic abilities (aestheticism or creativity), we recruited professional artists (n=3) in order to increase the range of drawing skills within our participants sample. All subjects were paid 20 euros for participating and were naïve about our hypotheses and about the purpose of the study. They all had normal or corrected-to- normal vision. Finally, they all gave their informed consent according to the University Paris Descartes Ethic committee. All participants except two performed in both experiments (these two participants did not reply to the announcement of the second experiment).

4.4.2.2.2 MATERIAL All the experiments presented in this study used the same apparatus. Original figures were presented on a 30-inch flat screen (Apple Cinema HD Display) with a resolution of 1920*1080 pixels and a refresh rate of 60Hz (main screen). Drawings were made on an interactive pen tablet (Wacom Cintiq 24HD) with a resolution of 1920*1080 pixels and a

132 refresh rate of 60Hz. Participant’s chin was held by a chinrest at a distance of 82 cm from the main screen’s center and of 50 cm from the pen tablet’s center. The chin rest’s height was held constant, so that the viewing angle was approximately orthogonal to both the main screen’s and the pen tablet’s surface.

4.4.2.2.3 ASSESSMENT OF DRAWING SKILL As we did in a previous study (Perdreau & Cavanagh, 2014), we assessed our participants’ drawing accuracy by asking them to perform a drawing task and then by ranking their drawings through a web-based experiment. Participants had 15 minutes to copy a grayscale picture of an inverted house as accurately as possible. Because of the many details in the picture and the limited time, we instructed our subjects to start by copying the house’s structure, then its details, and if they had enough time, to copy the house’s environment. This process made sure that all participants had at least drawn the house’s structure, which was used in our subsequent analysis.

Next, drawings were compared through an online experiment. Randomly selected pairs of drawings as well as the original picture were presented to independent observers [n=171, average age: 30.4 ± 0.9, 128 females] who had to choose which of the two drawings more precisely matched the original house. Each drawing pair has been compared a minimum 60 times by independent observers. Then, drawings were ranked using an ELO ranking system (Elo, 1978; Perdreau & Cavanagh, 2014. The web-application can be found on one of the authors' website: www.florianperdreau.fr). Higher ELO scores mean better drawing accuracy.

4.4.2.3 EXPERIMENT 1

4.4.2.3.1 STIMULI Target figures that our participants had to copy were randomly-generated polygonal shapes with 10 points each (Fig. 1). Each figure was initially designed to fall within a circle of 300 pixels of radius and was generated by randomly selecting points with a distance from the circle’s center greater than half of the radius and distant from the other points by at least 60 pixels. Moreover, every figure was designed to have a total area equal to 80% of that of the circle. A total of 72 figures were generated and used in this experiment.

133

EXPERIMENTS | Specialized visual memory engaged only during the drawing process

Figure 1. Example of change in a polygonal figure. A change was defined as a displacement of a single figure’s point. This displacement could be either inward or outward relative to the figure’s center. Despite the change was only local, we computed a percentage of change in the global structure as the percentage mean square error (%RMSE). This was our dependent variable in the experiment.

4.4.2.3.2 PROCEDURE All participants started the experiment with a practice drawing block in order to make them more familiar with the pen tablet. The task was to copy on the pen tablet the original figure displayed on the main screen as accurately as possible (i.e. respecting the shape, size and position). The drawing process was characterized by two alternating phases: an encoding and a drawing phase (Fig. 2). During the encoding phase, the original figure was shown for 1 second, while the drawing was blanked out from the pen tablet. In contrast, the original figure was not visible during the drawing phase and the drawing was displayed on the pen tablet until the participant drew a point [average drawing duration: 1514±69 msec]. These phases were repeated until the drawing was completed. Specifically, participants did not have to trace the line segments of the figure but only to mark the ending point, and they were only allowed to trace a point during each drawing phase. Each drawing trial started with the display of the target figure on the main screen with one of its points colored in red (chosen randomly on each trial), indicating the first point to be drawn, and an adjacent point colored in green indicating the next point to be drawn. On the next presentations of the original figure, the to-be-drawn point was colored in green. Each participant completed 12 drawing trials in this block and we measured the accuracy of the participant’s drawings.

Next, each participant ran in a practice block and then in six test blocks (12 drawings in each block), which used the same procedure. Although the task was identical to that of the

134 practice drawing block, participants were instructed that several changes could be made on the original figure and on their own drawing during all the copying process. Participants had to correct their drawing every time they observed a change in either the drawing or the original, so that their drawing always matched the original shape. Particularly, they were allowed to only correct a point at a time by clicking on its position with the pencil’s gum to erase it and then by drawing it again at the whished location. Changes consisted in the displacement of a figure’s point relative to the figure’s center, which could be either inward or outward (Fig. 1). Although the change was only local, it affected the whole figure’s shape so that we computed a global percentage of change resulting from this local modification (%RMSE). We varied this percentage of change from 0% (no change) to 4%. Finally, changes could occur at different locations relative to the last drawn point (-1, -2 or -4).

Finally, to encourage participants to focus on the drawing task, a feedback was presented at the end of each drawing trial (see Fig. 3). Feedback consisted in an horizontal bar with a gray line centered on it and indicating the participant’s average drawing accuracy computed during the drawing block, and with a blue line indicating the current drawing’s accuracy. Participants were told that the blue line should never be in the horizontal bar’s red area (i.e. the current drawing accuracy should never exceed the average drawing accuracy by more than two standard deviations). To reinforce this feedback, a smiley was displayed above the horizontal bar, which the emotion corresponded to this result (“happy” if the blue line was within the green area, “angry” if it was within the red area).

Therefore, our experimental design included three within-subject factors: the place of change (original, drawing), the location of change relative to the last drawn point (-1, -2, -4) and the percentage of change (6 intensities from 0% to 4% of global change); and participants’ drawing accuracy as between-subject factors. There were 10 trials in each condition and each participant ran in a total of 300 trials.

135

EXPERIMENTS | Specialized visual memory engaged only during the drawing process

Figure 2. A. Experimental procedure. Throughout the drawing process, the original and the drawing were displayed in alternation. While the original was displayed on the main screen (duration 1 second), the participant’s drawing was blanked out, and conversely. The participant was only allowed to draw a single point when the original was not presented. Once the participant had drawn a point, the drawing was removed from the pen tablet, and the original was then presented on the main screen. Each trial started with the display of the original on the main screen. A red point on the original figure indicated the first point to draw, and a green point indicated the next to-be-drawn point. B. Location of change. In addition to varying the amount of change (size of displacement relative to the overall figure’s size), we manipulated the location of change relative to the last drawn point in both the original and the drawing C. Feedback. A feedback was shown to participants at the end of each drawing, indicating their drawing’s accuracy relative to their average drawing accuracy (computed over all the drawing practice trials). Participants were explicitly told that their current accuracy should not fall within the red area (i.e. exceeding their average accuracy by more than 2 SD).

4.4.2.3.3 RESULTS First, in order to measure our participants’ performances, correct responses were defined as every time participants corrected a point that had actually just changed. Next, for each location of change relative to the last drawn point, we fitted cumulative Gaussian function to participant’s proportion of correct responses and we computed participant’s

136 proportion of correct response as a function of the percentage of change (Fig. 3), and we measured the threshold percentage of change needed so that the participant perceive a change in 50% of the cases. However, to make our results easier to read, we took the inverse of the measured threshold as a measure of sensitivity to change.

For each location of change relative to the last drawn point, we fitted cumulative Gaussian function to participant’s proportion of correct responses (Fig. 3), and we measured the threshold percentage of change needed so that the participant perceived a change 50% of the time. We computed participant’s sensitivity to change as the inverse of this threshold.

Next, in order to measure whether participants’ sensitivity to change varied according to the place of the change (“Drawing” vs. “Original”), to its location relative to the last drawn point (-1,-2,-4) or to participants’ drawing skill, we computed a linear mixed-effects model on participants’ log-sensitivity with all of these fixed factors, and with subjects as random factors to account for our repeated measures design. We found that overall, our participants’ sensitivity to change was higher for changes occurring in the original figure than in their own drawing [χ²(1)=275.4, p<.0001], as well as for changes occurring on locations closer to the last drawn point [χ²(2)=155.1, p<.0001]. As expected by our hypothesis, skilled participants were overall better at detecting changes regardless of their place (in the drawing or in the original) or of their location relative to the last drawn point [χ²(1)=22.8, p<.0001]. We found a significant interaction between the place of the change and its location [χ²(2)=14.2, p<.0008], as between drawing accuracy and the location of change [χ²(2)=12.3, p<.002], and a marginally significant interaction between drawing accuracy and the place of change [χ²(1)=2.94, p=.086]. However, these interactions should be considered as artefacts of the non-linearity of the decay of sensitivity as a function of the location of the changed point relative to the last drawn points. These interactions were no longer significant when participants’ sensitivity was log-scaled [Ps>.23], whereas the main effects previously described remained highly significant (Ps << .0001).

137

EXPERIMENTS | Specialized visual memory engaged only during the drawing process

Figure 3. Skilled participants detect more changes in drawing task. Left Panel. For each participant, we measured the threshold percentage of change leading to 50% of correct responses, for both place of change (Drawing and Original) and for every location of change relative to current drawing location. Next we computed the inverse of the threshold as sensitivity to change. Middle panel. We computed a linear mixed-effects model with participants drawing accuracy and the location and place of change as fixed factors. Sensitivity was higher for changes occurring in the original figure than in the drawing, and more skilled participants were overall more sensitive to changes regardless of these conditions. Right panel. Summary graph were participants have been divided in two levels of drawing skill relatively to the median drawing accuracy. Error bars represent the 95% confidence intervals of the means.

The presence of a decrease of sensitivity to change in all our participants as the distance between the changed point and the last drawn point increases may suggest that a complete representation of the object structure in memory does not contribute to drawing accuracy. One could argue that this effect only reflects a decay of memory performance with time. It should be noticed, though, that all the figure’s points were always simultaneously presented, so that this result may indicate an effect of the time of use rather than an effect of the delay from previous presentations. Only positions relevant for estimating the to-be-drawn feature’s relative position may be attended and encoded to visual memory, which may result in higher performances of change detection for those points (Rensink, O’Regan, & Clark, 1997; Simons & Rensink, 2005).

Interestingly, more skilled participants did show these same effects, although they were more sensitive than novices to changes occurring at these locations by about 40%.

138

This may suggest that drawing accuracy does not go with holding more features’ positions in memory but rather with a better resolution of their representation in memory (Zhang & Luck, 2008). To test that possible explanation, we measured the spatial errors made by our participants when they had to correct a point that had changed in their own drawing by repositioning the target point at the correct location. We then computed the correlation between log-transformed spatial errors and drawing accuracy. We found a significant relationship between participants’ average spatial errors and their drawing accuracy [r(19)=- 0.52, p<0.02, R²=.27] suggesting that more skilled participants made smaller spatial errors. This result would support the hypothesis of a better resolution of visual memory for spatial positions in skilled participants.

