Society for the Psychology ofMIst^^ (2)

-•'-.'. :... '.•"•'• VExperlmeifil| i^^

Edward Hopper, Cobb s barn 'dim' [distant houses, 1931

pPf^£&73

"**%m&****

\Wfikfiik§*i

T&&*

«u, n r w * ^ *

5Rll?

> * • • & •

S*j*4*'

BP^P^^^S^^^wwl

Edward Hopper. Cobbs barn and distant houses, 1930-33 See introduction for explanation

Whitney Museum

Bulletin of Psychology and the Arts Vol 4 (2)

Contents Experimental aesthetics: The state of the art - Paul J. Locher, Guest Editor

55 Introduction

Paul J. Locher

56 Prospects for a Cognitive Neuroscience of Visual Aesthetics

Anjan Chatterjee

61 Music and Neuroimaging: Technical Aspects

Julian Paul Keenan, Jennifer Romanowski, William Chistiana, and Gottfried Schlaug

66 How do Viewers look at Artworks? Calvin F. Nodine and Elizabeth A. Krupinski

70 Three Influences on the Eye of the Beholder: Exposure, Conditioning and Contrast Debra A. Zellner and Scott Parker

74 A Technique for Quantifying the Contribution of Pictorial Balance to the Creation of visual displays Paul J. Locher

78 The Role of Symmetry in the Expression of Valence, Arousal Balance, and in Interaction Urgency Design Stephan Wensveen and Kees Overbeeke

81 "Origins of Impressionism": Relating Visitor Behavior to Perceived Learning Jeffrey K, Smith and Lisa F. Smith

86 A Cross-Cultural Analysis of the Visual Rightness Theory of PicturePerception Leonor Lega, Luciane Paula-Pereira, Diana Giron, S ilvia Pastor, Paul J. Locher, and Graciela Hoyos

Div. 10 News 90 Message from the president

Diana Deutsch 90 Message from the president elect

Paul Locher 90 Message from the past president

Jerome L. Singer 91 Proposed Amendments to Division 10 Bylaws 91 International Congress on Aesthetics, Creativity, and Psychology

of the Arts

Vol 4 (2)

Inlroduclioii

Paul J. Lochcr Monlclair State University

It is 30 years since D. E. Bcrlyne (1974) published his now classic Stud-ies in the new experimental aesthetics: Steps toward an objective psychol-ogy of aesthetic appreciation. According to Berlyne, the main purpose of the work was "to give some idea of the techniques that are now available to the experimental aesthetician and of the kinds of questions that can and cannot be answered with their help" (p. vii). Those of us working in the field of experimental aesthetics are now able, due to advances in computer technology and neuroimaging techniques, to subject theoretical notions about the nature of an aesthetic episode to investigation with technological sophistication only dreamt of in the early seventies. Additionally, methods of inquiry described in this Issue for the study of the design of manufac-tured objects, museum visitor behaviors, and cross-cultural aspects of pic-ture perception have also become increasing sophisticated since the publi-cation of Berlyne's book. This Issue presents a sample of different contem-porary experimental approaches to the study of the forms of behavior that center around works of art and other acslhctic phenomena. As such, the collection of papers contained herein provides an overview of the "new experimental aesthetics."

The first two articles arc written by researchers interested in the poten-tial of ncuorimaging to gain an understanding of aesthetics. Anjan Chatlcrjec outlines the ways neuroimaging techniques arc starting to provide (he neu-ral basis for visual aesthetics. He presents a framework, adapted from the cognit ive ncuroscicncc of v is ion, from which hypotheses about neuroaesthctics might be tested. Next Julian Kccnan and his colleagues describe the potential of both structural Magnetic Resonance Imaging (sMRI) and function Magnetic Resonance Imaging (fMRI) to investigate brain-music relationships. Their paper provides an excellent description of the strengths and limitations of the various brain mapping techniques avail-able to investigate musical processing and other forms of aesthetic experi-ence.

The study of eye movements has been used as a tool by researchers for many decades to reveal covert perceptual and cognitive process that under-lie the perception of artworks. Pioneers in this field worked with eye-move-ment recording systems that lacked the precision of apparatus now in use. Modern systems, run by sophisticated computer programs that support their operation, enable one to record and analyze with a high degree of accuracy the aesthetic exploratory behaviors used by viewers to gather information from visual displays. Shown on the cover of this Issue are reproductions of two works by Edward Hopper. Superimposed on each picture is the record of the eye fixation pattern of a viewer whose task it was to examine the work for 15 s and then provide a title for the composition. Each viewer's fixation pattern consists of circles and lines. Each circle indicates the loca-tion of a fixation and its size indicates the length of time the observer directed his/her attention to the pictorial elements at that location in the pictorial field. The lines connecting the circles reflect the transport eye movements used by the observer to navigate from one area of interest in the pictorial field to the next. The two scanning patterns shown were gen-erated specifically for the cover of this Issue. They are presented as illus-trations of the way differences in the structural organization of two compo-sitions depicting different perspectives of the same landscape scene influ-ence the way observers look at them. In the third article of this Issue, Calvin Nodine and Elizabeth Krupinski describe the findings of early researchers who used eye-movement recording techniques to investigate picture per-ception. They then describe several recent experimental themes in the study of pictorial compositional design and discuss recent findings that support theoretical writings about the way people look at and react to paintings.

Debra Zellner and Scott Parker's paper addresses the question "What governs someone's assessment of how pleasant or unpleasant a stimulus is? They discuss the ways three processes - exposure, conditioning, and contrast - influence people's aesthetic and hedonic judgments of food stimuli, visual art, and music. The authors describe the problems for estab-lishing a psychological approach to aesthetics rooted simply in the proper-ties of aesthetic objects and in the personalities and discernment of those experiencing them.

Historically, attempts to document the role of balance as an important design principle of pictorial composition have determined its presence m i l 54

Bulletin of Psychology and the Arts paintings and other artworks qualitatively by examination of their struc-tural organization or "geometry." My paper describes a technique that pro-vides an ongoing record of the compositional strategy used by an artist to create a visual display from which a quantitative account of the strategy is generated. This system permits one to determine at any time during the creative process the location of the balancing center of a design, the distri-bution of physical (structural) weights about its axes, and the local and global dynamic qualities of its structural organization. Quantitative mea-sures of these design properties provide the types of "physical evidence" that Arnheim (1988) and others assume underlie their philosophical writ-ings on balance as the primary design principle of pictorial composition.

Stcphan Wensvccn and Kccs Ovcrbcckc's paper deals with the new field of produci design known as "interaction design." They explain that prod-ucts have become internetivc; they allow users to interact with the digital world instead of acting in the physical world. Their functionality has to be accessed not by doing things with fl product but by doing things to an inter-face. This intcrnction now makes possible the design of emotionally intel-ligent products, that is, products that respond to the emotional state of its user. Wensvcon and Overbooks describe their research that went into the development of an interactive nlorm clock'that is able to categorize the mood o user is in by analyzing the woy he/she sets the wake-up time with a set of 12 sliders. Of particular relevance to the focus of the present Issue is their finding of a strong relationship between the emotional slate of a user and parameters that describe the aesthetics of the slider pattern, for ex-ample, the symmetry and balance of the slider pattern created by a user.

How do people learn in art museums? This is the question addressed by Jeff and Lisa Smith. They studied the expectations, reactions, behaviors, and perceived learning of museum goers while they visited a special exhi-bition at The Metropolitan Museum of Art in New York City. The Smiths discuss the challenges faced by researchers who try to conduct rigorous research in non-controlled settings like museums and other cultural institu-tions.

In the last paper included in this Issue Leonor Lega and her colleagues report the way they have subjected the visual Tightness theory of pictorial composition to experimental scrutiny with respect to the influence of cul-tural factors on people's aesthetic preferences for paintings deemed "visu-ally right" (i.e., having "good" compositional organization). Their art stimuli consist of reproductions of paintings by renowned artists in the Western canon and computer-generated structurally less we 11-organized versions of them. Few earlier studies of cultural differences in aesthetic preference have utilized experimentally manipulated complex visual displays like those used by these researchers.

In closing, I wish to thank my colleagues for their excellent contribu-tions to this Issue. This collection of papers provides a glimpse of the rich-ness of work being carried out by those who would consider themselves experimental aestheticians following in the Berlyne tradition.

References Arnheim, R. (1988). The power of the center. Berkeley, CA: University of

California Press. Berlyne, D. E. (1974). Studies in the new experimental aesthetics: Steps to-

ward an objective psychology of aesthetic appreciation. New York, NY: John Wiley and Sons.

Bulletin of Psychology and the Arts

Prospects for a Cognitive Neuroscience of Visual Aesthetics

Anjan Chatterjee University of Pennsylvania

Abstract The neural basis for visual aesthetics is largely unknown. Yet, murmurings

within cognitive neuroscience suggest this wi l l soon change. In this re-view, I suggest several ways in which cognitive neuroscience might con-tribute to studies in aesthetics. First, I present a framework, adapted from the cognitive neuroscience of vision, from which hypotheses about neuro-aesthetics might be tested. Following that, I outline several ideas advanced by prominent neuroscientists that are provocative, but in need of experi-mental testing. Then I point to the effects of brain damage on artists, as contributing to our understanding of the neural bases of artistic produc-tion. Finally, I mention recent functional neuroimaging studies that are rel-evant to aesthetic concerns. These studies examine the neural response to beautiful faces and its relationship to affective systems within the brain. While it is too early to be sure, programmatic research in the cognitive neuroscience of aesthetics promises rich rewards by bringing new ways to approach empirical aesthetics. However, much work remains to be done.

