Hemispheric Specialization and Aesthetics John P. McLaughlin Since Paul Broca's famous observation in 1861 that a failure of speech was accompanied by a lesion in a patient's left hemisphere, evidence has accumulated at an ever-increasing rate that the brain's two hemispheres are specialized for different activities. Second in importance only to Broca's trailblazing observation were the reports on the "split-brain" patients that emerged in the early 1960s (cf Sperry, 1968). Behavioral tests on these patients strongly indicated that the two hemispheres were responsible for very different things. Recently, evidence has suggested that some portions of aesthetic experience depend more on one side of the brain than on the other. This chapter reviews that evidence. The two hemispheres of the brain are connected subcortically by large collections of nerve fibers, called cerebral commissures. These commissures serve the purpose of transferring information from one hemisphere to the other. In a "last resort" attempt to relieve the symptoms of psychomotor epilepsy, neurosurgeons cut these pathways to block interhemispheric transfer. The symptoms of this form of epilepsy are characterized by the gradual emergence of tremor in one part of one side of the body, followed by more severe tremor on the other side, involving a greater body area, followed in turn by even more severe tremor on the first side, involving an even greater body area and so on, back-and-forth. The surgeons reasoned that this pattern could be reduced, if not eliminated, were the commissures cut, because the normal path for the disruptive signals would be eliminated. They were proved correct. Patients showed dramatic reductions in seizure activity after the surgery. The effect of the surgery was to disassociate the two hemispheres. This provided interested scientists the opportunity to study what each side of the brain could do independently of the other. However, before getting to the research findings, a little more anatomy must be described so that the remaining connections between the hemispheres and the body are understood. Each hemisphere receives information from the various sensory receptors of the body. In the case of vision, the anatomical pattern is somewhat complicated, but its result is very simple. The right side of the brain receives the information about the left half of our visual world and the left side of the brain receives information about the right half of that world. The diagram (P. 35 in Solso's text) shows the relevant anatomy of the eyes and the optic nerves as if we were looking down from above. Consider the left eye. Receptor cells line the inside surface of the back half of the eye, forming the retina. The half that lie on the "earward" side of the retina send fibers in a bundle directly back to the left hemisphere . Notice which parts of the retina receive light from the right visual world. That light enters through the lens and falls on the "earward" receptors. Thus, from this eye, the left hemisphere gets information about the right visual world. Light entering from the left visual world falls on the "noseward" side of this retina. These receptors send a bundle of fibers across the head and then back to the right hemisphere. So, for this eye, the right hemisphere receives information about the left visual world. The area where this bundle crosses over is called the optic chiasm, which will be important a little later in this discussion. Now consider the right eye in the diagram. The "noseward" receptors send their fibers in a bundle (speckled in the diagram) across the optic chiasm to the left hemisphere, so again, the left hemisphere receives information about the right visual world. Finally, the "earward" receptors send their fibers directly back to the right hemisphere so that it gets information about the left visual world. The surgery to sever the commissures also severs the optic nerve fibers at the optic chiasm (crossover bundles), so the split-brain patients see with a smaller set of operative receptors. However, because noncrossover bundles are intact, these patients receive information about the whole visual world. In the case of the skin senses, e.g., touch, warmth, the sensory nerve fibers from the skin are part of the left and right ascending tracts of the spinal cord. Upon reaching the brain stem, these fibers cross over to the other side of the brain. Thus information about the right side of the body is received directly by the left side of the brain and vice versa. This same organization is found in the nerves that carry motor commands from the brain to the muscles. Movement of the right side of the body is directed by the left hemisphere and vice versa. None of these nerves, sensory or motor, are severed in the commissurotomy surgery. What is a split-brain person capable of doing ? If you met such a person and were ignorant of her medical history, you would never suspect that there was anything unusual about her. Differences between these patients and normal subjects only show up in very specific tests. The results of such tests indicate that there are differences in what each hemisphere can do. What I will describe here is true of right-handed subjects and of some left-handers. The left hemisphere controls speech and the right does not. If a