Figure 4. Most of the line segment is drawn blind. We computed the distance covered by the pencil on the pen tablet during the presentation of the original (blue) and the drawing (orange). In average, the pencil covered about 75% of the total length of the drawn line-segment while the original figure was displayed on the main screen and the drawing not visible. This hand movement pattern did not differ with drawing skill. Each point represents a participant and error bars show 95% confidence interval of the mean.

Better memory for relevant positions may also explain why our participants showed a dramatic drop of sensitivity for changes occurring in their production. To be suited to motor planning, the segment’s ending position has to be vectored and transformed in terms of amplitude and direction that are necessary to compute the kinematic attributes of the planned hand movement (Bock, Dose, Ott, & Eckmiller, 1990; Favilla, Hening, & Ghez, 1989; Kalaska & Crammond, 1992). This may require encoding the ending point spatial position relative to a particular reference frame, computing for example its distance and subtended

139

EXPERIMENTS | Specialized visual memory engaged only during the drawing process angle from other surrounding features (Valois & Lakshminarayanan, 1990). However, once this computation is done and the hand movement programmed, the only information that may be processed when looking at the drawing is that required for visual control of the hand trajectory (e.g. distance between hand position and target position; Tchalenko & Miall, 2009). Hence, we analyzed participant’s pencil movement in order to determine whether the overall drop of sensitivity we found for changes occurring in the drawing against those present in the original could be explained by a particular eye-hand strategy as described above (Fig. 4). Our data show that about 75% of the target segment was drawn [F(1,41)=31.57, p<0.0001] when the original figure was displayed than when the drawing was visible. These results are consistent with the “just-in-time” strategy previously reported (Tchalenko & Miall, 2009; Tchalenko, 2009).

Altogether, our results provide evidence of a relationship between drawing skill and a better memory resolution for features’ spatial positions. More skilled participants were not only overall better at identifying which point has changed, but they were also more accurate at remembering its previous position. Despite this advantage, more skilled participants did not differ from novices in the way of encoding visual information during the drawing progress.

4.4.2.4 EXPERIMENT 2: CHANGE DETECTION TASK The aim of the present experiment was twofold: 1. to examine whether the memory advantage we found for participants skilled in drawing would be specific to the constraints of the drawing task or related to a more general ability in detecting changes; and 2. to test whether the difference in sensitivity to change we observed between the Drawing and Original conditions in the first experiment could also be explained by the different availability of visual cues that might have helped our participants to detect the presence of a change. Particularly, changes are displacements of a point relative to the figure’s centroid, which result in a variation of the figure’s area. This cue could be also used to detect the presence of a change. However, computing the area might be more challenging for uncompleted figures (drawing in progress) than for completed figures (original). We therefore used both complete and incomplete figures in a change detection task not involving drawing.

4.4.2.4.1 PARTICIPANTS Participants were identical to those who participated in the first experiment. However, two of our previous participants were not able to participate during the test period.

4.4.2.4.2 STIMULI A total of 144 figures were generated like those used in the first experiment.

140

4.4.2.4.3 PROCEDURE

Figure 5. Experiment 2’s procedure. A probe stimulus (polygonal figure) was presented to the participant for 1000 ms, then disappeared for an inter-stimulus interval of 1000 ms and finally reappeared with a possible change or not. Participants had to answer, by clicking on the appropriate response box, whether a change has occurred or not. In order to match the Original vs. Drawing condition of the first experiment, we tested both complete and incomplete figures in this experiment.

The present experiment was divided in two blocks of trials that differed in that they either use complete or partial version of the figures. Both blocks used the same procedure.

Each trial started with a fixation cross centered on the screen, then replaced by the probe figure for 1000 ms. The probe figure was then immediately replaced by a blank screen for approximately 1000 ms. After this inter-stimulus interval, a target figure was displayed until the participant gives his or her response. Participants were asked whether a change occurred in the target figure or not comparatively to the probe figure. They had to give their response by clicking on the appropriate response box (“Present” if there was a change, or “Absent” in the opposite case). Response boxes were located on the side of the participant’s handedness. As in the first experiment, we varied the percentage of change induced by the displacement of one of the figure’s points from 0% to 4% of change.

We therefore had 2 within-subject factors: type of figure (drawing vs original), percentage of change (6 intensities), with a total of 12 trials per condition and a total amount of trials of 144 trials.

141

EXPERIMENTS | Specialized visual memory engaged only during the drawing process

4.4.2.4.4 RESULTS As in the first experiment, we fitted cumulative Gaussian function to participants’ proportion of correct responses plotted as a function of the percentage of change (Fig. 6). Next, for each participant and for each type of figure, we measured the sensitivity to change as the inverse of the threshold percentage of change leading to 75% of correct responses (2- AFC).

We computed a linear mixed-effects model with the Type of figure as fixed factors, with participants’ measured sensitivity as dependent variable, and participants as random factor. We found no significant effect of either the type of figure or of participants’ drawing accuracy on their measured sensitivity [Type of figure: χ²(1)=2.21, p=.14; Accuracy: χ²(1)=0.005, p=.94], and no significant interaction between these two factors [χ²(1)=1.05, p=.30].

Figure 6. No effect of drawing skill on change detection performances when drawing is not involved. Left Panel. For each participant, we measured the threshold percentage of change leading to 75% of correct responses in the change detection task and we computed the inverse of the threshold as a measure of participants’ sensitivity to change. We found no significant relationships between our participants’ sensitivity and their accuracy in drawing. Right Panel. Interaction plot showing participants’ sensitivity to change as a function of drawing skill and the type of figure used in the task (Drawing vs. Original). Errors bars are the bootstrapped 95% intervals of the means.

Taken together, these results suggest that drawing accuracy is related to better memory of spatial positions only when drawing is involved. Moreover, the present findings also rule out the possibility that the difference of detection performance found between the Original vs. Drawing conditions in the first experiment could be attributed to the presence of absence of visual cues that could have helped or prevent our participants to correctly detect changes during the drawing process.

142

4.4.2.5 GENERAL DISCUSSION What is the information actually encoded into visual memory during a drawing task? To what extent the visual memory process would be more efficient in skilled artists? Although the role of visual memory in drawing has been discussed in previous studies (Cohen, 2005; Glazek, 2012; McManus et al., 2010; Perdreau & Cavanagh, 2013b, 2014; Tchalenko & Miall, 2009; Tchalenko, 2009), it has never been directly tested during the drawing process itself yet. Here, we designed an interactive pen tablet experiment coupled with a change detection task that allowed us to measure visual memory performances and efficiency of participants while they were currently copying a figure. In a first experiment, participants of various drawing skill levels had to copy an original figure on a pen tablet and were instructed that several modifications could be brought to both the original figure and their own production. They had to correct their drawing every time they noticed a change so that it always matched the current version of the original, although the original and the drawing were never presented at the same time. We particularly varied the amount of overall change resulting from the displacement of a point as well as the location of the change relative to the last drawn point. In a second experiment, we asked whether better performances we expected to observe in participants more skilled in drawing were specific to the drawing context, or whether they could be attributed to more general aptitudes. Participants took part in a simple change detection tasks where drawing was not involved and with similar conditions regarding the type of stimuli we used in the first experiment.

In the first experiment, we found that participants were overall better at detecting a change when it occurred in the original figure than in their own drawing and that their sensitivity to change decreased as a function of the location of the change relative to the last drawn point. Because change detection performance indicates what information have been attended during the initial presentation (Rensink et al., 1997; Simons & Rensink, 2005), these findings may suggest that the processing of feature’s position is already optimized during the drawing process, regardless of drawing skill, as only features’ positions related to the encoding and the production of the to-be-drawn feature are attended and stored in visual working memory. Moreover, the difference of sensitivity observed for changes present in the original picture vs. in the drawing may show that encoding the to-be-drawn segment’s position needs more information than its positioning within the depiction. An explanation could be that the preparation of the production of the stroke needs the computation of both the amplitude and the direction of the subsequent hand movement (e.g. Bock et al., 1990; Favilla et al., 1989; Kalaska & Crammond, 1992). Particularly, computing the direction of movement requires encoding the angle between the starting and the ending position, which might be estimated within a particular reference frame defined by other surrounding features.

143

EXPERIMENTS | Specialized visual memory engaged only during the drawing process

Once these kinematic attributes are estimated, visual information used when looking at the drawing may only help the visual guidance of the pencil movement to reach the target position. In this line, our hand movement data showed that all our participants started to move the pencil toward the ending point when only the original figure was visible. This finding is consistent with previous studies providing evidence of a “just-in-time” eye-hand strategy in both novices and experts in drawing (Tchalenko & Miall, 2009; Tchalenko, 2009). However, the implication of visual working memory in the guidance of reaching movement has been disputed (Hesse & Franz, 2009; Milner & Goodale, 1998). In particular, monitoring the correct trajectory and positioning of the hand during the execution would only rely on a trace of the motor program (visuomotor transformation; Hesse & Franz, 2009). Our results clearly show that information used to program the movement is temporarily held in visual working memory. However, our study does not rule out this hypothesis as we are not able to address whether this information would be actually used to guide the ongoing action.

Finally, we found a strong relationship between participants drawing accuracy and their sensitivity to change, although this relationship did not interact with the previously mentioned effects. Interestingly, we did not observe a similar memory advantage for participants more skilled in drawing when drawing was not involved (Experiment 2). This suggests that more skilled participants may have a more accurate representation of both the original figure and their own drawing and that it cannot be accounted by a higher sensitivity to change in general (Exp. 2). However, the absence of an interaction between the effect of drawing skill and those of the location and the place of change (drawing vs. original) may tell that drawing accuracy does not depend on encoding more information in visual memory but rather on a more accurate representation of this stored information in memory (Zhang & Luck, 2008). In support of this hypothesis, we found that more skilled participants made fewer spatial errors when they had to reposition the changed point to its correct location.

What would explain the higher accuracy of more skilled participants and its absence in our second experiments, where drawing is not involved? The simplest explanation is that more skilled participants might allocate their attention more efficiently to locations in the original and the drawing that are critical for the task. This sharper focus of attention on fewer, more relevant locations may increase the resolution of stored visual representations and decrease the amount of irrelevant information (Awh, Jonides, & Reuter-Lorenz, 1998; Zhang & Luck, 2008). An alternative hypothesis would be that more skilled participants could call on a specialized form of visual memory while creating the drawing (e.g. Perdreau & Cavanagh, 2013b). We cannot discriminate between these two alternatives with our data here and further research will be required to disentangle them.