Prospects for a Cognitive Neuroscience of Visual Aesthetics

The cognitive neuroscience of aesthetics is in its infancy. Despite a re-cent spate of writings on art by neuroscientists, relatively little empirical work has been conducted. There are no central tenets from which cognitive neuroscientists can draw inspiration. With rare exceptions, it is not even clear that neuroscientists consider aesthetics a proper domain of inquiry. (Some aestheticians probably consider neuroscientific inquiry into aesthetics an abomination.) This is the "state of the art" vis-a-vis the cognitive neuro-science of aesthetics. Yet, there are reasons to be optimistic about pros-pects for a cognitive neuroscience of aesthetics.

In this review, I start with a general framework from visual cognitive neuroscience that is relevant to visual aesthetics. This framework provides a context within which empirical questions might be asked and hypotheses about neuro-aesthetics might be tested. Following the framework I wi l l outline recent ideas advanced by prominent neuroscientists that are pro-vocative, but need experimental grounding. Next, I touch on ways in which the neuropsychology of visual artists contributes to our understanding of artistic production. I wi l l then mention several neuroimaging experiments that are relevant to neuro-aesthetics, before making some concluding com-ments.

A Framework from Visual Cognitive Neuroscience The process by which humans visually recognize objects offers a frame-

work from which to consider visual aesthetics. Such a framework, adapted from cognitive neuroscience, rests on two assumptions (Chatterjee, 2002a). First, visual aesthetics, like vision, has multiple components. Second, an aesthetic experience is not a response to a single component. Rather, it is derived from responses to different components of a visual object. Investi-gations can be focused on any of these components or on their combinato-rial properties.

Beyond perceptual and cognitive aspects of visual aesthetics are Ihe emotional ones. The aesthetic experience has long been described as one of "disinterested interest," or one in which the viewer experiences pleasure without obvious utilitarian consequences of this pleasure. This experience contrasts with those to other visual objects that might give pleasure by appealing to basic drives such as the desire for food or sex. Aesthetic ob-jects presumably give pleasure without evoking additional desires, although the boundaries between the two emotional experiences may not always be clear (Santayana, 1896). The process by which humans react to stimuli and engage neural circuits that respond to pleasing or rewarding stimuli may offer a probe into the neural basis for " l iking without wanting."

The nervous system processes visual information both hierarchically and in parallel (Farah, 2000; Van Essen, Feleman, DeYoe, Ollavaria, & Knierman, 1990; Zeki, 1993). The levels of this processing can be classi-fied as early, intermediate and late vision (Marr, 1982). Early vision ex-tracts simple components from the visual environment, such as color, lu-minance, shape, motion and location (Livingstone & Hubel, 1987, 1988). These components are processed in different parts of the brain. Intermedi-ate vision segregates some components and groups others together to form coherent regions in what would otherwise be a chaotic and overwhelm-#5 c ing sensory array (Biederman & Cooper, 1991; Grossberg, Mingolla, &

Vol 4 (2)

Ros, 1997; Sajda & Finkel, 1995). Intermediate vision proceeds automati-cally, seemingly without effort. Late vision selects which coherent regions to scrutinize. It also evokes memories from which objects are recognized and meanings attached. In the aftermath of object recognition, emotions may be evoked and decisions about the objects (e.g., I wi l l eat it., or I wil l grasp i t ) can be made.

Individuals with brain damage may have selective deficits in any of these levels of visual processing. Thus, damage to parts of the occipital cortex can produce complete blindness or blindness for sectors of the visual field or blindness for selective elements of early vision, such as color or form or movement. Damage to intermediate vision results in an inability to group visual elements (Vecera & Behrmann, 1997). Such a person is overwhelmed by the experience of confronting a visual environment that cannot be orga-nized (Ricci, Vaishnavi, & Chatterjee, 1999). Damage to late vision takes varied forms. In some disorders, such as spatial neglect, some visual ob-jects are simply not selected for further processing because of their spatial location (Chatterjee, 2003). In others, such as associative visual agnosias, even when the visual percept of the object is discerned, the patient does not recognize the objeel (Farah, 2000). The percept is "stripped of meaning" as it no longer makes contact with the person's memories attached to that percept. This evidence for different levels of visual processing from human neuropsychological studies is corroborated by neurophysiologic studies.

How is this view of levels of visual processing reflected in empirical aesthetics? Thc'distinction between form and content common in aesthetic writings (e.g., Russell & George, 1990; Woods, 1991) is paralleled by the observations that early and intermediate vision process form and later v i-sion processes content. Figure 1 shows a working model of how the

ht Stimuli

--»

*"-

Early vision, Features,

(orientation, shape, color)

^

—*•

/ Intermediate

vision (grouping)

'

\ \ 1 Representational domain

(places, faces)

Figure 1. A general framework for the neurai underpinnings of visual aesthetics guided by visual neuroscience. neuroscience of visual aesthetics might be mapped. The early feature of an art object might be its color, which would be processed in parts of the occipital cortex, such as areas V4. Early features would be grouped to-gether to form larger visual units by intermediate vision. The neural basis of grouping is not well understood, but it likely involves extra-striatc cor-tex (Biederman & Cooper, 1991; Grossberg et al., 1997; Sajda & Finkel, 1995).

The process of grouping gives "unity in diversity," a fundamental fea-ture of compositional balance, which itself is a central idea about the for-mal structure of aesthetic objects. I f compositional form is apprehended automatically by intermediate vision, then sensitivity to such form should also be automatic. Subjects are sensitive to form "at a glance" wilh expo-sures as short as 50 msec (Locher & Nagy, 1996). Furthermore, preference for form predominates when images are shown over short exposure times. By contrast, preference for detail (which requires serial attcntionat pro-cessing) predominates when images are shown for slightly longer limes (Ognjenovic, 1991).

As a tentative proposal, I suggest the following sequence of processing within the nervous system. The visual attributes of art work initially arc processed like any other visual object. Various combinations of early and intermediate visual properties (e.g., color, shape, composition), especially i f they are balanced, engage frontal-pari eta I altcntional circuits. These attentional networks continue to modulate processing within the ventral visual stream (Humphreys, Riddoch, & Price, 1997; Moller, 1993, 1994; Pessoa, Kastner, & Ungerleider, 2003; Shulman cl al., 1997; Walanabc et al., 1998). This further modulation likely contributes to a more vivid expe-rience of the stimuli, both its attributes, such as colors or form as well as its content, such as faces or landscapes. Thus, a feed forward system is estab-lished in which the attributes of an aeslhctic object engage attention, and

attention further enhances the processing of these attributes. Aesthetic objects are probably one example of a class of objects that engage such a

feed forward system. Any objeel that is experienced vividly, such as an

Vol 4 (2)

emotional face, is probably mediated similarly. These other objects are dis-tinguished from aesthetic objects by the structural properties of the object itself and the subsequent emotional responses to them.

The emotional experience of viewing an aesthetic object is central to the aesthetic experience. Considerable evidence from neuroscience suggest that the anterior medial temporal lobe, medial and orbitofrontal cortices and subcortical structures mediate emotions in general, and reward systems in particular (Breiter, Aharon, Kahneman, Dale, & Shizgal, 2001; Delgado, Nystrom, Fissell, Nol l , & Fiez, 2000; Elliott, Friston, & Dolan, 2000; O'Doherty, Kringelbach, Rolls, Hornack, & Andrews, 2001; Schultz, Dayans, & Montague, 1997). Recently, the distinction between " l ik ing" and "wanting" has gained prominence in investigations of the neural bases of emotions (Berridge, 1996; Wyvell & Berridge, 2000). Wanting or the desire for rewards seems to be mediated by dopaminergic circuits includ-ing the ventral striatum and the nucleus accumbens. There is a long-held notion that a quintessential aesthetic experience involves disinterested in-terest. That is, the object gives pleasure but not for utilitarian reasons. The neural mediation of such an emotional experience, of liking without want-ing, has not been worked out. (Judgments about an aesthetic object might be considered outside the core aesthetic experience following the logic of disinterested interest. However, they feature prominently in experimental aesthetics in which subjects state preferences or make decisions aboul ob-jects. These judgments are likely to engage widely distributed circuits, most importantly the dorsolateral frontal and medial frontal cortices.)

Cognitive neuroscience, for the most part, focuses on properties of the nervous system that are shared rather than those that vary. What arc wc to make of the amazing diversity of art from different cultures and time peri-ods? Can one sort universal aspects of ncuroacslhctics from cultural ones? The analogous question of distinguishing universal from cultural aspects might be asked of vision. Is vision universal or relative? Early and inter-mediate vision are almost certainly universal. For example, individuals respond to the same light frequencies to perceive colors and group visual elements similarly. These processes are presumably "hardwired" in the brain. No amount of experience or cultural exposure will make a human perceive infrared light. Similarly, grouping mechanisms, such as seeing illusory contours, are likely to be obligatory. Later vision is different. Although the process of selecting and recognizing visual objects (linking the percept to memories and meanings) is likely to be universal, the specific memories and meanings evoked by visual percepts vary For example, what is the

Figure 2. A visual object that would be difficult to recognize by most viewers. This is an example of a percept stripped of meaning, object in Figure 2? While the percept may be clear, without some knowl-edge or memory of the class of which this object is a member, recogni-{ tion is at best a guess. Across cultures, individuals with similar brain

Bulletin of Psychology and the Arts

damage are likely to have object recognition deficits. However, their cul-ture affects which objects they do or do not recognize.