word is flashed up on a screen so that it falls in the patient's left visual field (LVF) the subject will not be able to say the word. Remember that information from the LVF goes to the right hemisphere. If, on the other hand, the word is flashed in the patient's right visual field (RVF), the patient can say the word. The right hemisphere can understand the word, however. It simply cannot produce the speech. If the patient uses his left hand to feel objects hidden from view, he can find the object corresponding to the word. The right hemisphere is better than the left in processing spatial information. If a drawing is flashed up in the LVF, the patient can draw the pattern rather well with her left hand. If, however, the pattern is flashed in the RVF and the patient draws with her right hand, the drawing is terrible. Since these early reports, many studies have demonstrated that data from normal subjects reveal differences in the performance of each hemisphere. In other words, even though the commissures are intact, manipulating variables like spatial location of information produces differences in behavior. For example, if a subject is asked to locate a brief event in space, his performance is better when the event occurs in the LVF. What has this to do with aesthetic experience ? A number of studies have shown that right- and left-handers differ in their preferences for paintings (Levy, 1976; Mead & McLaughlin,1992). The most likely interpretation of these differences is that the visuospatial ability of the right hemisphere is responsible, although new data seem to require a refinement of that explanation. If, for some reason, you see a painting in a mirror, or if a photograph is reversed in printing, the appeal of the painting or photograph may be reduced. This effect was first observed by an art historian, Woelfflin, and later by a psychologist, Mercedes Gaffron (1950). They both believed that paintings were scanned in a predictable way, with the observer beginning in the lower left of the picture space and proceeding up, to the right and into the space. If a painting was composed so that important segments of information fell along that path, the painting would be pleasing. If the painting was reversed, however, the flow of information would not be consonant with the scan-path and aesthetic appeal would suffer. Jerre Levy (1976) was the first to find that the effect was related to handedness. Using photographic slides of land and seascapes as stimuli, she asked right-handed subjects to choose between each original and its mirror-reversed version. A statistically significant proportion of these subjects chose one particular version over the other in 14 out of 97 pairs. An additional sample of right-handers also preferred those versions. A third group of subjects judged that the area of principal interest or heaviest weight was in the right half of the picture space for those 14 preferred versions. Left handers showed no particular preference for these versions nor did right handers for symmetric pictures. Why should that have happened ? Levy found an explanation in the visuospatial ability of the right hemisphere. There are data which suggest that increased activation of one hemisphere increases attention to the part of the visual world normally served by that hemisphere. Levy reasoned that engaging in a task involving pictures would increase right-hemisphere activation and thus boost attention to the left half of the world. She further argued that this bias could make a picture look lopsided unless the pattern in the picture could counteract the bias. A picture containing interesting material in its right half might do just that. The preference for such pictures, in this line of reasoning, is that they look better because they cancel distortions produced by the perceptual bias caused by the increased activation of the right hemisphere. An alternative explanation was offered by Beaumont (1985), although it too rests on the assumption that the right hemisphere is specialized for visuospatial processing. If a picture presents more interesting material in the right half of its space, two things will happen, he reasoned. First, the subject will look at that material and that will result in the material being processed by the most acute and efficient portion of the visual system. Second, looking at the right half of the picture puts more than half of the picture in the observer's LVF, thus assuring that most of the picture's visuospatial information will be sent directly to the right hemisphere. This composition of the picture, then, promotes aesthetic appeal because its information goes directly to its appropriate destination in the brain. McLaughlin, Dean and Stanley (1983) argued that Levy's selection of the 14 pairs left open the possibility that other pairs among the 97 might have been just as asymmetric in their composition but not have shown the preference effect. Therefore, they assembled sets of paintings judged either to contain areas of interest in off-center areas of the picture space or to be symmetric. In one experiment, right handers chose versions as Levy would have predicted, i.e., they chose more versions containing areas of principal interest in their right halves than in their left. Unlike Levy's findings, left handers in this