144

4.4.2.6 CONCLUSIONS This study aimed at investigating the role of visual working memory in the drawing process as well as its contribution to drawing skill. Although this issue has been discussed in previous studies (Cohen, 2005; Glazek, 2012; McManus et al., 2010; Rosenblatt & Winner, 1988; Tchalenko, 2009), the present study is the first to measure memory performances during an actual drawing task. Using an interactive pen tablet experiment coupled with a change detection task, we showed that all the feature positions are not equally represented in visual memory. First, spatial positions in the original are more accurately encoded to visual memory than those of the drawing itself, suggesting that an internal representation of the original would be adequate to guide the production. Spatial positions related to the segment currently being reproduced are more accurately represented than those of previous segments, which is consistent with the idea that only information relevant to the current hand movement are stored and maintained in visual working memory (Ballard et al., 1995, 1997). Finally, more skilled participants in drawing had a strong overall advantage in detecting changes in both the original and the drawing. This suggests that drawing skill may rely on a better resolution of spatial positions in memory (Zhang & Luck, 2008). Critically, this advantage was only present as long as active drawing was involved, which indicates a specialization of memory and / or attention mechanisms for the context of drawing.

4.4.2.7 ACKNOWLEDGMENTS This research was supported by an ANR grant to P.C. and a French Ministère de l’Enseignement Supérieur et de la Recherche grant to F.P.

4.4.2.8 REFERENCES

Alvarez, G. A., & Cavanagh, P. (2004). The capacity of visual short-term memory is set both by visual information load and by number of objects. Psychological Science, 15(2), 106– 11.

Baddeley, A., & Hitch, G. (1974). Working Memory. Psychology of Learning and Motivation, 8, 47–89.

Ballard, D., Hayhoe, M., & Pelz, J. B. (1995). Memory representations in natural tasks. Journal of Cognitive Neuroscience, 7(1), 66–80.

Bock, O., Dose, M., Ott, D., & Eckmiller, R. (1990). Control of arm movements in a 2- dimensional pointing task. Behavioural Brain Research, 40(3), 247–250.

145

EXPERIMENTS | Specialized visual memory engaged only during the drawing process

Chase, W. G., & Simon, H. A. (1973). Perception in chess. Cognitive Psychology, 4(1), 55– 81.

Coen-Cagli, R., Coraggio, P., Napoletano, P., Schwartz, O., Ferraro, M., & Boccignone, G. (2009). Visuomotor characterization of eye movements in a drawing task. Vision Research, 49(8), 810–8.

Cohen, D. J. (2005). Look little, look often: The influence of gaze frequency on drawing accuracy. Perception & Psychophysics, 67(6), 997–1009.

Cohen, D. J., & Jones, H. E. (2008). How shape constancy relates to drawing accuracy. Psychology of Aesthetics, Creativity, and the Arts, 2(1), 8–19.

Crawford, J. D., Medendorp, W. P., & Marotta, J. J. (2004). Spatial transformations for eye- hand coordination. Journal of Neurophysiology, 92(1), 10–9.

De Vito, S., Buonocore, A., Bonnefon, J.-F., & Della Sala, S. (2014). Eye movements disrupt spatial but not visual mental imagery. Cognitive Processing, 1(1), 1–7.

Draganski, B., & May, A. (2008). Training-induced structural changes in the adult human brain. Behavioural Brain Research, 192, 137–142.

Elliott, D., & Calvert, R. (1990). The influence of uncertainty and premovement visual information on manual aiming. Canadian Journal of Psychology, 44(4), 501–511.

Elo, A. E. (1978). The rating of chessplayers, past and present. London: Batsford.

Ericsson, K. a, & Lehmann, a C. (1996). Expert and exceptional performance: evidence of maximal adaptation to task constraints. Annual Review of Psychology, 47, 273–305.

Favilla, M., Hening, W., & Ghez, C. (1989). Trajectory control in targeted force impulses. Experimental Brain Research, 75(2).

Glazek, K. (2012). Visual and motor processing in visual artists: Implications for cognitive and neural mechanisms. Psychology of Aesthetics, Creativity, and the Arts, 6(2), 155– 167.

Gobet, F., & Simon, H. a. (1996). Templates in chess memory: a mechanism for recalling several boards. Cognitive Psychology, 31(1), 1–40.

146

Green, C. S., & Bavelier, D. (2003). Action video game modifies visual selective attention. Nature, 423(6939), 534–7.

Green, C. S., & Bavelier, D. (2007). Action-video-game experience alters the spatial resolution of vision. Psychological Science, 18(1), 88–94.

Hesse, C., & Franz, V. H. (2009). Memory mechanisms in grasping. Neuropsychologia, 47, 1532–1545.

Huette, S., Kello, C. T., Rhodes, T., & Spivey, M. J. (2013). Drawing from memory: hand-eye coordination at multiple scales. PloS One, 8(3), e58464.

Johansson, R., & Johansson, M. (2014). Look here, eye movements play a functional role in memory retrieval. Psychological Science, 25(October 2013), 236–42.

Kalaska, J. F., & Crammond, D. J. (1992). Cerebral cortical mechanisms of reaching movements. Science, 255(5051), 1517–23.

Land, M. F. (2006). Eye movements and the control of actions in everyday life. Progress in Retinal and Eye Research, 25(3), 296–324.

Lawrence, B. M., Myerson, J., Oonk, H. M., & Abrams, R. a. (2001). The effects of eye and limb movements on working memory. Memory (Hove, England), 9, 433–444.

Li, R., Polat, U., Makous, W., & Bavelier, D. (2009). Enhancing the contrast sensitivity function through action video game training. Nature Neuroscience, 12(5), 549–551.

Logie, R. H. R., & Marchetti, C. (1991). Visuo-spatial working memory: Visual, spatial or central executive?. Mental Images in Human Cognition, 1991.

Milner, A. D., & Goodale, M. A. (1998). The Visual Brain in Action. Brain, (27).

Ogawa, K., Nagai, C., & Inui, T. (2010). Brain mechanisms of visuomotor transformation based on deficits in tracing and copying. Japanese Psychological Research, 52(2), 91– 106.

147

EXPERIMENTS | Specialized visual memory engaged only during the drawing process

Perdreau, F., & Cavanagh, P. (2013). The artist’s advantage: Better integration of object information across eye movements. I-Perception, 4(6), 380–395.

Perdreau, F., & Cavanagh, P. (2014). Drawing skill is related to the efficiency of encoding object structure. I-Perception, 5(2), 101–119.

Postma, A., & De Haan, E. H. F. (1996). What was where? Memory for object locations. The Quarterly Journal of Experimental Psychology, 49A(1), 178–199.

Rensink, R. A., O’Regan, J. K., & Clark, J. J. (1997). To see or not to see: The need for attention to perceive changes in scenes. Psychological Science, 8(5), 1–6.

Rosenblatt, E., & Winner, E. (1988). The art of children’s drawing. Journal of Aesthetic Education, 22(1), 3–15.

Salway, A. F., & Logie, R. H. (1995). Visuospatial working memory, movement control and executive demands. British Journal of Psychology (London, England : 1953).

Simons, D. J., & Rensink, R. A. (2005). Change blindness: past, present, and future. Trends in Cognitive Sciences, 9(1), 16–20.

Smith, E., Jonides, J., & Koeppe, R. (1996). Dissociating verbal and spatial working memory using PET. Cerebral Cortex.

Tchalenko, J. (2007). Eye movements in drawing simple lines. Perception, 36(8), 1152– 1167.

Tchalenko, J. (2009). Segmentation and accuracy in copying and drawing: experts and beginners. Vision Research, 49(8), 791–800.

Tchalenko, J., & Miall, R. C. (2009). Eye–hand strategies in copying complex lines. Cortex, 45(3), 368–376.

Valois, K. De, & Lakshminarayanan, V. (1990). Discrimination of relative spatial position. Vision Research, 30(11), 1649–1660.

Van Sommers, P. (1984). Drawing and cognition: Descriptive and experimental studies of graphic production processes. New York: Cambridge Univ Press.

148

Wheeler, M. E., & Treisman, A. (2002). Binding in short-term visual memory. Journal of Experimental Psychology: General, 131(1), 48–64.

Winter, J. De, & Wagemans, J. (2004). Contour-based object identification and segmentation: Stimuli, norms and data, and software tools. Behavior Research Methods, Instruments, & …, 36(4), 604–624.

Zhang, W., & Luck, S. J. (2008). Discrete fixed-resolution representations in visual working memory. Nature, 453(7192), 233–5.

149

EXPERIMENTS | Specialized visual memory engaged only during the drawing process

150

151

EXPERIMENTS | Specialized visual memory engaged only during the drawing process

152

5 GENERAL DISCUSSION

153

GENERAL discussion | Specialized visual memory engaged only during the drawing process

154

In this last chapter, we will briefly review the main issues discussed in this thesis and will summarize the main results related to the corresponding studies. Finally, we will suggest future research to address issues rose by our studies.

5.1 MAIN ISSUES OF THE THESIS

The main purpose of this thesis was to investigate the brain mechanisms involved in observational drawing. Drawing is a very challenging task for most of us, and intensive practice is often required to reach a good accuracy. We therefore compared skilled artists vs. novices in order to determine how and to what extent already existent perceptual, memory and motor mechanisms, initially not tuned to the purpose of drawing, could develop according to the requirements of the drawing task.

We first examined whether drawing skill was related to the ability to access visual representations of objects unaffected by perceptual constancy mechanisms that ordinarily correct for changes in viewing conditions, which may prevent us to depict objects as they are perceived. Although we found no evidence for a relationship between this ability and drawing skill, previous studies did. Using a visual search task implying amodal completion, we were able to demonstrate that drawing skill may not rely on the ability to access earlier, uncorrected visual representations, but rather that it may depend on post-perceptual corrections and on a particular visual analysis of scenes and objects’ layout.

A critical aspect of drawing accuracy is the correct positioning of features according to the global object’s spatial organization, which may require a good representation of the object’s structure. However, such representation might need to be built from and across each fixation made on the object while scanning it or comparing it to the copy. In two other studies, respectively using a gaze-continent window task, a visual masking task and a categorization task in periphery, we were able to show that drawing skill was related to a better sequential integration and a more efficient encoding of structural information within and across fixations. Taking together, these studies suggested that drawing skill required a better encoding of object structure regardless of viewing conditions.

Better encoding of structural organization is also thought to be at the basis of expert chess players’ memory performances. In a final study, we demonstrated that this mechanism was most evident in the context of drawing production, and that an accurate visual memory representation of both the drawing and the copy stored in long-term memory would provide an advantage in guiding the drawing process.