As in object recognition, aesthetic objects are likely to evoke both uni-versal and relative responses to different components of the object. Re-sponses to early and intermediate visual components of an art object are likely to be universal. Thus probes for preferences for early visual ele-ments, such as color (Eysenck, 1941) or compositional form (Goetz, Borisy, Lynn, & Eysenck, 1979), which unifies diverse elements, are likely to be universal. By contrast, probes for preferences for aspects of later vision, such as the content of an image are likely to be relative (Gombnch, 1960).

Recent Comments on A r t by Cognitive Neuroscientists Recently, the relationship between the biology and beauty of vision has

captured the imagination of prominent vision neuroscientists. Zeki (1999a, b) argues that no theory of aesthetics is complete without an understanding of its neural bases. He suggests an important parallel between the goals of the nervous system and that of artists. Both are driven to understand essen-tial attributes of the world. The nervous system decomposes visual infor-mation received through the retina into different components, such as color, luminance and motion. Similarly, many artists, particularly within the last century have honed in on different visual attributes, such as color by Matisse or motion by Caldcr. Zeki suggests that artists, like visual neuroscientists, endeavor to discover attributes of the visual world, which correspond to processing components of the visual brain. With respect to the general frame-work described in the previous section, Zeki highlights the importance of the modular nature of early vision in its role in visual art, and the way in which elements from early vision evoke aesthetic experiences.

Livingstone (Livingstone, 2002) focuses on artists' use of interactions between the dorsal (where) and the ventral (what) visual stream. The dis-tinction between dorsal and ventral processing is central in visual cognitive neuroscience (Ungerleider & Mishkin, 1982). The dorsal stream is sensi-tive to contrast differences, motion and spatial location. The ventral stream is sensitive to simple form and color. Livingstone suggests that the shim-mering quality of water or the sun low on the horizon seen in some impres-sionistic paintings (e.g., the sun and surrounding clouds in Monet's Im-pression Sunrise) is produced by isoluminant objects distinguishable only by color. The dorsal stream is insensitive to isoluminant color differences of the image. Since the dorsal stream identifies motion (or the lack thereof) and spatial location, Livingstone argues that this is why these isoluminant forms are not fixed with respect to motion or spatial location and are expe-rienced as unstable or shimmering. Conversely, since shape can be derived from luminance differences, she argues that artists can use contrast to pro-duce shapes, leaving color for expressive, rather than descriptive purposes (as in Derain's portrait of Matisse). Like Zeki (1999b), Livingstone also focuses on features of early vision, but highlights their interactions. Artists use these interactions as part of their tool-kit to achieve specific aesthetic effects.

Ramacjiandran and Hirstein (1999) propose several perceptual principles underlie aesthetic experiences. Of this list of perceptual principles, they emphasize the "peak shift" phenomena as offering insight into the aesthet-ics of abstract art and rely on Tinbergen's (1954) work. Tinbergen demon-strated that sea gull chicks beg for food from their mothers by pecking on a red spot near the tip of the mother's beak. It turns out that a disembodied long thin stick with three red stripes near the end evokes an exaggerated response from these chicks. Ramachandran and Hirstein propose that neu-ral structures that evolved to respond to specific visual stimuli respond more vigorously (a shift in their peak response) to underlying primitives of that form even when the subject is not aware of the primitive. Their insight is that abstract art may be tapping into such visual primitives, although they are not specific about the neural mechanisms that might account for this intriguing hypothesis.

These three attempts to link visual neuroscience and aesthetics reflect the emerging recognition by neuroscientists that visual aesthetics are an important part of human visual experience. As such, visual aesthetics ought to conform to principles of neural organization.

The Neuropsychology of Visual Artists I f visual aesthetics conform to principles of neural organization, what

happens to artists with brain damage? Several factors make it difficult to know what to make of the art produced by brain-damaged individuals. The data are observational rather than experimental. Artists vary in their tal-ents, raising questions of whether general principles can be extracted from

something that is already so variable across individuals. Beyond the ques-c£ n tion o f talent lies the problem o f artistic styles and content which may

change with brain damage. When considering different artistic tradi-

Bulletin of Psychology and the Arts

tions, is the same kind of behavior even being scrutinized? Despite these limitations, I suggest that the neuropsychology of artists contributes to the empirical studies of the arts. I have developed this thesis in detail else-where (Chatterjee, submitted). Here, I focus on three questions. Because of specialized visuo-motor skills, are artists spared the kind of visual defi-cits seen after brain damage in other people? Does brain damage alter art-ists' styles? Can brain-damaged artists contribute to our understanding of the roles of "knowing" and of "seeing" in producing art?

Artists with brain damage that affects their visual system do not appear to be spared deficits experienced by others. Rather, because of their skills they often express their deficits with particular eloquence. Sacks (1995) described an artist with achromatopsia, a selective loss of color perception. Before the traumatic brain injury, which produced this deficit, the artist painted colorful abstractions. Following his brain injury, everything ap-peared "dirty gray" to him. Initially he applied color in a haphazard man-ner, before resigning himself to black and white renderings. Eventually he reintroduced color to his paintings, but with an extremely limited palette. Thus, this patient's achromatopsia changed how he painted, bul did not prevent him from continuing to be a successful artist.

Unilateral spatial neglect is a disorder in which individuals appear to be unaware of objects or images in space contralateral lo their brain lesion (Chatterjee, 2003). This disorder is more common following right than left brain damage. Several successful artists, including Lovis Corinth, Anton Radersheidt, Reynolds Brown and Loring Hughes all developed neglect following right brain damage (Heller, 1994; Jung, 1974; Schnider, Regard, Benson, & Landis, 1993). These artists demonstrate dramatic contralesional neglect in their paintings, with minimal brush strokes and detail in contralesional space. The great Italian f i lm director, Fellini, developed left neglect following a right hemisphere stroke (Cantagallo & Sala, 1998). Fellini was also an accomplished cartoonist, and his spontaneous cartoons demonstrated left-sided neglect (Figure 3). Some patients with neglect can

y * t —

*£

Figure 3. An example of Fellini's spontaneous cartoons when asked to bisect lines. As can be seen he drew his adornments on the right of the line even as he mis-bisected the line to the right of true center. Reprinted with permission (Cantagallo & Sala, 1998). process left sided stimuli to some degree short of full awareness. Fellini demonstrated such partial awareness in some of his drawings. For example, he drew a cartoon (Figure 4) of an elephant on the left asking the question of a doctor behind a desk on the right, "Can I stay here, doctor?"

Vol 4 (2)

Visual agnosias are disorders in which patients are unable to recognize objects visually. Lissauer (1890), at the end of the nineteenth century made the classic distinction between apperceptive and associative agnosias, to highlight the fact that a patient might not identify an object because of a deficit anywhere along a perceptual to conceptual continuum. Appercep-tive agnosics have deficits that lie closer to perceptual processing (De Renzi & Lucchelli, 1993; Warrington & James, 1988). Whereas associative agnosics deficits lie closer to conceptual knowledge of objects (Riddoch & Humphreys, 2003). The few artists with visual agnosias who have been described demonstrate problems of either apperception or association in their art. Wapner and colleagues (Wapner, Judd, & Gardner, 1978) described an artist, JR. with an apperceptive agnosia. JR's drawings after his stroke retained some of his premorbid techniques, such as the use of shadows and perspective. However, he often lost his place in the middle of drawings. Because he could not recognize his own drawing, he often omitted some details and elaboraied others. Sometimes he redrew features, such as five legged rhinocerous or a plane with many propellers. At times he would use a conceptual strategy to guide his drawings. For example, in drawing a telephone, he would say "It needs a base for it to stand on, a place to speak into, something to hear with a wire to plug in for communication and place to dial." This strategy was not particularly effective. His piecemeal vision was accompanied by a fragmented approach to his drawings. He drew fea-tures accurately, but did not fit these features into an overall composition.

JR's deficits contrast with those of two other artists, who had similar associative agnosias (Franklin, van Sommers, & Howard, 1992; Schwartz & Chawluck, 1990). The quality of their drawings varied dramatically de-pending on the context in which they drew. When presented with a rich visual model, such as a picture or actual person, their drawings remained remarkably skilled and beautiful (see Figure 5). However, associating a

Figure 4. A cartoon drawn by Fellini showing an elephant on the left asking if it can stay there. Reprinted with permission (Cantagallo & Sala, 1998)

Figure 5. A drawing copied from an example by a patient with an associative visual agnosia. Reprinted with permission (Franklin et al., 1992). word to a visual form proved to be very difficult, and this difficulty was expressed in their art. When given names of objects to draw, their drawings became simple, crude and sometimes unrecognizable.