experiment chose significantly more versions with areas of interest in their left halves. The two subject groups were not choosing opposite versions of the same pairs, however. Where one group showed a definite preference for one version of a pair, the other group was likely to be neutral, raising the possibility that right- and left-handers were attending to different cues. A second experiment duplicated the finding for left-handers, but failed to detect a definite preference for right-handers. Clearly, variability is a problem for the picture-reversal effect. Other investigators (Freimuth and Wapner, 1979) failed to find an effect on preference when location of interest was varied, but did find that the presence of cues to motion did affect preference. Versions of paintings that implied left-to-right motion were preferred over their mirror-reversed counterparts by right-handers. Using a larger set of paintings, Mead and McLaughlin (1992) also found that right-handers preferred versions with left-to-right implied motion. These paintings had been selected because they were judged to contain motion cues and because they were symmetric in the location of areas of interest. Right-handers also preferred versions that were asymmetric in the location of weight in the composition, but they preferred the extra weight in the left half of the picture space, rather than the right. Unlike the earlier work, this experiment also detected preferences among some left-handers that were similar to those of right-handers. Some left-handers, called inverters, exhibit a distinctive writing posture in which the wrist is bent and the pen points down the page, toward the writer. Typically, the hand is above the line being written. Levy and Reid (1976) demonstrated that inverters perform similarly to right-handers on some simple tests of lateralization. Noninverters, whose writing hand is below the line and whose pen point up the page, performed in opposite fashion on those tests. Mead and McLaughlin (1992) found that the inverters also preferred versions with left-to-right motion cues and that noninverters displayed no particular preference. The good news here is that differences in preference paralleled performance differences on other tests of laterality, indicating that the aesthetic judgments are controlled by some lateralized mechanism. The bad news is the lack of agreement from one study to the next on findings from left-handed subjects. It has been suggested that the left-handed population is even more diverse than the writing posture criterion indicates, so variability in findings may reflect hidden sources of variability among subjects. Another possibility may be in variability of the stimulus sets. There may be particular conjunctions of cues that are important for preference that are not uniformly present in the stimuli. At this point in the story, we ought to consider the theoretical explanations advanced by Levy (1976) and by Beaumont (1985). Left-to-right motion cues might draw attention and eye fixations to the rightmost portions of the painting. Were that so, then most of the picture would be placed in the LVF. Beaumont's (1985) model, then, can account for these preference patterns. Levy's (1976) explanation assumes attention is deployed to the left and is counteracted by areas of interest in the right. Cues to motion could play that role also, so her model can also account for the data from paintings that imply motion. Mead and McLaughlin's (1992) finding that extra weight was preferred in the left portions of paintings seems inconsistent with Levy's model, however. Banich, Heller and Levy (1988) argued that Levy's original method of stimulus selection was the best way to discover exactly what combination of stimulus variables determine preference. They showed that individual pictures for which one version was significantly preferred over the other by right-handers had greater ratings for interest asymmetry and for implied motion than paintings that had been selected for the presence of one or the other asymmetry. These latter paintings attracted mean preference judgments which were significantly different from zero, but not as great as the preference scores for the former set. Banich, et al (1988) suggested that the reduced preference scores might have been due to conflicting asymmetries in those pictures. They also found that left-handers choices were quite different from those of right-handers. The preferences expressed by their subjects were different from those reported by Freimuth and Wapner (1979) and by Mead and McLaughlin (1992) for stimuli with implied motion. Banich, et al (1988) reported preferences for right-to-left motion. At this writing, it is not clear what procedural differences might account for the discrepancies in results. Nevertheless, continued observation that right- and left-handers are different in their patterns of choice preserves the interpretation that a lateralized mechanism is involved in these aesthetic judgements. A very different set of data suggest that hemispheric specialization may affect art and aesthetic preferences in another way. Portrait painters rarely depict the full face of a sitter, but rather portray the person in some degree of profile. A number of investigators, e.g., McManus and Humphrey (1973) have reported