155

GENERAL discussion | Discussion of the main results

5.2 DISCUSSION OF THE MAIN RESULTS

5.2.1 IS AN ARTIST’S PERCEPTION MORE VERIDICAL? In the first study of this thesis we investigated whether drawing skill could be accounted for by the involvement of a particular perceptual, more veridical mode – an “innocent-eye” (Ruskin, 1912). We started with the idea that observational drawing requires an extra focus on the particular viewing conditions under which the original object is perceived. Such a perceptual mode may require extra practice as it would not be in frequent use for our visual system which focuses more on recognizing objects regardless of viewpoints, distances or lighting conditions. We therefore tested professional artists, art students and novices in perceptual matching tasks involving size and brightness constancies, as well as in a visual search task within the context of amodal completion. We found no advantages (no reduction in the effect of visual constancies) for artists and art students over novices in those tasks. The absence of differences between artists and non-artists in our first experiment (size constancy) could be attributed to the fact that our stimuli did not trigger a strong effect of size constancy (about 10%). For example, a later study by Ostrofsky and colleagues (Ostrofsky et al., 2012) also designed a size constancy size effect of about 20% and found a 3% smaller effect in participants more skilled in drawing. In this line, our brightness constancy task triggered an effect of about 30% in our participants, and we found a numerical, but non-significant 10% advantage for art students and professional artists. Although this could be indicative of a more veridical perception in artists, this result should be taken carefully as several failures in replicating the artist’s advantage in visual constancies tasks (shape constancy) have been reported (McManus et al., 2011; Ostrofsky et al., 2012). Even if there is a small (non-significant here) perceptual benefit for artists, our findings call into question their ability to directly access earlier, uncorrected visual representations. First, in the two first experiments, art students and professional artists took about twice as long to make their perceptual judgments, suggesting that the tasks were not easier for them. Secondly, artists were not faster at finding the target within the context of amodal completion in the third experiment.

Taken together, our results strongly suggest that artists do not have access to an uncorrected retinal image but instead apply post-perceptual corrections on the basis of their explicit knowledge about visual effects to guess the proximal property of the stimulus. For instance, some of professional artists and art students we tested reported knowing the visual illusions we used, as they are taught in art schools: a square projected in perspective must be represented as a trapezoid, and surfaces lying within a cast shadow appears brighter than they actually are. One might therefore set matching values on the basis of such explicit

156 knowledge, decreasing the size or the luminance of the test stimulus accordingly. The use of explicit knowledge, rather than more veridical perception, could also account for discrepancies between results of studies investigating the relationship between drawing skill and perceptual constancies: more skilled participants may or not know these effects depending on whether they had a formal training in art school or were self-taught. The relative contribution of explicit knowledge and perceptual abilities in these tasks could be investigated by instructing both novices and drawing experts about these visual effects and examining whether an advantage for more skilled participants would still remain.

To conclude, observational drawing requires one to focus on the particular viewing conditions under which the original is perceived. In particular, this may require an extra focus on proximal properties of the original, which both novices and professional artists are capable of (Carlson, 1966; Rock, 1983). Observational drawing does not require accessing a greater “innocence of the eye” (Ruskin, 1912), which would only be the privilege of trained artists. In contrast, artists act as “copyists”, making progressive sketches and continuously making corrections until their perception of the depiction matches their perception of the original (Gombrich, 1960).

5.2.2 DRAWING SKILL AS PERCEPTUAL EXPERTISE? A main assumption of the “innocent eye” hypothesis is that artists could access a raw percept, unaffected by perceptual inferences and object knowledge. This requires that low- level representations of the scene could be accessed by our conscious perception. However, this view is not consistent with the functional organization of the visual system.

Visual information arriving on the retina is further processed throughout the hierarchy of the visual system: from the detection and processing of basic units (orientation, colors, positions, motion, etc.) to their integration into higher-level descriptions, such as objects, categories and concepts. Despite this bottom-up stream of visual processing, our conscious perception is at the top of this implicit hierarchy (Ahissar & Hochstein, 2004). Starting with coarser representations (the gist of the scene) and not with details allows faster recognition of objects with which we must interact. However, the way the content of this final percept will be subsequently analyzed depends on the requirements of the task at hand. For instance, if one’s task is to prepare a jam-butter sandwich, he or she will not attend and use the same objects (e.g. butter, bread, jam, and knife) as if he or she were to wash the dishes (e.g. Ballard, Hayhoe, & Pelz, 1995). This requires knowing what the object looks like and selecting this identified object while ignoring irrelevant ones. Therefore, visual perception is driven by both memory and attention. This also holds for expertise: radiologists are more efficient than novices at detecting abnormalities in a x-ray film displayed for 200 msec

157

GENERAL discussion | Discussion of the main results

(Myles-Worsley, Johnston, & Simons, 1988). Although the information arriving on both experts and novices’ retinas is the same and follows the same path through the stages of the hierarchical process, experts know better what the differently contrasted regions of the film “mean” (e.g. a bone, a tissue, a chamber, an organ) and can better discriminate normal patterns from abnormal ones. For these reasons, it has been suggested that “what typically limits naïve performance is the accessibility of task-relevant information rather than the absence of such information within neuronal representations” (Ahissar & Hochstein, 2004, p.457). What would be the task-relevant information in observational drawing?

Because observational drawing requires one to accurately and carefully depict his or her visual percept, it could be argued that any encoded visual information is therefore relevant. In that sense, expertise in observational drawing may be an exceptional case of perceptual expertise, which is by definition a specialization of processes relative to a particular domain of vision (e.g. object categories, orientation, etc.). However, the medium (pencil, paper, brush, etc.) necessarily restricts the amount and the type of information that can be reproduced. Hence an optimal selection of the most relevant information is required (Graham & Meng, 2011): understanding what visual information is relevant or ignored by the visual system (Cavanagh, 2005) and what is used in order to trigger a good recognition (Biederman & Kim, 2008; Biederman, 1987; Kozbelt et al., 2010). However, although these abilities and the resulting knowledge may characterize drawing experts and undoubtedly contribute to their production quality, we argued in this thesis that it may not be the most fundamental skill required to reach an expert level of drawing.

5.2.3 THE CHUNKING HYPOTHESIS Probably more important than selecting the relevant information (features, vertices, etc.) is the correct spatial positioning of these selected units according to the original’s spatial layout. As we suggested in this thesis, this requires an accurate encoding and representation of the relative spatial relationships between the features, which may be disrupted by the many eye and hand movements made throughout the drawing process. Understanding drawing expertise according to both aspects, perception and action, is critical as drawing is essentially a visuomotor task. Any expertise is thought to be a specialization for the particular constraints of the trained task. Hence, only focusing on perceptual aspects of drawing expertise may lead to findings of perceptual abilities not necessarily linked to the acquisition of drawing skill (Ericsson & Lehmann, 1996a).

In this thesis, we hypothesized that drawing skill would relate to a particular chunking of visual information more suited to the subsequent motor production. Chunking of information is an encoding mechanism characterizing forms of perceptual expertise whereby

158 individual features are grouped together according to meaningful relations and accessed as holistic units (Ahissar & Hochstein, 2004). For instance, when looking at a chessboard, experts chess players will group local spatial organizations of several pieces as meaningful legal moves (perceptual chunks), whereas novices will encode those same pieces independently (Chase & Simon, 1973; Reingold & Charness, 2005). After having identified and encoded local chunks on the chessboard, they can integrate them into a more complex representation according to their spatial relation. A similar mechanism would enable drawing experts to sequentially build an internal representation of the object’s spatial organization from local structures (grouped features or lines) encoded from each fixation made on the original and the copy throughout the drawing process. Because these local structures are integrated according to their relative spatial positions (e.g. according to eye-movements amplitude and direction), this would allow drawing experts to build an internal representation more robust to changes in retinal inputs and reference frames that can guide the production.

In a second study, we designed a gaze-contingent window experiment wherein our participants had to categorize an object’s structure as possible or impossible, while they were only allowed to see a portion of the object centered on their gaze position. Because discriminating possible from impossible objects requires identifying local features as violations regarding the global spatial organization of the object (Williams & Tarr, 1997), this forced our participants to sequentially build an internal representation of the whole object despite the limited amount of information available during each fixation. We found that more skilled and more experienced participants were able to perform in this task with smaller samples of foveal information, which is consistent with the hypothesis of a better integration of local information into a more complex internal representation. However, as we said earlier, building such a representation would require integrating local samples of information according to their spatial relationships. In order to characterize this process more deeply, it would be interesting to determine the nature of these spatial relationships: are they object- centered or eye-centered? In the latter case, amplitude and direction of an eye-movement made between two features could be used in order to measure their relative spatial positions. The implication of eye-movements in the formation and retrieval of internal representation has been previously documented (de Vito, Buonocore, Bonnefon, & Della Sala, 2014; Huette, Kello, Rhodes, & Spivey, 2013; Laeng & Teodorescu, 2002). This could be addressed by using a similar gaze-contingent window tasks but manipulating online features positions according to eye-movements direction and amplitude between an encoding and a test phase (e.g. same-different judgment task).

Two other predictions of the chunking hypothesis can be made on the basis of expert chess players performances. First, chunking information allows a faster, more efficient

159

GENERAL discussion | Discussion of the main results processing (Reingold, Charness, Schultetus, & Stampe, 2001). Second, size of perceptual chunks is thought to grow with experience and skill, which results in larger extent of the visual field that can be processed at a single glance (larger visual span; Rayner, 1998; Reingold, Charness, Pomplun, & Stampe, 2001a). This would allow drawing experts to process the global structure while focusing a particular part of the original or of the copy.

In a third study, we therefore designed a backward visual masking task using again a possible/impossible object decision task, wherein an object was briefly displayed centered on the gaze position (center of the screen) of participants. We varied the presentation duration of the object to measure the encoding efficiency as well as its overall size around the fixation location in order to measure participants’ visual span. We found that drawing experts were more efficient at encoding the object’s structure, although they were as much affected by the object size as novices. Hence, our findings would support our first prediction but not the second. Moreover, it raises the question of how relational (structural) information is encoded from each fixation. It has been found that expert chess players would encode local chunks in parallel whereas novices and intermediate players would use a serial processing (Reingold, Charness, Schultetus, et al., 2001). However, our skilled participants were as much affected by the object size as novices. This finding, consistent with a previous study (Donnelly et al., 1999), indicates a serial search where individual features are visited and continuously compared to the global organization of the object until a local structural violation is found. The greater efficiency of our more skilled participants could be attributed to several different factors that this experiment could not disentangle: a faster displacement of spatial attention between features, a faster encoding of features, or a faster switching between global and local aspects of the object. However, the serial search we observed could be attributed to our task (finding local structural violations) and may not rule out the hypothesis of an automatic processing of structural information.