While artists are not spared the kinds of deficits experienced by others following brain damage, sometimes the changes in their artwork can be surprisingly appealing. Brain damage that renders accurate depiction diff i-cult can sometimes free artists to be more expressive. The artist Loring Hughes had trouble with spatial relationships of lines after she suffered right brain damage. Heller (1994) described Hughes' experience as fol-lows: ".. .after damage to her brain.. .two things contributed to a new found freedom from inhibition. First, her difficulties with spatial relationships made it impossible for her to replicate the world on paper, and this forced her to accept and explore the world of her own emotions...she gave up trying to reproduce things and turned, instead to her imagination.... To her surprise, when she did start showing, the feedback from the artistic com-

munity was much more encouraging than before. Her paintings 'deliver ! an emotional wallop' according to art critic Eileen Watkins." Similarly

Vol 4 (2)

Reynolds Brown and Lovis Corinth were judged to be more expressive in their paintings after their right brain damage. They used bolder strokes and less detail The critic Alfred Kuhn (quoted in Gardner, p. 23, 1975) charac-terized Corinth after his brain damage as having "shifted from the ranks of the great painters into the circle of great artists."

DeKooning is perhaps the best-known artist whose style changed fol-lowing neurologic disease (Garrels, 1995; Storr, 1995). He developed Alzheimer's disease, probably in the late 1970s. With the help and support of his ex-wife and assistants, he continued lo paint until 1988. There is general agreement among experts thai this late period, exhibited in San Fransisco in 1996, constituted a new and coherent style. These paintings were abstract and successively simpler, using moslly primary colors such as reds and blues on white. Gary Garrels (1995), senior curator at the San Francisco Museum of Modern Art comments: "In the 1980s works, the essential procedures and techniques were not changed, but simplified, and the vocabulary of forms was retained but clarified. Particularly in the works from 1984, the results are paintings of an openness and freedom not seen before, paintings that arc extraordinarily lyrical, immediately sensual, and exhilarating."

Do artists produce what they apprehend directly, or do they produce what they think is true of the world? Gombrich (1960) underscores the importance of this question in the history of Western art when he explores reasons why styles of representation have changed dramatically over the years. He suggests that artist bring tremendous top-down information to bear on their perceptions. Sensations trigger hypotheses about what it is they arc seeing. In turn, the hypotheses are based on internal representa-tions. Consequently, artists are more often aware of their own internal rep-resentations than they are of their direct sensations. An exaggerated ex-ample of this point is evident in the drawings of children. Children draw roads receding into the horizon with parallel lines rather than lines con-verging into the vanishing point because they know that the sides of the road are parallel and do not meet in the distance. This perceptual hypoth-esis testing occurs automatically, and artists need considerable practice to better "sec" the world. What happens when artists have impoverished in-ternal representations? Germane to this question are observations of autis-tic children with exceptional drawing skills.

About 10% of autistic children have islands of exceptional skills (Rimland & Fein, 1988). Some of these skilled children are gifted visual artists. Selfe (1977) reported the first detailed description of such a child, Nadia. Nadia had several developmental abnormalities. She did not respond to her mother and she lacked expressions of social sympathy. As she got older, her rela-tionships with other children seemed like an obsessive concern for their presence, rather than consisting of any substantial interpersonal interac-tion. Her language development was delayed and her speech was frequently echolalic and ritualistic. Despite these abnormalities, she was amazingly skilled at drawing. At three and a half years of age she was drawing re-markably life-like horses in perspective (see Figures 6 and 7). Unlike other children, she did not go through a stage of drawing simplified schematic

Bullet in of Psychology and the Ar ts

Figure 6. Drawing by the autistic child, Nadia, when she was 3 and a half years old. Reprinted with permission (Selfe, 1977). images of horses before drawing them more realistically. Her skills were* highly developed at the outset, and although the objects she drew changed<

Figure 7. A typical drawing of a person on a horse that by a six year old child. Reprinted with permission (Selfe, 1977). over time, her abilities themselves did not improve substantially. Her move-ments when drawing were deft, rapid and without hesitation. She drew intensively for a few minutes at a time and produced line drawings without any interest in color. She tended to draw specific subjects like horses and riders and she seemed oblivious to the boundaries of the page. Her draw-ings were modeled after other drawings. However, she did not copy the models. Rather, her drawings were often composites of previous images she had seen, and sometimes had subtle shifts in their orientation. She never looked at the original once she started to draw.

While Nadia was exceptional, she is not unique. Other such autistic art-ists have been reported. Thus, in the setting of impoverished conceptual development, but preserved or superior visuo-motor skills, children can produce remarkable drawings with depth, rather than the simplified sym-bolic drawing produced by most children. Consistent with the view that Nadia drew what she saw rather than what she knew, she would sometimes begin her drawings at bizarre points, such as at the neck of a horse and not its head where most people start. I f language development is a marker for conceptual development, then it is especially intriguing that Nadia's draw-ing skills became prosaic as she acquired more language.

It is worth repeating that inferences made from the neuropsychology of artists must be made with considerable caution. Nonetheless, these examples of the dramatic phenomenology seen in artists with brain damage show that such patients contribute to our understanding of visual aesthetics.

Functional Neuroimaging and Visual Aesthetics: The Example of Beautiful Faces

Functional magnetic resonance imaging (fMRI) techniques are likely to advance our understanding of the neural underpinnings of perceptual and emotional aspects of aesthetic experiences. fMRI uses changes in blood flow in response to stimuli or cognitive processes to make inferences about the relevant underlying neural activity. The inferential logic in fMRI stud-ies can be complicated (Aguirre & D'Esposito, 1999). I f one knows the components of a cognitive operation which can be isolated, then one might test hypotheses about which parts of the brain are engaged by this compo-

nent. Alternatively, i f one has a clear sense of the function of a particular region of the brain, then one can test the hypothesis that a cognitive pro-cess includes that function. Domains of inquiry that are relatively early

Bulletin of Psychology and the Arts

in their evolution, such as neuro-aesthetics, pose challenges to testing hy-potheses using fMRI. Currently, the specific components of a process (the aesthetic experience) may not be isolated and knowledge of relevant brain functions might not be sufficiently worked out.

Despite these limitations, a few recent fMRI studies relevant to the cog-nitive neuroscience of beauty, particularly in the context of faces, have been reported. Faces are a category of visual objects with special saliency within the nervous system. Faces convey tremendous amounts of informa-tion, such as identity, emotion, personal and cultural attributes and so on. Patients with focal brain damage, usually to the occipital-temporal junc-tion can have a disorder called prosopagnosia, a selective inability to rec-ognize faces (Young, Newcombe, de Haan, Small, & Hay, 1993). fMRI studies have shown that in most normal individuals a region within the fusiform gyrus is more responsive when subjects look at faces than when they look at other objects such as houses (Allison, McCarthy, Nobre, Puce, & Belger, 1994; Kanwisher, McDermott, & Chun, 1997). The importance o f faces for humans is also reflected in art. There is a long history of por-traits being used to communicate information about the individual depicted (Chatterjee, 2002b).

The fMRI studies in facial attractiveness focus on the difference between an aesthetic judgment of the beauty in faces and the degree to which they are desirable. Aharon and colleagues (2001) studied young heterosexual men's responses to attractive and average male and female faces. These subjects also indicated how long they wished to view a face by repeated key-presses. This procedure allowed the investigators to distinguish be-tween an aesthetic judgment and the desire to keep looking at the face, because they found a dissociation of attractiveness rating and key press across gender. These young men chose to view attractive women longer than average looking women, but did not choose to view attractive men longer than average looking men or average looking women. The main finding in the study was that the desire to keep looking at faces correlated with activity within the nucleus accumbens and closely related subcortical structures. These structures are postulated to be part of a more general reward circuitry, associated with monetary, drug and homeostatic rewards and thought to be mediated by dopaminergic systems. While the neural structures activated by aesthetic judgments per se were not identified, the authors suggest that judgments of attractiveness and the pleasure derived from attractive objects might be neurally dissociable.

In a related study, Kampe, Frith, Dolan, and Frith (2001) found that direction that attractive faces were gazing also modulated activity within reward circuits, close to the nucleus accumbens. Again, they did not find neural effects of the perception of attractiveness per se, but they did find attractiveness when paired with eye contact activated the ventral striatum. They postulate that in the setting of social contact, the return of eye-gaze from an attractive face is rewarding, and consequently activity of dopamin-ergic reward systems are enhanced. By contrast, O'Doherty and colleagues (2003) found that attractive faces activated the medial orbito frontal gyrus even when subjects made an unrelated judgment of the faces, in this case gender. They also found that this activity was further enhanced i f the face was smiling. Unlike the two previous studies, they did not find significant activity within the nucleus accumbens or the ventral striatum. They specu-late that the orbitofrontal activation reflects the direct affective response to beautiful faces and not expectations of rewards that might be mediated by the circuits involving the nucleus accumbens and the ventral striatum.

These studies are cited as an example of how functional neuroimaging might contribute to empirical studies in aesthetics. There is something para-doxical about them, with respect to aesthetics. These studies use attractive faces as a probe to understanding the neural bases for rewards. Such stud-ies are critical to understanding drives in normal and pathological states. The focus is usually on wanting rather than liking. In some conditions, such as in drug addiction, one might find wanting without liking. It is pre-cisely the converse of this, liking without wanting, if such circuitry could be delineated, that would be of most interest to empirical aesthetics.

Conclusions The prospects for a cognitive neuroscience of aesthetics is promising.