that the most common pose is that which exposes the sitter's left cheek and partially or completely hides the right cheek. The data that indicate the left cheek is more often exposed are based on counts of paintings either held in museum collections or published in books. For example, McManus and Humphrey (1973) surveyed 1474 paintings from both sources. They reported that 60% of the paintings in the sample showed more of the left cheek than the right. Moreover, there were effects attributable to the sex of the sitter. The left cheek was shown more often for women, 68%, than for men, 56%. McManus and Humphrey (1973) entertained a number of explanations, and then dismissed all but two. They suggested that the sex effect could be due to a possible preference for the left face that might develop in infancy. If a right-handed mother is likely to carry her baby in the left arm, then the infant will see the left cheek most of the time. Thus, a preference based on familiarity might develop. A second possibility is that the exposure of one cheek or the other is a symbolic communication of some sort, but their data did not offer any insight into what that might be. In another study, Humphrey and McManus (1973 ) reported data which they interpreted from the latter point of view. Restricting themselves to portraits by Rembrandt, they found that a majority of his self-portraits exposed the right cheek, but that the frequency of right-cheek exposure in portraits then fell off progressively across the sitters' categories of male kin, male non-kin, female kin and female non-kin. Humphrey and McManus perceived these categories as reflective of a dimension ranging from socially-like-myself to socially-unlike-myself and concluded that Rembrandt was conveying a message about his social distance from the sitter. Gordon (1974) pursued this idea in an examination of 295 portraits by Goya. Like the findings reported by McManus and Humphrey (1973), Gordon found that Goya's portraits exposed the left cheek more often and that this effect was more pronounced for female sitters than male. There was no evidence, however, for such a bias when kin and non-kin were compared. Based on 933 portraits in central European museums, Gruesser, Selke and Zynda (1988) also reported the sex difference for portraits of men and women, although the magnitude of the bias seems to have declined over the last six centuries. Dividing the number of portraits exposing the left cheek by that of portraits exposing the right, they showed, for portraits of women painted in the 15th Century, the ratio was approximately 8.0. However, the ratio rather systematically declined century- by-century until it was barely greater than 1.0 for portraits done in the 20th Century. For portraits of men, the 15th Century value was approximately 3.0, but all of the rest hovered above and below 1.0. Noting that portraits of women that did expose the right cheek tended to be "official" portraits of, for example, sovereigns, they concluded that the general trend to show women's left cheeks resulted from a sex bias. They suggested that the observed historical trend in the portraits is the result of improvements in the status of women in Western culture. What could the bias be ? It has been demonstrated that the left face, controlled for the most part by the right hemisphere, is the more expressive side of the face (c.f. Rhodes, 1985). Perhaps artists over the years have tried to portray women as more emotional or more expressive than men. Indeed, this may be the symbolic communication suggested by McManus and Humphrey (1973). The familiarity hypothesis suggested by them also remains a possibility. This hypothesis, of course, makes the assumption that the left-cheek portraits are preferred over their mirror-reversed versions, but, to date, only one study has sought evidence on this point. Recently, Benjafield and Segalowitz (1993) reported some analyses of portrait drawings by Leonardo da Vinci. The stimuli were eight profiles, four of women and four of men, half showing the right cheek and half the left. Subjects rated these drawings and their enantiomorphs on ten Semantic Differential Scales (Osgood, Suci and Tannenbaum, (1957) that loaded primarily on either Evaluation, Potency or Activity. Their most interesting finding was that judgments on the Potency and Activity scales discriminated originals showing one cheek or the other. Original drawings showing the right cheek and their reversed copies were seen as more potent and more active than originals showing the left cheek and their reversed copies. Put another way, Leonardo drew more potent-and active-appearing faces when he drew profiles showing the right cheek. Whatever cues contribute to those perceptions were effective whether the subjects looked at the original or at its reversed version. This effect did not interact with the sex of the sitter, but women's faces were rated as less potent and more positive than those of men. However, no effects of original cheek-orientation were observed for the Evaluation scales. Clearly, these data support the idea that artists, da Vinci at least, include additional information when a profile orientation is chosen. Since the female, left-cheek originals were not preferred, the familiarity hypothesis seems unsupported. However, the