In a second experiment, we found that more skilled draftspersons were also able to discriminate smaller structures located in visual periphery, which could not be accounted by a better visual acuity. In contrast to our first experiment that required covertly scanning a larger portion of the visual field around the fixation location, this second experiment involved the ability to scan closely spaced individual features in periphery. This situation may more closely match the reality of observational drawing. It has been shown that drawing experts segment the original into meaningful chunks of lines that can be drawn in a continuous hand movement (set of lines; Tchalenko, 2009), whereas novices copy isolated, unrelated lines. Interestingly, drawing experts’ fixations are not made directly on the segmented line but within a range of 5°. This may indicate that they encode surrounding, perifoveal features. However, closely spaced features in periphery are harder to access because of visual

160 crowding, which impair identification, recognition and position coding of features due to a limit of the spatial resolution of attention (Bouma, 1970; He et al., 1996, 1997). The better performances of more skilled participants in this experiment may therefore indicate that training in drawing may diminish the effects of crowding.

Taken together, these studies demonstrated that drawing skill is related to an enhanced processing and representation of objects structure, regardless of presentation conditions. We first hypothesized that a more efficient encoding of structural organization would allow drawing experts to build a more robust internal representation of both the original and the copy that could guide the drawing process.

In a final study, we designed an interactive drawing experiment coupled with a change detection task in order to examine how such a representation would intervene during the drawing process. Particularly, it remained unclear whether this internal representation would include the whole object structure or only local structures relevant to the being drawn segment. This study showed that positions directly related to the target segment currently being drawn were better remembered that any others and that original’s positions were more accurately represented by both novices and more skilled participants. The former finding is consistent with the idea of an offload of visual working memory in interactive visuomotor tasks, where only information relevant to the guidance of the ongoing action is held in visual working memory (Ballard et al., 1995). Interestingly, more skilled participants had a much more accurate memory representation for all the tested positions. However, this advantage was no longer observed in a change detection task where drawing was not involved. This may indicate a specialization of particular memory mechanisms for the constraints of observational drawing. In particular, both the visual monitoring and execution of memory- guided hand movements are thought to recruit similar representations in spatial working memory (Quinn & Ralston, 1986; Salway & Logie, 1995). This causes interference and impairment of spatial memory accuracy. Drawing experts could avoid this interference by either relying on more automatic processes or by recruiting non-spatial memory representations. The hypothesis of a chunking process in drawing experts would support the latter interpretation: drawing experts could retrieve individual positions from an internal representation of the object structure stored in long-term memory. However, because both alternatives are equally likely and because our experiment could not rule either out, further research should be conducted to examine whether the execution of hand movements during the drawing process is more automatic in drawing experts than in novices. This issue could be addressed by the use of online perturbations of the visual feedback during the production of drawings.

161

GENERAL discussion | Discussion of the main results

5.2.4 DRAWING IS NOT PLAYING CHESS: THE SPECIFICITY OF EXPERTISE

IN DRAWING So far, we have systematically compared perceptual and memory mechanisms involved in observational drawing to those shown by experts in chess. Of course, this does not mean that expertise in chess would transfer to drawing, or vice et versa. Acquisition of expertise resulting in higher performances in a domain is described as a progressive adaptation of brain mechanisms (perceptual, memory, or motor) to the constraints of the task (Ericsson & Lehmann, 1996b). What are the main differences between expertise in chess and expertise in drawing?

Expert chess players perceptual and memory performances are strictly limited to the context of legal chess game. Most of studies reported that when presented with random organization of chess pieces, experts’ performances dropped at those of novices (Chase & Simon, 1973; Gobet & Simon, 1996b; but see Bilalić, Langner, Erb, & Grodd, 2010). It has therefore been suggested that structural processing of pieces by experts chess players would be domain specific and not related to enhanced ability of processing spatial relationships (e.g. in a random pattern). Particularly, perceptual and memory performances of expert chess players would be attributed to the use of chess game templates accumulated over games played in the past and stored in long-term memory: expert players recognize structural organization of pieces in the current chess game evoking chess moves previously seen in past games (Gobet & Simon, 1996a). In that sense, expertise in chess is the acquisition of a greater vocabulary of coherent patterns given chess rules (Gobet & Simon, 1998). This is consistent with the general characterization of acquisition of expertise as an attentional bias toward task-relevant information and driven by explicit and implicit knowledge stored in long-term memory, which results in more efficiently tuned high-level perceptual representations (Ahissar & Hochstein, 2004). However, it also raises the several questions: what are the attentional units (chunks) in observational drawing and what knowledge would drive their selection?

At a first glance, objects would seem the most natural perceptual units (perceptual chunks) upon which drawing experts operate. However, drawing is a sequential process requiring a correct segmentation of objects into simpler sets of lines. The meaningfulness of the relation between lines included in a chunk may depend on whether this particular set of lines forms a relevant featural group (e.g. junction), but also on whether it can be drawn in a single and continuous hand movement (Tchalenko, 2009; Van Sommers, 1984). This would suggest a link between perceptual chunking and motor characteristics of hand movements. If this link should be confirmed, it would be important to determine whether the size of the

162 perceptual chunks depends on the length of motor sequence that can be programmed at once by draftspersons.

We have seen that drawing experts know better what information must be selected and represented in order to trigger a good recognition of the scene and to increase its realism (Cavanagh, 2005; Kozbelt et al., 2010). However, we suggested that this explicit knowledge may not be at the core of drawing expertise. Processing spatial layout may be more critical is one wishes to make accurate copies of scenes and objects. In that sense, perceptual units selected by experts in drawing would be spatial structures characterizing spatial relationships of relevant features (e.g. vertices, junctions). Consistently, our studies clearly showed that drawing experts had an advantage in processing structural information. Nevertheless, this more efficient processing may not rely on the recruitment of stored templates in working memory in order to match the current layout to previously seen ones. Indeed, a peculiarity of observational drawing is that the spatial layout may change from drawing to drawing, since it is strictly dependent on viewpoint. Interestingly, this advantage in processing structure was also seen for impossible structures, which would correspond to illegal move situations in chess where experts lose their advantage. A possible but speculative account for this finding is that, due to their extensive observation of scenes and objects’ structures, drawing experts acquired a greater knowledge of structural environmental statistics (Brady, Konkle, & Alvarez, 2009) – how objects and features spatially relate in natural images. An extensive explicit analysis of objects may give extra knowledge of how perceptual units could be organized in objects. This would give artists greater understanding of the syntax of vision and therefore would make artists more efficient in detecting violations to these syntactic rules.

5.2.5 TALENT VS TRAINING Our studies demonstrated that drawing skill is related to specific perceptual and memory mechanisms more adapted to the particular constraints of observational drawing: a more efficient encoding of structural visual organization and the ability to build visual internal representations more robust to changes in visual inputs and reference frames. Our studies did not establish, though, whether these abilities would be the byproduct of the intensive training in drawing that characterizes drawing experts, or whether they would be only present in innately talented people (Winner & Drake, 1996). For instance, innate talent (due to genetic factors) could explain why only some children choose to continue practicing drawing in adolescence and adulthood. Research on the acquisition of expertise and skill has suggested that practice could indeed improve performances of any individual, but that this improvement would be limited by an idiosyncratic upper asymptote (Sternberg, 2012). The

163

GENERAL discussion | Conclusion distance between this asymptote and expert performance would indicate what is due to innate talent. However, there is no direct evidence for this hypothesis and most of studies investigating expertise and skills support the hypothesis of a maximal adaptation of already existent physiological and psychological apparatus to the particular task constraints as long as it is supported by a strong motivation and deliberate practice (Ericsson & Charness, 1994; Ericsson & Lehmann, 1996b). Beyond the metaphysical debate between innate vs. acquired abilities, investigating the improvement of particular brain mechanisms through practice is critical in order to address whether identified mechanisms in experts would be true invariants of the brain functioning or only specific to some individuals. Although we did find a strong relationship between performances in our gaze-contingent window experiment (study 2) and years of practice in drawing, our studies were not designed to address these legitimate issues. Potentially, longitudinal studies would be required: training novices in drawing and measuring to what extent the identified perceptual and memory abilities would improve with time compared to control, untrained participants. A recent unpublished study found that training students in gestures drawing over 4 months resulted in improvements in drawing accuracy as well as changes in brain structures (Schlegel et al., 2012). But again, this study cannot tell whether training alone suffices to reach expert performance and accuracy levels. Expertise in drawing as in other disciplines is often related to daily intensive practice over dozens of years, but unfortunately most of longitudinal studies are only conduced over much shorter time-scales.

5.3 CONCLUSION

Observational drawing involves many different cognitive, perceptual, memory and motor skills. In this thesis, we focused on the most fundamental aspect of drawing expertise: the processing and robust internal representation of spatial relationships. Our studies demonstrated that these aspects of drawing skill could be accounted for by perceptual and memory mechanisms thought to underlie other forms of expertise: a chunking process.

164

165

GENERAL discussion |

166

6 BIBLIOGRAPHY

167

ERROR! USE THE HOME TAB TO APPLY TITRE 1 TO THE TEXT THAT YOU WANT TO APPEAR HERE. | Error! Use the Home tab to apply Titre 2 to the text that you want to appear here.

168

Adelson, E. H. (1993). Perceptual organization and the judgment of brightness. Science, 262(December), 2042–4.

Adelson, E. H. (2000). Lightness Perception and Lightness Illusions. Perception, 3, 339–351.

Adi-Japha, E., & Freeman, N. H. (2001). Development of differentiation between writing and drawing systems. Developmental Psychology, 37(1), 101–114.

Agostini, T., & Galmonte, A. (2002). Perceptual organization overcomes the effects of local surround in determining simultaneous lightness contrast. Psychological Science, 13(1), 89.

Ahissar, M., & Hochstein, S. (2004). The reverse hierarchy theory of visual perceptual learning. Trends in Cognitive Sciences, 8(10), 457–64.

Aks, D. J., & Enns, J. T. (1996). Visual search for size is influenced by a background texture gradient. Journal of Experimental Psychology. Human Perception and Performance, 22(6), 1467–81.

Alvarez, G. A., & Cavanagh, P. (2004). The capacity of visual short-term memory is set both by visual information load and by number of objects. Psychological Science, 15(2), 106– 11.

Ames, A. (1925). Depth in pictorial art. The Art Bulletin, 8(1), 4–24.

Arend, L. E., & Spehar, B. (1993). Lightness, brightness, and brightness contrast: 1. Illuminance variation. Attention, Perception, & Psychophysics, 54(4), 446–456.

Arend, L. E., & Spehar, B. (1993). Lightness, brightness, and brightness contrast: 2. Reflectance variation. Perception & Psychophysics, 54(4), 457–68.

Arnheim, R. (1954). Art and Visual Perception: A Psychology of the Creative Eye (p. 508). Berkeley: University of California Press.

Awh, E., Jonides, J., & Reuter-Lorenz, P. (1998). Rehearsal in spatial working memory. Journal of Experimental Psychology: Human Perception and Performance, 24(3), 780– 790.