Adapting ideas from visual neuroscience may help guide testing of hypoth-esis about the perceptual, cognitive and affective aspects of neuro-aesthet-ics. Intriguing ideas proposed by prominent neuroscientists are poised to be tested empirically. Insights into artistic production can be gained by observations of artists with brain damage. Hopefully, new observations wi l l continue to educate us in this regard. These observations need to be ac companied by experimental work in patients with brain damage. Patient^

Vol 4 (2) Kanwisher, N. , McDermott, J., & Chun, M. (1997). The fusiform face area; a

module in human extrastriate cortex specialized for perception o f faces. Jour-na l o f Neuroscience, 17,4302-4311.

Lissauer, H. (1890). Ein Fall von Seelenblindheit Nebst Einem Beitrage zur Theori derselben Archiv f u r Psychiatric und Nervenkrankheiten, 21,222-270.

Livingstone, M, (2002). Vision and ar t : The biology o f seeing. New York: Abrams.

Livingstone, M. , & Hubel, D. H. (1987). Psychophysical evidence for sepa-rate channels for the perception of form, color, movement, and depth. The Jour-nal o f Neuroscience, 7, 3416-3468.

Livingstone, M. & Hubel, D. (1988). Segregation of form, colour, movement, and depth; Anatomy, physiology, and perception. Science, 240, 740-749.

Locher, P., & Nagy, Y. (1996). Vision spontaneously establishes the percept o f pictorial balance. Empir ical Studies o f the Arts, 14( 1), 17-31

Marr, D. (1982). Vision. A Computational Investigation into the Human Rep-resentation and Processing o f Visual Information. New York; WH Freeman and Company.

Motter, B. C. (1993). Focal attention produces spatially selective processing in visual cortical areas V I , V2 ; and V4 in the presence of competing stimuli Journal o f Neurophysiology, 70, 909-919.

Motter, B. C. (1994). Neural correlates ol 'al tcnt ivc selection for color or luminance in extrastriate area V4. . / . Neurosci., 14, 2178-2189.

O'Doherty/J., Kringelbach, M „ Rolls, E., Hornack, J., & Andrews, C. (2001). Abstract reward and punishment representations in the human orbitofrontal cortex. Nature Neuroscience, 4, 95-102.

O'Doherty, J., Winston, J., Crilchley, H., Perret, D., Burt, D., & Dolan, R (2003). Beauty in a smile: the role o f orbitofrontal cortex in facial attractive-ness. Neuropsychologia, 4 1 , 147-155.

Ognjenovic, P. (1991). Processing o f aesthetic information. Empir ica l Stud-ies o f the Arts, 9(1), 1-9.

Pessoa, L., Kastner, S., & Ungerleider, L. (2003). Neuroimaging studies o f attention: from modulation o f sensory processing to top-down control. The Journal o f Neuroscience, 23, 3990-3998.

Ramachandran, V. S., & Hirstein, H. (1999). The science o f art: a neurologi-cal theory of aesthetic experience. Journal o f Consciousness Studies. 6 ,15-51.

Ramachandran, V. S., & Hirstein, W. (1999). The science o f art: A neurologi-cal theory o f aesthetic experience. Journal o f Consciousness Studies, 6(6-1), 15-51.

Ricci , R., Vaishnavi, S., & Chatterjee, A . (1999). A deficit o f preattentive vision: experimental observations and theoretical implications. Neurocase, 5(1), 1-12.

Riddoch, M. , & Humphreys, G. (2003). Visual Agnosia. Neurol Cl in, 2 / , 501-520.

Rimland, B., & Fein, D. (1988). Special talents o f autistic savants. In L. Obler & D. Fein (Eds.), The Exceptional Bra in (pp. 474-492). New York; Guil ford.

Russell, P. A., & George, D. A. (1990). Relationships between aesthetic re-sponse scales applied to paintings. Empi r ica l Studies o f the Arts, 5(1), 15-30.

Sacks, O. (1995). The case o f the color b l ind painter. In An Anthroplogist on Mars (pp. 3-41). New York: A l f red A. Knopf, Inc.

Sajda, P., & Finkel, L. H. (1995). Intermediate-level visual representations and the construction o f surface perception. Journal o f Cognitive Neuroscience, 7,267-291.

Santayana, G. (1896). The Sense o f Beauty. Being the Outline o f Aesthetic Theory. New York: Dover Publications, Inc.

Schnider, A., Regard, M. , Benson, D. F., & Landis, T. (1993). Effect o f a right-hemisphere stroke on an artist's performance. Neuropsychiatry, Neurop-sychology, & Behavioral Neurology, 6(4), 249-255.

Schultz, W., Dayans, P., & Montague, P. (1997). A neural substrate of predic-tion and reward. Science, 275, 1593-1599.

Schwartz, M. , & Chawluck, J. (1990). Deterioration o f language in progres-sive aphasia: a case study. In M. Schwartz (Ed.), Modular Deficits in Alzheimer-Type Dementia (pp. 245-296). Cambridge, M A : The M I T Press.

Selfe, L (1977). Nadia A case o f extraordinary drawing abil i ty in an autis-tic chi ld. New York: Academic Press.

Shulman, G. L., Corbetta, M., Buckner, R. L., Raichle, M . E., Fiez, J. A., Miezin, F. M. , et al. (1997). Top-down modulation o f early sensory cortex. Cerebral Cortex, 7, 193-206.

Storr, R. (1995). A t last light. In Willem de Kooning. The Late Paintings, The 1980s (pp. 37-79). Minneapolis: San Franscisco Museum of Modern art and Walker Arts Center.

Tinbergen, N. (1954). Curious Natural ist. New York: Basic Books. Ungerleider, L. G , & Mishkin, M . (1982). Two cortical visual systems. In D.

J. Ingle, M. A. Goodale & R. J. W. Mansfield (Eds.), Analysis o f visual behav-ior (pp. 549-586). Cambridge: M I T Press.

Van Essen, D. C., Feleman, D. J., DeYoe, E. A., Ollavaria, J., & Knierman, J. (1990). Modular and hierarchical organization o f extrastriate visual cortex in the macaque monkey. Cold Springs Harbor Symp Quant Biol, 55, 679-696.

Vecera, S., & Behrmann, M . (1997). Spatial attention does not require preattentive grouping. Neuropsychology, 1 1 , 30-43.

Wapncr, W., Judd, T., & Gardner, H. (1978). Visual agnosia in an artist Cortex, 14, 343-364.

Warrington, E., & James, M . (1988). Visual apperceptive agnosia: a chnico

Bulletin of Psychology and the Arts anatomic study o f three cases. Cortex, 24, 13-32.

Watanabe, T., Sasaki, Y , Miyauchi , S., Putz, B., Fuj imaki , N. , Nielsen, M. , et al. (1998). Attention-regulated activity in human primary visual cortex. Jour-nal o f Neurophysiology, 79, 2218-2221.

Woods, W. A. (1991). Parameters o f aesthetic objects: Appl ied aesthetics. Empir ica l Studies o f the Arts, 9(2), 105-114.

Wyve l l , C , & Berridge, K. (2000). Intra-accumbens amphetamine increases the conditioned incentive salience o f sucrose reward: enhancement of reward "want ing" without enhanced " l i k i n g " or response reinforcement. Journa l o f Neuroscience, 20(8122-8130).

Young, A. W., Newcombe, F., de Haan, E. H. F., Small, M. , & Hay, D. C. (1993). Face perception after brain injury: Selective impairments affecting iden-tity and expression. Brain, 116, 941-959.

Zeki , S. (1993). A Vision o f the Bra in. Oxford: Blackwell Scientific Publica-tions.

Zeki , S. (1999a). Art and the Brain. Journal o f Consciousness Studies, 6, 76-96

Zeki , S (1999b). Inner Vision1 An Exploration o f Ar t and the Bra in. New York: Oxford University Press

Anjan Chatterjee Universi ty o f Pennsylvania

Department o f Neurology 3 West Gates, 3400 Spruce Street Phi ladelphia, PA 19104. emai l : anian(ajmail,med.upcnn.edu

Mus i c and N e u r o i m a g i n g : Technical Aspects

Julian Paul Keenan, Jennifer Romanowski , and Wi l l i am Chistiana

Montc la i r State Universi ty

Got t f r ied Schlaug

Harvard Medica l School

Abs t rac t

Due to technical advances, a signi f icant number o f researchers have be-come interested in the potential o f neuroimaging to gain an understanding o f the complex brain - music interactions. An ima l models generally fa i l to answer questions specif ic to human musical processing. Both structural M R I and funct ional M R l have been employed over the last years to inves-tigate brain - music relationships. Here structural M R I and funct ional M R I are examined in terms o f more technical aspects. Employ ing both struc-tural M R I (Keenan, Thangaraj, Halpern, & Schlaug, 2001) and funct ional M R I example (Gaab, Keenan, & Schlaug, 2003), the underpinnings o f these methods are examined. The elucidat ion o f methodological l imi tat ions w i l l aid in both the understanding and construction o f neuroimaging and music - bra in relationships.

Neuro imag ing and the B r a i n Discover ing the brain correlates o f music w o u l d prov ide signif icant i n -sight into both the work ings o f the brain and the basic components o f musical processing. Further, such discoveries wou ld add to our under-

standing o f memory, language and cogni t ive categorization. Given these

Bulletin of Psychology and the Arts

implications, it would seem evident that significant attention should be given to neuroimaging and musical processing. It is the purpose of this paper to discuss how a relatively simple and important construct (i.e., lo-calizing musical processing in the brain) may not be as simple as initially imagined. These issues, though presented in terms of neuroimaging and music, likely apply to any issue involving brain localization, and in some instances, aesthetic experimentation in general, This manuscript is prepared for the person with little-to-no neuroimaging experience that may be inter-ested in interpreting the findings of others or in perhaps running his or her own research.

