By Kevin Niall Dunbar, Ph.D., University of Toronto at Scarborough
The goal of our research was to determine whether there are cognitive differences between performing arts and non-performing arts students, and to discover what the brain-based differences are that underlie the cognitive differences. We compared performing arts students in theater and music to students who are not involved in the performing arts. We investigated performance on a variety of reasoning tasks and investigated whether there were differences in patterns of brain activity of the students. We used functional Magnetic Resonance Imaging (fMRI) to investigate this question.
fMRI makes it possible to monitor the activity of the brain. The basic idea of fMRI is that regions of the brain that are being used in a specific mental activity will use more oxygen than regions of the brain that are not used in the task, and fMRI is sensitive to changes in oxygen uptake in the brain. Thus, fMRI is an index of how active specific regions of the brain are in particular tasks.
The hypothesis being tested in our Dana-funded research is that if performing arts students are using their brains in a different way from the non-performing arts students, then we should see differences in brain activation patterns between these different groups. It may be the case that specific regions will show increased activation in the performing arts students and that other regions may show decreased activation. Of course it may be that we see no differences between the performing arts students and non-performing arts students in the tasks that we used. The goal of the research was to obtain first ever information on this important issue and we began the research with open expectations as to what we might discover.
The focus of our research was twofold. The first was to determine the types of brain-based changes that occur as a function of being exposed to a performing arts education. The second was to postulate the brain-based mechanisms that might lead to these improvements. This second aim is particularly important since previous attempts to investigate whether the arts have important effects on the brain have not posited specific neural mechanisms that might be involved. Thus, the goal of the research was to go beyond vague and general claims about the effects of a performing arts education to testing specific hypotheses.
In particular, the research conducted in the Dunbar laboratory was to study the effects of the performing arts on key cognitive processes involved in reasoning, such as generating novel and creative concepts, and being able to map information from one context to another, very different, context. This is known as transfer. The overarching question here is whether abilities that are acquired in the performing arts will transfer to other domains. Our goal was to investigate whether education in the performing arts influences abstract reasoning ability.
We investigated whether students in theater and music reasoned differently from non-arts students and what the brain-based changes that underlie these differences might be. Another question that we investigated is whether arts students are intrinsically different from non-arts students. We investigated this question by comparing performing-arts students’ genetic makeup to non-performing arts students genetic makeup. This allowed us to investigate the possibility that differences between performing arts students and non-performing arts students’ might be due to underlying genetic differences that predispose students to prefer and choose the performing arts rather than other areas of education.
Year 1 Research Summary
In the first research year, we conducted an fMRI (functional magnetic resonance imaging) investigation, comparing brain activity of theater students with non-theater students as they worked on a new variant of a widely used task of creative thinking. One common claim about the effects of training in the performing arts is that it fosters creativity. This task—The Uses of Objects task- provides a measure of divergent thinking. This task is deceptively simple: particpants are given the name of an object, such as a brick, and are asked to generate as many uses for the object as they can think of. Previous researchers have found that this task is a useful measure of creative thinking (e.g., Carlsson, Wendt, & Risberg, 2000). Furthermore, researchers have found that arts students perform differently from non-arts students; they produce more varied and creative uses than students not in the arts (see Guilford, 1967 and Hudson, 1967 for the classic study on creativity and the arts). Postdoctoral student Dr. Jinathan Fugelsang, graduate student Adam Green, and undergraduate student Raphael Lizcano were actively involved in this research.
In our research, we investigated changes in creativity using a standardized test that measures the ability to reason creatively: or the Uses of Objects task. We hypothesized that training in the performing arts should lead to increased performance on the Uses of Objects task. We investigated the Uses of Objects task in arts and non-arts students. In year 1, we focused on using Guilford’s Uses of Objects test (Guilford, 1967), which we adapted for use in fMRI. In the original version of this task, participants must generate as many uses for an object as they can. Participants who generate unusual items are called divergers and those who generate standard uses are called convergers. We modified this task for the fMRI experiment that compared performing arts students with non-performing arts students. We used photographs of objects to present stimuli to the participants and we devised a button-pressing technique for responses that eliminated the need for writing a response. Of course we compared our new task to the more standard task and found good correlations between performance on the two different versions of the task. Our hypothesis was that performing arts students would generate more divergent answers than non-performing arts students and that the different groups would display different patterns of brain activation. This is a sensitive task that has been used in assessing creativity in many domains and occupations.