169

ERROR! USE THE HOME TAB TO APPLY TITRE 1 TO THE TEXT THAT YOU WANT TO APPEAR HERE. | Error! Use the Home tab to apply Titre 2 to the text that you want to appear here.

Baddeley, A. (1992). Working memory. Science, 255, 556–559.

Ballard, D., Hayhoe, M., & Pelz, J. B. (1995). Memory representations in natural tasks. Journal of Cognitive Neuroscience, 7(1), 66–80.

Ballard, D., Hayhoe, M., Pook, P. K., & Rao, R. P. N. (1997). Deictic codes for the embodiment of cognition. Behavioral and Brain Sciences, 20, 723–767.

Bates, D., Maechler, M., & Bolker, B. (2012). lme4: Linear mixed-effects models using S4 classes.

Bennett, D. J., & Warren, W. (2002). Size scaling: retinal or environmental frame of reference? Perception & Psychophysics, 64(3), 462–77.

Biederman, I. (1987). Recognition-by-components: A theory of human image understanding. Psychological Review, 94(2), 115–117.

Biederman, I., & Kim, J. G. (2008). 17 000 Years of Depicting the Junction of Two Smooth Shapes. Perception, 37(1), 161–164.

Bilalić, M., Langner, R., Erb, M., & Grodd, W. (2010). Mechanisms and neural basis of object and pattern recognition: a study with chess experts. Journal of Experimental Psychology. General, 139(4), 728–742.

Bock, O., Dose, M., Ott, D., & Eckmiller, R. (1990). Control of arm movements in a 2- dimensional pointing task. Behavioural Brain Research, 40(3), 247–250.

Bollen, K. A., & Jackman, R. W. (1985). Regression Diagnostics: An Expository Treatment of Outliers and Influential Cases. Sociological Methods & Research, 13(4), 510–542.

Bouma, H. (1970). Interaction effects in parafoveal letter recognition. Nature, 226, 177–178.

Brady, T. F., Konkle, T., & Alvarez, G. a. (2009). Compression in visual working memory: using statistical regularities to form more efficient memory representations. Journal of Experimental Psychology. General, 138(4), 487–502.

170

Brainard, D. H. (1997). The Psychophysics Toolbox. Spatial Vision, 10(4), 433–436.

Broderick, P., & Laszlo, J. I. (1987). The drawing of squares and diamonds: a perceptual- motor task analysis. Journal of Experimental Child Psychology, 43(1), 44–61.

Bruno, N., Bertamini, M., & Domini, F. (1997). Amodal completion of partly occluded surfaces: Is there a mosaic stage? Journal of Experimental Psychology. Human Perception and Performance, 23(5), 1412–1426.

Cai, R. H., Pouget, A., Schlag-Rey, M., & Schlag, J. (1997). Perceived geometrical relationships affected by eye-movement signals. Nature, 386(6625), 601–604.

Calabrese, L., & Marucci, F. S. (2006). The influence of expertise level on the visuo-spatial ability: differences between experts and novices in imagery and drawing abilities. Cognitive Processing, 7(S1), 118–120.

Canny, J. (1986). A Computational Approach to Edge Detection. IEEE Transactions on Pattern Analysis and Machine Intelligence, 8(6), 679–698.

Carlson, J. A. (1966). Effect of Instructions and Perspective-Drawing Ability on Perceptual Constancies and Geometrical Illusions. Journal of Experimental Psychology, 72(6), 874–879.

Carlson, V. R. (1960). Overestimation in Size-Constancy Judgments. The American Journal of Psychology, 73(2), 199–213.

Carlson, V. R. (1962). Size-constancy judgments and perceptual compromise. Journal of Experimental Psychology, 63(1), 68–73.

Carson, L., Millard, M., Quehl, N., & Danckert, J. (2012). Drawing expertise predicts not just quality but also accuracy. Perception, 41(ECVP Abstract Supplement), 231.

Carson, L., Quehl, N., Aliev, I., & Danckert, J. (2013). Angle-based Drawing Accuracy Analysis and Mental Models of Three-Dimensional Space. Art & Perception, 0, 1–30.

Cavanagh, P. (2005). The artist as neuroscientist. Nature, 434(7031), 301–7.

171

ERROR! USE THE HOME TAB TO APPLY TITRE 1 TO THE TEXT THAT YOU WANT TO APPEAR HERE. | Error! Use the Home tab to apply Titre 2 to the text that you want to appear here.

Cavanagh, P., & Leclerc, Y. G. (1989). Shape from shadows. Journal of Experimental Psychology. Human Perception and Performance, 15(1), 3–27.

Chakravarthi, R., & VanRullen, R. (2011). Bullet trains and steam engines : Exogenous attention zips but endogenous attention chugs along. Journal of Vision, 11(4), 1–12.

Chamberlain, R., & McManus, I. C. (2013). Subjective and Objective Measures of Drawing Accuracy and their Relationship to Perceptual Abilities. Perception, 42(ECVP Abstract Supplement), 106.

Chamberlain, R., McManus, I. C., Brunswick, N., Rankin, Q., Riley, H., & Kanai, R. (2014). Drawing on the right side of the Brain: A Voxel-based Morphometry analysis of observational Drawing. NeuroImage.

Chase, W. G., & Simon, H. A. (1973). Perception in chess. Cognitive Psychology, 4(1), 55– 81.

Chatterjee, A. (2004). The neuropsychology of visual artistic production. Neuropsychologia, 42(11), 1568–83.

Coen-Cagli, R., Coraggio, P., Napoletano, P., Schwartz, O., Ferraro, M., & Boccignone, G. (2009). Visuomotor characterization of eye movements in a drawing task. Vision Research, 49(8), 810–8.

Cohen, D. J. (2005). Look little, look often: The influence of gaze frequency on drawing accuracy. Perception & Psychophysics, 67(6), 997–1009.

Cohen, D. J., & Bennett, S. (1997). Why can’t most people draw what they see? Journal of Experimental Psychology: Human Perception and Performance, 23(3), 609–621.

Cohen, D. J., & Earls, H. (2010). Inverting an image does not improve drawing accuracy. Psychology of Aesthetics, Creativity, and the Arts, 4(3), 168–172.

Cohen, D. J., & Jones, H. E. (2008). How shape constancy relates to drawing accuracy. Psychology of Aesthetics, Creativity, and the Arts, 2(1), 8–19.

172

Cook, D. R. (1979). Influential in Linear Observations Regression. Journal of the American Statistical Association, 74(365), 169–174.

Cowie, R., & Perrott, R. (1993). From line drawings to impressions of 3D objects: developing a model to account for the shapes that people see. Image and Vision Computing, 11(6), 342–352.

Crawford, J. D., Medendorp, W. P., & Marotta, J. J. (2004). Spatial transformations for eye- hand coordination. Journal of Neurophysiology, 92(1), 10–9.

Curby, K. M., Glazek, K., & Gauthier, I. (2009). A Visual Short-Term Memory Advantage for Objects of Expertise. Journal of Experimental Psychology. Human Perception and Performance, 35(1), 94 –107.

Cutzu, F., & Edelman, S. (1994). Canonical views in object representation and recognition. Vision Research, 34(22), 3037–56.

Day, R. (1972). The basis of perceptual constancy and perceptual illusion. Investigative Ophthalmology, 11(6), 525–32.

De Vito, S., Buonocore, A., Bonnefon, J.-F., & Della Sala, S. (2014). Eye movements disrupt spatial but not visual mental imagery. Cognitive Processing, 1(1), 1–7.

De Winter, J., & Wagemans, J. (2006). Segmentation of object outlines into parts: a large- scale integrative study. Cognition, 99(3), 275–325.

Del Giudice, E., Grossi, D., Angelini, R., Crisanti, a F., Latte, F., Fragassi, N. a, & Trojano, L. (2000). Spatial cognition in children. I. Development of drawing-related (visuospatial and constructional) abilities in preschool and early school years. Brain & Development, 22(6), 362–7.

DiCiccio, T. J., Efron, B., Hall, P., Martin, M. A., Canty, A. J., Davison, A. C., … Young, G. A. (1996). Bootstrap confidence intervals. Statistical Science, 11(3), 189–228.

Donnelly, N., Found, a, & Müller, H. J. (1999). Searching for impossible objects: processing form and attributes in early vision. Perception & Psychophysics, 61(4), 675–90.

173

ERROR! USE THE HOME TAB TO APPLY TITRE 1 TO THE TEXT THAT YOU WANT TO APPEAR HERE. | Error! Use the Home tab to apply Titre 2 to the text that you want to appear here.

Draganski, B., & May, A. (2008). Training-induced structural changes in the adult human brain. Behavioural Brain Research, 192, 137–142.

Drake, J. E. (2013). Is superior local processing in the visuospatial domain a function of drawing talent rather than autism spectrum disorder? Psychology of Aesthetics, Creativity, and the Arts, 7(2), 203–209.

Elliott, D., & Calvert, R. (1990). The influence of uncertainty and premovement visual information on manual aiming. Canadian Journal of Psychology, 44(4), 501–511.

Elo, A. E. (1978). The rating of chessplayers, past and present. London: Batsford.

Ericsson, K. A., & Charness, N. (1994). Expert performance: Its structure and acquisition. American Psychologist, 49(8), 725.

Ericsson, K. A., & Lehmann, A. C. (1996a). Expert and exceptional performance: Evidence of maximal adaptation to task constraints. Annual Review of Psychology, 47, 273–305.

Ericsson, K. A., & Lehmann, A. C. (1996b). Expert and exceptional performance: Evidence of maximal adaptation to task constraints. Annual Review of Psychology, 47, 273–305.

Favilla, M., Hening, W., & Ghez, C. (1989). Trajectory control in targeted force impulses. Experimental Brain Research, 75(2).

Feder, K. P., & Majnemer, A. (2007). Handwriting development, competency, and intervention. Developmental Medicine and Child Neurology, 49(4), 312–7. x

Freud, E., Avidan, G., & Ganel, T. (2013). Holistic Processing of Impossible Objects: Evidence from Garner’s speeded-classification task. Vision Research.

Freud, E., Ganel, T., & Avidan, G. (2013). Representation of possible and impossible objects in the human visual cortex: evidence from fMRI adaptation. NeuroImage, 64, 685–92.

Geisler, W. S., Perry, J. S., & Najemnik, J. (2006). Visual search: the role of peripheral information measured using gaze-contingent displays. Journal of Vision, 6(9), 858–73.

174

Gilchrist, A. L. (1988). Lightness contrast and failures of constancy: a common explanation. Perception & Psychophysics, 43(5), 415–24.

Gilchrist, A. L. (2006). Seeing Black and White. European Journal of Neurology (Vol. 14). New York: Oxford University Press, USA.