I (JPK) often begin my lectures on physiological psychology by point-ing out to my students that a toolbox contains 'tools' and no) 'tool' (Keenan, Gallup, & Falk, 2003). Whatever the job may be, Ihe contractor does not generally rely on one tool. He or she wil l not try to hammer with a screw-driver, nor wi l l he or she refuse to take on a job because there is a need for using multiple tools for a given job. Instead, any reasonable person wi l l use multiple tools to complete whatever work needs be done. The same holds true in the case of discovering the brain correlates of any behavior or action. It is a poor craftsman or ncuroscientist that relies on one tool or method. While both the screwdriver and the hammer are great tools, they are both only ideal in performing certain tasks. The same holds true for both structural Magnetic Resonance Imaging (sMRI) and functional Mag-netic Resonance Imaging (fMRl). No imaging technique on its own wi l l suffice in accurately delineating any brain/behavior relationship.

This issue becomes critical when one is presented with images depicting color maps of the brain from fMRI (or related) experiments. It is easy to become enamored with the 'scientific1 feel that these images have without understanding the basic methodology that undergoes the generation of such a map. These issues are critical and are often overlooked or not understood by even the most educated scientists. Examining the science behind the tools of the neuroscientists is advantageous.

There is a further issue that requires us to examine the technical aspects of neuroimaging in terms of music. Both the brain and musical processing are highly complicated and multi-leveled. Thus, any serious approach to discovering how these two interact must require a basic understanding of the instruments employed.

The brain can be conceptualized as a mobile, with an infinite number of possible states. I came upon this analogy while reading an article describ-ing Ted Williams (Updike, 1999). Williams was a baseball legend for a number of decades in Boston. However, he was injury plagued towards the end of his career. Updike, watching Williams in one of his last baseball appearances described Williams as looking like a Calder mobile with one of the threads cut. Updike was obviously trying to depict a person who was not quite right. After reading this, I thought that I might apply the mobile analogy to the brain. As one conceptualizes the brain in terms of it compli-cated interactions, the mobile analogy provides a powerful example of how we need to approach the science of neuroimaging:

With its billions of neurons and connections, the brain may for-ever elude complete and accurate analysis. But we can also view the brain like a Calder mobile, beautiful, delicate and interde-pendent in its many internal elements and connections. To de-scribe with precision the contribution of each piece of a Calder mobile is difficult even when the elements are few and the con-nections are minimal. The weights, sizes and locations of each of the pieces are intra-dependent, held together with a seem-ingly infinite combination of torques and forces. Like Calder's sculpture, the brain escapes its own form in a way that tran-scends the sum of its parts. It is fluid and free-form, yet made up of modules and networks that function in a predictable fashion. And, as with a Calder mobile, the key is balance. When consid-ering the abilities and complexities of the brain, one is struck by the incredible efficiency and splendor expressed in gray and white matter. A l l that we know and all that we are is derived from this three-pound mass, created on a design well beyond our compre-hension, let alone our ability to describe it. One either gains faith in the work of God, evolution or a bit of both. At this moment in neuroscience, a complete description of any brain-behavior re-lationship is not possible even for simple behaviors. When you see a diagram with two or three boxes with arrows used to ex-plain some brain function, be warned that the model is vague at best and errs towards oversimplification. Even with advanced techniques like CAT scans and MRIs, we can only describe gross areas under specific conditions and generate conclusions based

Vol 4 (2)

on only statistical grounds (Keenan et al., 2003, pp. 99-100). The complication and interdependence of the brain is paramount to keep

in mind as one examines brain/behavior relationships. Neuroimaging

There is no agreed upon definition of'neuroimaging.' Most neuroscien-tists identify 'non-invasive' as being a key element such that the brain can be observed in living, awake, humans. This is not to say that neuroimaging is limited to humans or the awake brain, but rather that these are options. Obviously, i f we are interested in analyzing the brain correlates of musical processing, gaining information about the human brain is vital. Therefore, there has been and continues to be great excitement about neuroimaging in studies related to music. However, even the definition of'non-invasive' is an issue. Techniques commonly thought of as non-invasive including SPECT and PET involve the use of radioactive ligands that need to be injected into the participant. Therefore, some might suggest that these procedures are far from non-invasive.

Brain mapping techniques are many, with the major techniques includ-ing: CT (Computed Assisted Tomography), sMRI (Structural Magnetic Resonance Imaging), SPECT (Single Photon Emission Tomography), PET (Positron Emission Tomography), fMRI (functional Magnetic Resonance Imaging), ERP (Event Related Potential), EEG (Electroencephalogram), TMS (Transcranial Magnetic Stimulat ion), and rTMS (repetit ive Transcranial Magnetic Stimulation). Each of these methods has advantages and disadvantages in elucidating brain / behavior correlates.

The choice of the technique employed wi l l depend on the hypothesis that one has. For example, i f one is curious in discovering what brain re-gion differentiates a tone from a voice, temporal resolution may be impor-tant. That is, one might want to examine what is occurring at different brain regions at 150ms after the stimuli are presented. The proper technique would be ERP as opposed to fMRI.

Two of the most popular techniques are sMRI and fMRI. These imaging techniques are widely used now in the study of musical processing. Be-cause of this, they wi l l be examined in some detail. Structural vs. Functional

Neuroimaging comes in two basic 'flavors.' One can examine either the anatomical or functional (or both) underpinnings of any behavior, in this case, music. The anatomical underpinnings involve the structure, or basic physical makeup of a brain correlate. For example, discovering that the frontal cortex of musicians is bigger than the frontal cortex of non-musi-cians would be a typical anatomical study. For these studies, we generally turn to sMRI. '

Functional studies examine actual brain activity. Whereas structural stud-ies examine the underlying structure, functional studies investigate some type of activity in the brain as it relates to a given behavior. For example, one might want to see i f certain brain regions are 'active' when one hears a trumpet as compared to hearing a cello. Typically, the goal is to determine the underlying neuronal activity, but this is not always the case. MRI and Music

The process of magnetic imaging is quite complex. However, at a very basic level, MRI works by measuring different magnetic properties in a steady state magnetic field. Different parts of the brain have different mag-netic properties. By tracing different magnetic properties in three-dimen-sional space, a 'map' of the brain is created. This map can be viewed a number of ways such as in the axial, sagital, and coronal planes (i.e., from the top, side and front). Typically, these 'slices' are taken in resolutions as high as 1mm which provides an excellent view of gross structures, but leaves researchers quite far from viewing actual neurons.

In examining the value of sMRI and the anatomical technique, consider a recent paper from our group (Keenan et al., 2001). This paper focused on the Planum Temporale (PT) and extended previous results (Schlaug, Jancke, Huang, & Steinmetz, 1995). The PT is a region often associated with lan-guage and auditory processing. We were examining the PT in regards to Absolute Pitch (AP). We thought that there might be right / left differences in terms of the PT in AP musicians based on previous results. The PT is found to be asymmetric in normal samples, with a greater leftward bias, that is, the left side is larger than the right. The PT has been used as a marker of hemispheric dominance. There arc studies that have found dif-ferences in PT asymmetries depending on handedness (Foundas, Leonard, Gilmore, Fennell, & Heilman, 1994; Shapleskc, Rossell, Woodruff, & David, 1999; Steinmetz, Volkmann, Jancke, & Freund, 1991; Tzourio, Crivello,

Mellet, Nkanga-Ngila, & Mazoyer, 1998) There is evidence (for and against) in terms of gender and PT asymmetries (Foundas, Faulhaber,

Kulynych, Browning, & Weinberger, 1999; Idc, Rodriguez, Zaidel, &

Vol 4 (2)

Aboitiz, 1996; Jancke & Steinmetz, 1993; Kulynych, Vladar, Jones, & Weinberger, 1994), There has also been research examining possible asym-metries in terms of schizophrenia (Barta et al., 1997; Kleinschmidt et al., 1994; Shapleske et al.) and dyslexia (Leonard et aL, 1993; Shapleske et al.). These studies typically revolve around the implications of hemispheric asymmetries of brain regions related to language and auditory processing.

To determine i f there were underlying brain differences, we examined three groups: musicians with AP, musicians without AP and non-musicians. SMRIs were taken of all of the participants.

There were a number of interesting findings: 1) It was found that there was a significantly greater leftward asymmetry for the AP musicians when compared to the non-AP musicians and non-musicians; 2) The AP group had a significantly smaller right PT compared to both groups; 3) The left PT was only marginally larger in the AP group; and, 4) The size of the right, and not left PT, was a better predictor of music group membership.

These results indicated that it was the right hemisphere that might be driving differences between AP musicians and others. Because Ihc left / right difference was found to be bigger in AP musicians, it was always assumed that the AP musicians had larger left PTs. Instead, it appears that the smaller right is driving group differences. My colleagues and I assumed that a process known as pruning might be occurring. fMRI and Music

I wi l l provide a second example from our group to demonstrate how fMRJ can be employed to investigate musical processing. In this study, we were asking a seemingly simple question: Arc there gender differences in the processing of musical stimuli in terms of brain function?