absence of an effect of cheek orientation on the evaluative scales is somewhat surprising from the perspective of the symbolic-communication hypothesis also. While painters throughout the centuries may have been sending sexist signals when emphasizing women's left cheeks, it seems reasonable to presume that: (a) they believed they were creating better art by doing it that way, and/or (b) their audiences and patrons preferred those versions. McLaughlin and Murphy (1994) obtained preference data for portraits using forced-choice procedures in which the subject chose between an original portrait and its mirror-reversed version. Both male and female observers preferred versions in which the right cheek was exposed, although men expressed a stronger preference. Similarly, both left- and right-handed subjects preferred the right-cheek versions, with the left- handers expressing a stronger preference. Neither the sex of the sitter for the portrait nor the cheek exposed in the original portrait affected preference. One explanation for those findings was that subjects chose the right-cheek versions because those versions are generally more salient to observers, who are perceptually biased to emphasize the cheek that tends to be in their left visual fields. There is ample evidence that the right hemisphere of the brain, at least for dextral subjects, is specialized to process information about faces. Gilbert and Bakan (1973) reported evidence that the salience of the right side of the face does not result from physical differences between the two sides of faces, but rather from a left visual-field perceptual bias attributed to the hemispheric specialization. This finding was replicated by Rhodes (1985). Other investigators (Levy, Heller, Banich and Burton, 1983; Morris and Hopkins, 1993) have employed chimeric faces as stimuli and reached conclusions that are consistent with Gilbert and Bakan and Rhodes. A chimeric face is a composite created by pairing, in a photographic print, one side of a face expressing an emotion with the other side expressing neutrality or a different emotion. Levy, Heller, Banich and Burton (1983) reported that, for smiling-neutral composites, subjects reported the expression displayed by the cheek of the composite in their left visual fields as the emotion of the whole display. Evidence of this bias in chimpanzees has recently been reported also, further reinforcing the explanation that there is a biological basis for the effect (Morris and Hopkins, 1993). The bias in emphasizing the left cheek of women demonstrated by generations of artists was not favored by these 20th century subjects. In that gender of the sitter did not determine preference, they seem to have been possessed by the same happy, egalitarian spirit that Gruesser, et al (1988) attributed to 20th century artists. On the other hand, those same artists should be informed about these strong preferences for the right cheek. References Banich, M. Heller, W. and Levy, J.(1989) Aesthetic preference and picture asymmetries. Cortex, 25, 187-195. Beaumont, J. G. (1985) Lateral organization and aesthetic preference. Neuropsychologia, 23, 103-113. Benjafield, J. and Segalowitz, S.J. (1993) Left and reight in Leonardo's drawings of faces. Empirical Studies of the Arts, 11, 25-31. Freimuth, M. and Wapner, S. (1979) The influence of lateral organization on the evaluation of paintings. British Journal of Psychology, 73, 211-218. Gaffron, M. (1950)Left and right in pictures. Art Quarterly, 13, 312-321. Gilbert, C. and Bakan, P. (1973) Visual asymmetry in perception of faces. Neuropsychologia, 11, 355-362. Gordon, I. (1974) Left and right in art. in D. O'Hare(ed.) Psychology and the Arts. Harvester, Sussex. Greusser, O-J.,Selke, T. and Zynda, B. (1988) Cerebral lateralization and some implications for art, aesthetic perception and artistic creativity. in I. Rentschler, B. Herzberger and D. Epstein(eds) Beauty and the Brain. Birkhauser, Basel, 257-293. Humphrey, N.K. and McManus,C.(1973) Status and the left cheek. New Scientist, 59, 437-439. Levy, J. (1976)Lateral dominance and aesthetic preference. Neuropsychologia, 14, 431-445. Levy, J. and Reid,M.(1976) Variations in writing posture and cerebral organization. Science, 194, 337-339. Levy, J., Heller, W., Banich, M. and Burton,L.A.(1983) Asymmetry of perception in free viewing of chimeric faces. Brain and Cognition, 2, 404-419. Mead, A.M. and McLaughlin, J.P.(1992) The roles of handedness and stimulus asymmetry in aesthetic preference. Brain and Cognition,20,300-307. McLaughlin, J. P., Dean, P. and Stanley, P.(1983) Aesthetic preference in dextrals and sinistrals. Neuropsychologia, 21,147-153. McLaughlin,J.P. and Murphy, K.E. (1994) Preference for profile orientation in portraits. Empirical Studies of the Arts, 12, 1-7. McManus,I.C. and Humphrey,N.K. (1973) Turning the left cheek, Nature, 243,271-272. Morris, R.D. and Hopkins,W.D. (1993) Perception of human chimeric faces by chimpanzees. Brain and Cognition, 21, 111-122. Osgood,C.E. Suci, G.J. and Tannenbaum,P.(1957) The Measurement of Meaning. Univ. of Illinois Press, Urbana. Rhodes, G.(1985) Perceptual asymmetries in face recognition. Brain and Cognition,4,197-218. Sperry,R.W. (1968) Hemisphere deconnection and unity in conscious awareness. American Psychologist, 23, 723-733.