Glazek, K. (2012). Visual and motor processing in visual artists: Implications for cognitive and neural mechanisms. Psychology of Aesthetics, Creativity, and the Arts, 6(2), 155– 167.

Gobet, F., & Simon, H. a. (1996a). Templates in chess memory: a mechanism for recalling several boards. Cognitive Psychology, 31(1), 1–40.

Gobet, F., & Simon, H. a. (1998). Expert chess memory: revisiting the chunking hypothesis. Memory (Hove, England), 6(3), 225–55.

Gobet, F., & Simon, H. A. (1996b). Recall of rapidly presented random chess positions is a function of skill. Psychonomic Bulletin & Review, 3(2), 493–503.

Goldstone, R. L. (1998). Perceptual learning. Annual Review of Psychology, 49, 585–612.

Gombrich, E. (1960). Art and Illusion : A Study in the Psychology of Pictorial Representation Summary (Fifth edit.). Oxford: Phaidon Press Limited.

Gowen, E., & Miall, R. C. (2006). Eye-hand interactions in tracing and drawing tasks. Human Movement Science, 25(4-5), 568–85.

Graham, D. J., & Meng, M. (2011). Artistic representations: clues to efficient coding in human vision. Visual Neuroscience, 28, 1–9.

Green, C. S., & Bavelier, D. (2003). Action video game modifies visual selective attention. Nature, 423(6939), 534–7.

Green, C. S., & Bavelier, D. (2007). Action-video-game experience alters the spatial resolution of vision. Psychological Science, 18(1), 88–94.

175

ERROR! USE THE HOME TAB TO APPLY TITRE 1 TO THE TEXT THAT YOU WANT TO APPEAR HERE. | Error! Use the Home tab to apply Titre 2 to the text that you want to appear here.

Green, C. S., & Bavelier, D. (2008). Exercising your brain: a review of human brain plasticity and training-induced learning. Psychology and Aging, 23(4), 692–701.

Greene, M. R., & Oliva, A. (2009). Recognition of natural scenes from global properties. Cognitive Psychology, 58(2), 137–176.

Harris, C., & Stephens, M. (1988). A Combined Corner and Edge Detector. In Proceedings of the 4th Alvey Vision Conference (pp. 147–152). Manchester, UK: Alvey Vision Club.

Hayes, S., & Milne, N. (2011). What’s wrong with this picture? an experiment in quantifying accuracy in 2D portrait drawing. Visual Communication, 10(2), 149–174.

Hayhoe, M., Mruczek, R., & Pelz, J. B. (2003). Visual memory and motor planning in a natural task. Journal of Vision, 3, 49–63.

Hayward, W. G. (2003). After the viewpoint debate: where next in object recognition? Trends in Cognitive Sciences, 7(10), 425–7.

He, S., Cavanagh, P., & Intriligator, J. (1996). Attentional resolution and the locus of visual awareness. Nature, 383(3), 334–337.

He, S., Cavanagh, P., & Intriligator, J. (1997). Attentional resolution. Trends in Cognitive Sciences, 1(3), 115–21.

He, Z. J., & Nakayama, K. (1992). Surfaces versus features in visual search. Nature, 359(6392), 231–233.

Hesse, C., & Franz, V. H. (2009). Memory mechanisms in grasping. Neuropsychologia, 47, 1532–1545.

Hochstein, S., & Ahissar, M. (2002). View from the Top:: Hierarchies and Reverse Hierarchies in the Visual System. Neuron, 36(5), 791–804.

Holm, S. (1979). A simple sequentially rejective multiple test procedure. Scandinavian Journal of Statistics, 6(2), 65–70.

176

Hubel, D., & Wiesel, T. (1970). The period of susceptibility to the physiological effects of unilateral eye closure in kittens. The Journal of Physiology, 206, 419–436.

Huette, S., Kello, C. T., Rhodes, T., & Spivey, M. J. (2013). Drawing from memory: hand-eye coordination at multiple scales. PloS One, 8(3), e58464.

Kakei, S., Hoffman, D. S., & Strick, P. L. (2003). Sensorimotor transformations in cortical motor areas. Neuroscience Research, 46(1), 1–10.

Kalaska, J. F., & Crammond, D. J. (1992). Cerebral cortical mechanisms of reaching movements. Science, 255(5051), 1517–23.

Kanizsa, G. (1979). Organization in vision: Essays on Gestalt perception. New York: Praeger.

Kanizsa, G. (1985). Seeing and thinking. Acta Psychologica, 59(1), 23–33.

Kellman, P. J., & Shipley, T. F. (1991). A theory of visual interpolation in object perception. Cognitive Psychology, 23(2), 141–221.

Kennedy, J. (1974). A psychology of picture perception. Perception. San Francisco: Jossey- Bass Publishers.

Kesten, H. (1958). Accelerated Stochastic Approximation. The Annals of Mathematical Statistics, 29(1), 41–59.

Kosslyn, S. M. (1987). Seeing and imagining in the cerebral hemispheres: a computational approach. Psychological Review, 94(2), 148–75.

Kozbelt, A. (2001). Artists as experts in visual cognition. Visual Cognition, 8(6), 705–723.

Kozbelt, A., & Seeley, W. P. (2007). Integrating art historical, psychological, and neuroscientific explanations of artists’ advantages in drawing and perception. Psychology of Aesthetics, Creativity, and the Arts, 1(2), 80–90.

177

ERROR! USE THE HOME TAB TO APPLY TITRE 1 TO THE TEXT THAT YOU WANT TO APPEAR HERE. | Error! Use the Home tab to apply Titre 2 to the text that you want to appear here.

Laeng, B., & Teodorescu, D.-S. (2002). Eye scanpath during visual imagery reenact those of perception of the same visual scene. Cognitive Science, 26, 207–231.

Land, M. F. (2006). Eye movements and the control of actions in everyday life. Progress in Retinal and Eye Research, 25(3), 296–324.

Laszlo, J. I., & Bairstow, P. J. (1985). Perceptual-motor behaviour : developmental assessment and therapy. New York, NY: Praeger.

Lawrence, B. M., Myerson, J., Oonk, H. M., & Abrams, R. a. (2001). The effects of eye and limb movements on working memory. Memory (Hove, England), 9, 433–444.

Lee, H., & Vecera, S. P. (2005). Visual cognition influences early vision: the role of visual short-term memory in amodal completion. Psychological Science : A Journal of the American Psychological Society / APS, 16(10), 763–8.

Lee, M. (1989). When is an object not an object? The effect of “meaning” upon the copying of line drawings. British Journal of Psychology, 80, 15–37.

Leibowitz, H., & Harvey, L. O. (1967). Size matching as a function of instructions in a naturalistic environment. Journal of Experimental Psychology, 74(3), 378.

Leibowitz, H. W., & Harvey, L. O. (1969). Effect of instructions, environment, and type of test object on matched size. Journal of Experimental Psychology, 81(1), 36–43.

Li, R., Polat, U., Makous, W., & Bavelier, D. (2009). Enhancing the contrast sensitivity function through action video game training. Nature Neuroscience, 12(5), 549–551.

Locher, P. (2010). How Does a Visual Artist Create an Artwork? In J. C. Kaufman & R. J. Sternberg (Eds.), The Cambridge Handbook of Creativity (pp. 131–144). Cambridge: Cambridge University Press.

Marr, D., & Nishihara, H. K. (1978). Representation and recognition of the spatial organization of three-dimensional shapes. Proceedings of the Royal Society of London. Series B, Containing Papers of a Biological Character. Royal Society (Great Britain), 200(1140), 269–94.

178

Martelli, M., Majaj, N. J., & Pelli, D. G. (2005). Are faces processed like words? A diagnostic test for recognition by parts. Journal of Vision, 5, 58–70.

Matthews, W. J., & Adams, A. (2008). Another reason why adults find it hard to draw accurately. Perception, 37(4), 628–630.

Meng, X., Rosenthal, R., & Rubin, D. (1992). Comparing Correlated Correlation Coefficients. Psychological Bulletin, 111(1), 172–175.

Miall, R. C., Gowen, E., & Tchalenko, J. (2008). Drawing cartoon faces--a functional imaging study of the cognitive neuroscience of drawing. Cortex, 45(3), 394–406.

Miall, R. C., & Tchalenko, J. (2001). A Painter’s Eye Movements: A Study of Eye and Hand Movement during Portrait Drawing. Leonardo, 34(1), 35–40.

Milner, A. D., & Goodale, M. A. (1998). The Visual Brain in Action. Brain, (27).

Mitchell, P., Ropar, D., Ackroyd, K., & Rajendran, G. (2005). How Perception Impacts on Drawings. Journal of Experimental Psychology. Human Perception and Performance, 31(5), 996 –1003.

Mitchell, P., & Taylor, L. M. (1999). Shape constancy and theory of mind : is there a link ? Cognition, 70, 167–190.

Moore, C. M., & Brown, L. E. (2001). Preconstancy information can influence visual search: The case of lightness constancy. Journal of Experimental Psychology: Human Perception and Performance, 27(1), 178–194.

Myles-Worsley, M., Johnston, W. a, & Simons, M. a. (1988). The influence of expertise on X- ray image processing. Journal of Experimental Psychology. Learning, Memory, and Cognition, 14(3), 553–7.

179

ERROR! USE THE HOME TAB TO APPLY TITRE 1 TO THE TEXT THAT YOU WANT TO APPEAR HERE. | Error! Use the Home tab to apply Titre 2 to the text that you want to appear here.

Navon, D. (1977). Forest before trees: The precedence of global features in visual perception. Cognitive Psychology, 9, 353–383.

New, B., Pallier, C., Brysbaert, M., & Ferrand, L. (2004). Lexique 2 : A new French lexical database. Behavior Research Methods, Instruments, & Computers, 36(3), 516–524.

Ogawa, K., Nagai, C., & Inui, T. (2010). Brain mechanisms of visuomotor transformation based on deficits in tracing and copying. Japanese Psychological Research, 52(2), 91– 106.

Ostrofsky, J., Kozbelt, A., & Kurylo, D. (2013). Perceptual grouping in artists and non-artists: A psychophysical comparison. Empirical Studies of the Arts, 31(2), 131–143.

Ostrovsky, Y., Andalman, A., & Sinha, P. (2006). Vision Following Extended Congenital Blindness. Psychological Science : A Journal of the American Psychological Society / APS, 17(12), 1009–1015.

Pelli, D. G. (1997). The VideoToolbox software for visual psychophysics: transforming numbers into movies. Spatial Vision, 10(4), 437–442.

Penrose, L. S., & Penrose, R. (1958). Impossible object: A special type of visual illusion. British Journal of Psychology, 49(1), 31–33.

Pepperell, R. (2006). Seeing without objects: visual indeterminacy and art. Leonardo.