There is evidence that females are less laterali/.cd than males in terms of cortical brain anatomy (Coney, 2002; McGlonc, 1980; Voyer, 1996) and function (Coney). These studies indicate that males (end to use either the right or the left hemisphere for a given task whereas females tend to use the right and the left together on more tasks. Not all studies, however, agree with this idea (Frost et al., 1999; Kcrlesz & Benke, 1989; Speck et al., 2000).

We therefore hypothesized that there might be (musical processing dif-ferences in the brain in terms of gender. The experiment employed a pitch memory task during fMRI scanning. Wc found a number of differences in terms of gender, the brain, and music processing: 1) Male participants had greater laleralizcd activations (left > right) in anterior and posterior perisylvian regions; 2) Wc found a trend for males to have more cerebellar activation than females; and, 3) The female participants showed more promi-nently posterior cingulalc activation compared to males. Our overall con-clusion was that similar to language studies males rely more on left lateral-ized hemispheric processing even for musical tasks. That is, males tended to activate more left regions and we did not observe a real bilateral activa-tion in the males as compared to females.

Notice how this study differs from the anatomical study of Keenan et al. (2001). In Gaab et al. (2003), the brain 'in action' is captured. The partici-pants were scanned and different pitches were presented. We captured the activity of the brain during pitch memory tasks, and therefore derived dif-ferences based on the function of the brain.

Issues of Neuroimaging I have presented two basic neuroimaging experiments. The first (Keenan

et al., 2003) examined left/right anatomical differences with regard to a behavioral characteristic (e.g., AP). The second study (Gaab et al., 2003) examined functional differences in pitch processing depending on gender. These studies provide a backdrop to discuss pertinent issues in neuroimaging in regards to imaging. One must understand the underlying methods be-fore one is able to interpret the data. Control Condition

The first issue that must be examined is the idea of a control condition. The control condition in neuroimaging is critical for the interpretation of the results. This being said, it is interesting to note that people outside of neuroimaging rarely consider the importance of the control condition in interpreting these data.

Consider the AP paper that was presented. The question revolved around the basic idea that absolute pitch musicians have brains that are unique. The critical question becomes, "Unique in comparison to what?" That is, i f the brains of AP musicians are different, how are we deciding this and who or what are we comparing the AP group to.

Consider first the idea of having no controls. One would examine an area of the brain (Figure la) and determine its size, shape, volume,, etc However, without a control condition, there is nothing to compare the( group

Bulletin of Psychology and the Arts

Anatomical

Functional

Figure 1. Two portrayals of an anatomical (A) MRI and a functional MRI (B). Both demonstrations convey the need for control conditions. Simply measuring the volume, size, etc. of a brain region does little to tell us about that brain region. Understanding of the measure almost always depends on a comparison condition. to. In our case, the experimental group (AP musicians) needed a control condition. In our experiment, we compared the AP musician group to a non-AP musician group and a non-musician group. These controls are v i -tal. In fact, the controls wind up being as important as the experimental group. We found that overall laterality differed between the AP musicians and both groups. Critical to the study, however, was also the comparison between the non-AP musicians and the non-musicians. These groups did not differ which makes one think that the laterality in the brain is not due to just being a musician, but rather due to AP.

The same control issue applies for fMRI as well. It is tempting to think that we can just scan a brain while a person listens to Bach and then derive the 'Bach center' (Figure lb). Again, one must have controls. I f a group of participants are placed in an fMRI , and we play Bach, certain brain regions wi l l become active. However, the brain may be responding to the noise of the scanner, the fear of being in a tight space, the temperature of the room, some type of mental imagery, etc. Therefore, similar to MRI, one must have a control fMRI condition.

I f we are interested in finding the 'Melody center,' we wil l need to com-pare activity listening to a melody with some other condition. This other condition might be an atonal melody (vs. a tonal melody) or perhaps the same piece of music in which a subject attends to the rhythm as opposed to attending to the melody. One may, alternatively, wish to have a melodic music condition compared to a no music condition. The reality is that the resulting comparison wi l l usually differ dramatically depending on the con-trol condition employed (Figure 2).

Tonal Melody No Music Unique to Tonal Melody

Tonal Melody Atonal Melody Unique io Tonal Melody

. ,_ Figure 2. Two imagined fMRI studies in which one is looking for areas of 6 2 1 activation associated with listening to a tonal melody. In the first study (A),

activation associated with listening to no music is subtracted from the tonal

Bu l l e t i n o f Psychology and the A r t s

melody condition. In the second study, activation associated with an atonal melody is subtracted (B). Notice the differences in the final image. In A, we would determine that a tonal melody is associated with activity in the dorsal and ventral frontal cortex. In B, we would assume that a tonal melody is associated with dorsal activity only. These examples demonstrate the importance of the control condition.

The proper control cannot be emphasized enough. Continuing the melody example, one must ask what is desired in performing this experiment. For example, we might have chosen to examine melody because we are inter-ested in discovering i f certain regions respond to melodies that are atonal or tonal. I f this were the case, one might want to compare a piece that has the same number of notes, the same rhythm, the same length and tempo but only differs on the notes being played in certain bars. For example, one might put in a flat fifth to replace a major fifth. However, it should be pointed out that the control condition here can be modified further. For example, it would certainly make a difference i f the tonal melody we are altering is known (e.g., 'Jesu, Joy of 'Man's Desire') or not known (an obscure composition). However, the degree of alteration (which would be the control condition) would effect the eventual subtraction. Therefore, i f we altered only one note of 'Jesu' and compared it to the original, we would find one pattern of brain activity. However, i f we changed every major third to a minor third in the piece, and compared it to the original, we would get a completely different pattern of brain activity.

These issues can continue throughout entire lines of research. I f we truly Want to know i f tonal melodies activate certain brain areas, we wil l have to run many experiments and the control conditions wi l l have to vary. In the above example, we could compare the original 'Jesu' to a slightly altered version as well as a dramatically altered version. However, any difference we find between the conditions might be due to any change in any piece of music. That is, the brain areas that are 'lighting up' may be a change center rather than a change in melody center. Therefore, more controls wi l l be needed such as comparing a tonal melody that is original to an altered version that is still tonal. Rarely is an area implicated in one set of studies. Subject Histoiy

It is important to consider subject history in terms of neuroimaging. The disadvantage of the post-mortem exam (which neuroimaging is suppos-edly replacing, at least in part) is that many subject attributes may be re-sponsible for a given deficit observed. For example, a patient examined with amusia may also have a condition prior to, concurrent with, or follow-

' ing the development of the disorder. These conditions may lead to brain abnormalities that might be mistakenly interpreted as having a direct rela-tionship to amusia.

Because sMRI has similarities with the autopsy method, it is important to consider subject history. For example, we can also perform an sMRI on a person that developed amusia. However, the same considerations must be used as would be with a post-mortem exam. For example, i f the sMRI takes place 4 years following the development of the condition, it becomes difficult to attribute amusia to any underlying brain correlates.

In other words, in many studies it would be advantageous to have a baseline condition. For example, in the amusia example, there is rarely a 'before amusia' condition. In other words, we rarely have pre/post data on any given patient. This basic experimental consideration is important lo remember when interpreting the results of any anatomical MRI study in-volving the onset of an injury or accident Further, subjeel history is an important consideration when running any neuroimaging study Tcsling a homogeneous group is paramount in minimizing within groups error. In our fMRI study on pitch processing (Gaab ct al., 2003), we made sure that the participants were matched on musical ability. Causality

Every psychology undergraduate can recite the phrase "Correlation does not equal causation." This simple rule is critical in interpreting neuroimaging results. Neuroscientists caution that just because brain activity is associ-ated with a given behavior, that alone is not evidence of causality. For example, i f there is an increase in activity in the superior temporal sulcus (STS) when one hears a trumpet, we cannot assume that this is a 'trumpet' area or that the STS causes or allows for the recognition of a trumpet. A l l that is established is that activity in the STS is associated with trumpet playing.

This fact can be difficult to convey to those not in the neuroimaging community. It is tempting to assume that areas that 'light up' are 'music areas,' 'Bach regions' or 'tonal/atonal areas' and there is a direct, causal, 1:1 relationship. However, this is not the case. No amount of increase in resolution wil l make MRI or fMRI a tool capable of determining causal-f 63 ity.

Vol 4 (2)

Our paper that demonstrated anatomical PT differences in AP musicians is a good example of this (Keenan et al., 2001). Just because there is a difference in PT laterality and size, one does not know i f that is causing AP. Further, it is not known i f AP causes a change in PT size or vice versa. Further, it is possible that this is a spurious correlation and that it is another brain region that is critical for AP. It is here that other techniques can ben-efit the neuroimaging researcher. For example, we may wish to examine patients with PT damage to see i f they lose their AP abilities, gain AP abili-ties, etc. From our toolbox, we may move beyond correlation. However, it is rare that fMRI or sMRI can establish causality. What is Activation?

Most of what you see in an fMRI is a statistical illusion. Looking at an fMRI image, one imagines that the image represents which groups of neu-rons are active. However, this is not the case.

FMRI, in fact, does not measure neural activity directly. Instead, fMRI measures changes in the magnetic field. Neuronal activity does not change the magnetic field to any great extent. Oxygen, however, does change the magnetic field. It turns out that there are differences in oxygen levels of blood depending on differences of regional neural activity. It is these dif-ferences in oxygen concentrations that are measured in most fMRI studies.