Pepperell, R., & Haertel, M. (2014). Do artists use linear perspective to depict visual space? Perception, 1–22.

Perdreau, F., & Cavanagh, P. (2011). Do artists see their retinas? Frontiers in Human Neuroscience, 5(171), 1–10.

Perdreau, F., & Cavanagh, P. (2013a). Is Artists’ Perception more Veridical? Frontiers in Neuroscience, 7(6), 1–11.

Perdreau, F., & Cavanagh, P. (2013b). The artist’s advantage: Better integration of object information across eye movements. I-Perception, 4(6), 380–395.

180

Perdreau, F., & Cavanagh, P. (2014). Drawing skill is related to the efficiency of encoding object structure. I-Perception, 5(2), 101–119.

Piaget, J., & Inhelder, B. (1967). The Child’s Conception of Space (p. 490). New York: W.W. Norton.

Picard, D., & Durand, K. (2005). Are young children’s drawings canonically biased? Journal of Experimental Child Psychology, 90(1), 48–64.

Posner, M., Petersen, S., Fox, P., & Raichle, M. (1988). Localization of cognitive operations in the human brain. Science, 240(4859), 1627–1631.

Prins, N. (2012). The psychometric function: the lapse rate revisited. Journal of Vision, 12(6), 1–16.

Quinn, J. G., & Ralston, G. E. (1986). Movement and attention in visual working memory. The Quarterly Journal of Experimental Psychology Section A, 38(4), 689–703.

Rand, C. W. (1973). Copying in Drawing: The Importance of Adequate Visual Analysis versus the Ability to Utilize Drawing Rules. Child Development, 44(1), 47.

Rayner, K. (1998). Eye movements in reading and information processing: 20 years of research. Psychological Bulletin, 124(3), 372–422.

Rayner, K., McConkie, G. W., & Zola, D. (1980). Integrating information across eye movements. Cognitive Psychology, 12(2), 206–226.

Reingold, E. M., & Charness, N. (2005). Perception in chess: Evidence from eye movements. In Cognitive processes in eye guidance.

Reingold, E. M., Charness, N., Pomplun, M., & Stampe, D. M. (2001a). Visual Span in Expert chess player: Evidence From Eye Movements. Psychological Science, 12(1), 48.

Reingold, E. M., Charness, N., Pomplun, M., & Stampe, D. M. (2001b). Visual Span in Expert Chess Players: Evidence From Eye Movements. Psychological Science, 12(1), 48–55.

Reingold, E. M., Charness, N., Schultetus, R. S., & Stampe, D. M. (2001). Perceptual automaticity in expert chess players: parallel encoding of chess relations. Psychonomic Bulletin & Review, 8(3), 504–510.

181

ERROR! USE THE HOME TAB TO APPLY TITRE 1 TO THE TEXT THAT YOU WANT TO APPEAR HERE. | Error! Use the Home tab to apply Titre 2 to the text that you want to appear here.

Rensink, R. A., & Enns, J. T. (1998). Early completion of occluded objects. Vision Research, 38(15-16), 2489–505.

Rensink, R. A., O’Regan, J. K., & Clark, J. J. (1997). To see or not to see: The need for attention to perceive changes in scenes. Psychological Science, 8(5), 1–6.

Roccasecca, P. (2009). Teaching in the Studio of the “Accademia del Disegno dei pittori, scultori e architetti di Roma.” Conservation Research, 34, 123–159.

Rock, I. (1983). The Logic of Perception. Cambridge, MA: MIT press.

Rosenblatt, E., & Winner, E. (1988). The art of children’s drawing. Journal of Aesthetic Education, 22(1), 3–15.

Ruskin, J. (1912). Elements of drawing. New York. London: J.M. Dent & sons.

Salway, A. F., & Logie, R. H. (1995). Visuospatial working memory, movement control and executive demands. British Journal of Psychology (London, England : 1953).

Schacter, D. L., Cooper, L. A., & Delaney, S. M. (1990). Implicit memory for unfamiliar objects depends on access to structural descriptions. Journal of Experimental Psychology. General, 119, 5–24.

Schlegel, A., Fogelson, S., Li, X., Lu, Z., Alexander, P., Meng, M., & Tse, P. (2012). Visual art training in young adults changes neural circuitry in visual and motor areas. Journal of Vision, 12(9), 1129–1129.

Schneider, W. X. (1999). Visual-spatial working memory, attention, and scene representation: a neuro-cognitive theory. Psychological Research, 62(2-3), 220–36.

Seamon, J., & Carrasco, M. (1999). The effect of study time on priming possible and impossible figures in the object decision test. Psicothema, 11(4), 801–813.

182

Seeley, W. P., & Kozbelt, A. (2008). Art, Artists, and Perception: A Model for Premotor Contributions to Perceptual Analysis and Form Recognition. Philosophical Psychology, 21(2), 149–171.

Shepard, R. N., & Judd, S. a. (1976). Perceptual illusion of rotation of three-dimensional objects. Science (New York, N.Y.), 191(4230), 952–4.

Simons, D. J., & Rensink, R. A. (2005). Change blindness: past, present, and future. Trends in Cognitive Sciences, 9(1), 16–20.

Soldan, A., Hilton, H. J., & Stern, Y. (2009). Bias effects in the possible/impossible object decision test with matching objects. Memory & Cognition, 37(2), 685–692.

Solso, R. (2001). Brain activities in a skilled versus a novice artist An fMRI study. Leonardo, 34(1), 31–34.

Solso, R. L. (1996). Cognition and the Visual Arts (p. 312).

Sternberg, R. J. (2012). Giftedness as developing expertise. In K. A. Heller, F. J. Mönks, R. J. Sternberg, & R. F. Subotnik (Eds.), International Handbook of giftedness and talent (pp. 55–66). Oxford, UK: Elsevier Science Ltd.

Stuart, G. W., Bossomaier, T. R. J., Johnson, S., & If, G. W. S. (1993). Preattentive processing of object size: implications for theories of size perception. Perception, 22(10), 1175–93.

Taylor, L. M., & Mitchell, P. (1997). Judgments of apparent shape contaminated by knowledge of reality: Viewing circles obliquely. British Journal of Psychology, 88(4), 653–670.

Tchalenko, J. (2007). Eye movements in drawing simple lines. Perception, 36(8), 1152– 1167.

Tchalenko, J. (2009). Segmentation and accuracy in copying and drawing: experts and beginners. Vision Research, 49(8), 791–800.

Tchalenko, J., Dempere-Marco, L., Hu, X. P., & Yang, G. Z. (2003). Eye Movement and Voluntary Control in Portrait Drawing. In The Mind’s Eye: Cognitive and applied Aspects of Eye Movement Research (pp. 705–728). Elsevier Science Ltd.

183

ERROR! USE THE HOME TAB TO APPLY TITRE 1 TO THE TEXT THAT YOU WANT TO APPEAR HERE. | Error! Use the Home tab to apply Titre 2 to the text that you want to appear here.

Tchalenko, J., & Miall, R. C. (2009). Eye–hand strategies in copying complex lines. Cortex, 45(3), 368–376.

Thorpe, S., Fize, D., & Marlot, C. (1996). Speed of processing in the human visual system. Nature, 381, 520–522.

Thouless, R. H. (1931). Phenomenal regression to the real object. I. British Journal of Psychology. General Section, 21(4), 339–359.

Thouless, R. H. (1932). Individual differences in phenomenal regression. British Journal of Psychology. General Section, 22(3), 216–241.

Todorovic, D. (2002). Constancies and illusions in visual perception. Psihologija, 35(3-4), 125–207.

Todorovic, D. (2010). Context effects in visual perception and their explanations. Review of Psychology, 17(1), 17–32.

Todorović, D. (2002). Constancies and illusions in visual perception. Psihologija, 35(3-4), 125–207.

Toet, A., & Levi, D. (1992). The two-dimensional shape of spatial interaction zones in the parafovea. Vision Research, 32(7), 1349–1357.

Treisman, A., & Gelade, G. (1980). A feature-integration theory of attention. Cognitive Psychology, 12(1), 97–136.

Treisman, A. M., & Gelade, G. (1980). A feature-integration theory of attention. Cognitive Psychology, 12(1), 97–136.

Tse, P. U. (1999). Volume completion. Cognitive Psychology, 39(1), 37–68.

Tyler, C. W. (2011). Paradoxical perception of surfaces in the Shepard tabletop illusion. I- Perception, 2(2), 137–41.

Valois, K. De, & Lakshminarayanan, V. (1990). Discrimination of relative spatial position. Vision Research, 30(11), 1649–1660.

Van Sommers, P. (1984). Drawing and cognition: Descriptive and experimental studies of graphic production processes. New York: Cambridge Univ Press.

184

Van Sommers, P. (1989). A system for drawing and drawing-related neuropsychology. Cognitive Neuropsychology, 6(2), 117–164.

Vuilleumier, P., Henson, R. N., Driver, J., & Dolan, R. J. (2002). Multiple levels of visual object constancy revealed by event-related fMRI of repetition priming. Nature Neuroscience, 5(5), 491–9.

Wheeler, M. E., & Treisman, A. (2002). Binding in short-term visual memory. Journal of Experimental Psychology: General, 131(1), 48–64.

Whitney, D., & Levi, D. M. (2011). Visual crowding: a fundamental limit on conscious perception and object recognition. Trends in Cognitive Sciences, 15(4), 160–8.

Wichmann, F. A., & Hill, N. J. (2001a). The psychometric function: I. Fitting, sampling, and goodness of fit. Perception & Psychophysics, 63(8), 1293–1313.

Wichmann, F. A., & Hill, N. J. J. (2001b). The psychometric function: II. Bootstrap-based confidence intervals and sampling. Perception & Psychophysics, 63(8), 1314–1329.

Winner, E., & Drake, J. E. (1996). The rage to master: The decisive role of talent in the visual arts. In The road to excellence: The acquisition of expert performance in the arts and sciences, sports and games (pp. 271–301). Hillsdale, NJ: Erlbaum.

Wolfe, J. M., & Horowitz, T. S. (2004). What attributes guide the deployment of visual attention and how do they do it? Group, 5(June), 1–7.

185

ERROR! USE THE HOME TAB TO APPLY TITRE 1 TO THE TEXT THAT YOU WANT TO APPEAR HERE. | Error! Use the Home tab to apply Titre 2 to the text that you want to appear here.

Zesiger, P., Martory, M., & Mayer, E. (1997). Writing without Graphic Motor Patterns : A Case of Dysgraphia for Letters and Digits Spar ing Shor thand Writing. Psychology, 14(5), 743–763.

Zhang, W., & Luck, S. J. (2008). Discrete fixed-resolution representations in visual working memory. Nature, 453(7192), 233–5.

186