Notice that we are not measuring neuronal activity, nor are we measur-ing blood flow. We are measuring differences in oxygen concentrations. We assume that the oxygen demands of an area of the brain involved in a given activity are greater than an area that is not involved. It is important to note that activity is not measured directly.

Further, and equally important, is the idea of the 'brain map' that is usu-ally presented as a figure in an fMRI study. The image is usually a brain that has different colors such as red, yellow and blue. Seeing this brain, one assumes that the red stands for regions that are highly active, yellow somewhat active and blue might mean a negative activation.

These maps are actually statistical constructions (Figure 3). Magnetic changes are recorded for a given condition and compared to magnetic

m Tonal Atonal

Time point 4

Time point 5

If significance is found at a .05, we color the pixel yellow If significance is found at a.01, we color the pixel red. If no significance is found, the pixel is left blank In this case, p

Vol 4 (2)

Tonal vs. Atonal t-test was for a single pixel. Depending on the experiment and the equipment, one can divide the brain up into 100,000 pixels. Figure 3 depicts, therefore, only 1 of a possible 100,000 comparisons that might be needed to analyze the entire brain. This means that we have to carry out comparisons over 100,000 pixels. That means 100,000 tonal vs. atonal t-tests wi l l need to be performed. Given this, one would have a type I error rate of 5,000%, slightly above the desired 5%! Therefore, there is obvi-ously a chance for type I error.

One could make a Bonferroni adjustment. The new alpha rate would be .0000005. This would now present the problem of increased type I I error rate. It would become quite difficult to find any significant pixels at this alpha. ROI

One of the issues worth examining is how we defined the PT in our AP study (Keenan et al., 2001). In comparing structure to structure, it becomes extremely difficult to automate the process. That is, it is nearly impossible to program a computer to find the PT on an MRI image. Therefore, the size (shape or volume as well) of a given area is typically done via hand trac-ing. This is where an expert anatomist plays a role, knowing (he bound-aries of a given region. In the case of the PT, this is a fairly straightforward process as it is one of the most studied regions of the brain. However, even with all of the equipment at our disposal, defining the PT relied on actual tracing. While the computer tabulated all of the values, it was a human that needed to define the region of interest (ROI).

Therefore, it is possible (and likely) that human error is introduced into the process. I f we are interested in examining the prefrontal cortex, at some point a human is going to have to clclcnninc what defines the prefrontal cortex. Like the maps in fiVTRr, this fact hopefully serves as a warning in terms of functional imaging and music. One must read carefully in the methods about the manner in which regions were defined or measured. Other Issues

Noise, Both sMRI and fMRI require 'noise' to acquire images. Anyone who has had an MRI can attest to the loud banging and knocking that occurs. In almost every MRI center, people are required to wear protective ear plugs or head phones. Obviously, this w i l l play a large role in both conducting and interpreting music studies. Several groups have invented special fMRI techniques (e.g., sparse temporal sampling techniques) to circumvent the influences of scanner noise on the task and on auditory activation (Gaab ct al., 2003).

Movement. Again, both sMRI and fMRI require the participant to re-main still. The MRI scanner does not 'know' what it is scanning and it wi l l scan anything that is in a given area. Therefore i f one moves, the MRI does not adjust. Just like movement wil l blur a photograph, movement wi l l dra-matically alter an MRI image. Therefore, many studies have a bite bar (an object to bite on) to stabilize the participant. When considering a study, this is important. Even moving the arm (e.g., a piano simulation) must be controlled so thai the head does not move.

Metal. Because one is measuring a magnetic field, there can be no metal in the scanning environment. Al l studies must be planned with this in mind.

Inhibition vs. Excitation. Beyond the scope of this paper is an examina-tion of the issue of inhibition vs. excitation. However, fMRI does not typi-cally differentiate between neuronal activity that excites and neuronal ac-tivity that inhibits. Therefore, active can mean an increase in inhibition.

Conclusion There are many issues one must deal with when focusing on neuroimaging

and the brain. The first is the complication of the brain itself. While it is tempting to ascribe a single region to a certain function, this is rarely how the brain works. The brain, a complex mobile, is a highly intra-dependent entity. Activity in one region is often dependent upon other regions. The brain should thus be looked upon as a balanced system.

Neuroimaging has the potential to provide a tremendous amount to our understanding of the relationship between music and the brain. However, understanding the limitations of the techniques themselves wi l l benefit the individual either reading the literature or planning a study. The complexi-ties are many and pitfalls exist. It is important to keep these in mind when examining brain - music relationships.

References Barta, P. E., Pearlson, G. D., Brill, L. B., 2nd, Royall, R., McGilchrist, I. K.,

Pulver, A. E., Powers, R. E., Casanova, M. R, Tien, A. Y , Frangou, S., & Petty, R. G. (1997). Planum temporale asymmetry reversal in schizophrenia: replica-tion and relationship to gray matter abnormalities. American Journal ofPsy chiatry, J54(5), 661 -7.

Coney, J. (2002). Lateral asymmetry in phonological processing: Relating

Bulletin of Psychology and the Arls behavioral measures to neuroimaged structures. Brain and Language, 80, 355-365.

Foundas, A. L., Faulhaber, J. R., Kulynych, J. J., Browning, C. A., & Weinberger, D. R. (1999). Hemispheric and sex-linked differences in Sylvian fissure morphology: A quantitative approach using volumetric magnetic reso-nance imaging. Neuropsychiatry, Neuropsychology, and Behavioral Neurol-ogy, 72(1), 1-10.

Foundas, A. L., Leonard, C. M., Gilmore, R., Fennell, E., & Heilman, K. M. (1994). Planum temporale asymmetry and language dominance. Neuropsychologia, 32(10), 1225-31.

Frost, J. A., Binder, J. R., Springer, J. A., Hammeke, T. A , Bellgowan, P. S., Rao, S. M., & Cox, R. W. (1999). Language processing is strongly left lateral-ized in both sexes: Evidence from functional MRI. Brain, 722(Pt 2), 199-208.

Gaab, N., Kccnan, J. P., & Schlaug, G. (2003). The effects of gender on the neural substrates of pitch memory. The Journal of Cognitive Neuroscience, 15(6), 810-20.

Idc, A , Rodriguez, E., Zaidel, E., & Aboitiz, F. (1996). Bifurcation patterns in the human sylvian fissure: Hemispheric and sex differences. Cerebral Cor-tex, 6(5), 717-25.

Janckc, L., & Sleinmetz, H. (1993). Auditory lateralization and planum temporale asymmetry. Neuroreport, 5(2), 169-72.

Kccnan, J. P., Gallup, G G., & Falk, D. (2003). The face in the mirror- The search for the origins of consciousness. New York, NY: HarperCollins/Ecco.

Keenan, J. P., Thangaraj, V., Halpern, A., & Schlaug, G. (2001). Absolute pitch and the planum temporale. Neurolmage, 14, 1402-08.

Kertesz, A., & Benke, T. (1989). Sex equality in intra-hemispheric language organization. Brain and Language, 37(3), 401-408.

Kleinschmidt, A., Falkai, P., Huang, Y , Schneider, T , Furst, G, & Steinmetz, H. (1994). In vivo morphometry of planum temporale asymmetry in first-epi-sode schizophrenia. Schizophrenia Research, 12(\), 9-18.

Kulynych, J. J., Vladar, K., Jones, D. W., & Weinberger, D. R. (1994). Gender differences in the normal lateralization of the supratemporal cortex: MRI sur-face- rendering morphometry of Heschl's gyrus and the planum temporale. Cerebral Cortex, 4(2), 107-18.

Leonard, C. M., Voeller, K. K., Lombardino, L. J., Morris, M. K., Hynd, G. W., Alexander, A. W., Andersen, H. G, Garofalakis, M., Honeyman, J. C , Mao, J., & et al. (1993). Anomalous cerebral structure in dyslexia revealed with magnetic resonance imaging [see comments]. Archives of Neurology, 50(5), 461-9.

McGlone, J. (1980). Sex differences in human brain asymmetry: A critical survey Behavioural Brain Science, 3, 215-227.

Schlaug, G, Jancke, L., Huang, Y , & Steinmetz, H. (1995). In vivo evidence of structural brain asymmetry in musicians. Science, 257, 699-701.

Shapleske, J., Rossell, S. L., Woodruff, P. W., & David, A. S. (1999). The planus temporale: A systematic, quantitative review of its Structural, functional and clinical significance. Brain Research Brain Research Reviews, 29(1), 26-49.

Speck, O., Ernst, T , Braun, J., Koch, C, Miller, E., & Chang, L. (2000). Gender differences in the functional organization of the brain for working memory. Neuroreport, 77(11), 2581-2585.

Steinmetz, H., Volkmann, J., Jancke, L , & Freund, H. J. (1991). Anatomical left- right asymmetry of language-related temporal cortex is different in left-and right- handers. Annals of Neurology, 29(3), 315-9.

Tzourio, N., Crivello, R, Mellet, E., Nkanga-Ngila, B., & Mazoyer, B. (1998). Functional anatomy of dominance for speech comprehension in left handers vs right handers. Neurolmage, 8(1), 1-16.

Updike, J. (1999). Hub fans bid kid adieu. In D. Halberstam (Ed.), The best American sports writing of the century (pp. 304-17). New York, NY. Houghton Miffl in.

Voyer, D. (1996). On the magnitude of laterality effects and sex differences in functional brain asymmetries. Laterality, I , 51-83.