This first-year research provides the first fMRI comparison ever undertaken of performing arts students with non-performing arts students. We analyzed the task-related BOLD (blood oxygenation level dependent) fMRI imaging response for conditions in which participants were judging the Uses of Objects minus the materials judgments. As can be seen from Figures 1a and 1b, we found differences in activation levels between the performing arts and non-performing arts groups. Regions in the medial fusiform gyri were selectively activated for the Uses of Objects task for the non-performing arts majors and not for the performing arts majors. Performing arts majors showed increased activation in two frontal areas: the left inferior frontal gyrus and in the left superior frontal gyrus. Prior work on word generation tasks has demonstrated that the left inferior frontal gyrus is involved in generating names in language processing tasks. The finding that performing arts students have increased activation in this area suggests that they are taking a more linguistic approach to the task, whereas the non-performing arts students are taking a more perceptual approach to the task. This result is important, as our other research on scientific thinking and expertise indicates that expertise can lead to increased activation in linguistic areas that are associated with conceptual thinking (Dunbar and Nelson, in preparation).
Our work on relational reasoning indicated that a region of the prefrontal cortex, frontal polar cortex, should show increased activation in the Uses of Objects task, and we therefore conducted a region of interest analysis of the frontal poles for the Uses of Objects task. We indeed found increased activation in the frontal poles (See also, Green, Fugelsang, Kraemer, Lizcano, & Dunbar, in preparation). However, there were no significant differences between the performing arts students and our controls in activation levels. Thus, while we have found differences between performing arts and controls, these differences appear to be in areas of the brain associated with linguistic symbolic retrieval from memory, rather than mapping from one semantic domain to another. This finding is important as it indicates a more precise locus for the effects of performing arts expertise than has been previously offered.
|(all images courtesy of Dunbar Lab)|
Year 2 Research Summary
Research in Year 2 concentrated on three main levels, prompted by the meeting of the consortium investigators in fall 2005. A goal of the research was to use many of the same tasks that other laboratories were using. Thus, we developed a battery of tasks that measure performance on attention, working memory, and reasoning that could be administered within a 1-hour period. Postdoctoral fellow James Nelson, graduate student Adam Green, and undergraduates Raphael Lizcano and Oya Nuzumlali were involved in the development of these tasks.
In a second level, a key point that arose at the consortium meeting was the need to tease apart, at multiple levels, the effects of a performing arts education on the brain. The multiple levels approach is particularly important when adults are being studied, since our adult performing arts participants may have self-selected into performing arts; differences between performing arts students and non-performing arts students may be due, therefore, to underlying trait differences rather than to the effects of a performing arts education per se. This is a centuries old question for this type of research. The problem is that if a researcher finds differences between students in the performing arts and non-performing arts students it may be due to the training in the performing arts, or to inherent genetic differences. Teasing apart these two types of reasons is notoriously difficult. One way of addressing this centuries-old conundrum is to have a measure of student’s underlying genetic traits. Dr. Posner has espoused this view at the Dana meetings, and this is an important way of determining whether any differences between performing arts students and non-performing arts students is due to underlying genetic differences or to the effects of a performing arts education. With this goal in mind, my graduate student Adam Green spent the winter term of 2006 at the Sackler Institute in New York learning to conduct DNA genotyping using buccal swabs (i.e., saliva). This is a relatively inexpensive and easy to obtain technique that we used in our 2006-2007 research. Rather than postulating a genetic or an environmental approach, a third possible hypothesis is that an interaction between environmental and genetic mechanisms may be at the root of differences or between performing arts and non-performing arts students. We are currently developing a DNA-microarray technology to address this question. Using this technique we are able to determine how training in a domain influences the expression of a wide variety of genes in the human genome. Dunbar & Petitto, (in preparation).
The third component of our Year 2 research was to develop tasks that we could use to assess differences, through imaging, in brain activities of performing arts and non-performing arts students. We refined our analogical reasoning tasks and focused on the frontopolar cortex as a key area for region of interest analyses. We also refined and developed the Uses of Objects task that we have been using to assess differences between performing arts students and non-performing arts students. The new Uses of Objects task was a categorization task, in which participants assess how good each of various uses is, for a particular object. This new modified Uses of Objects task provides more data and decreased error variance in the data, resulting in a more stable measure of divergent thinking. Dartmouth acquired a new 3.0 Tesla fMRI scanner during our second research year, and in the interim before the scanner became operational, we developed our DNA genotyping skills, the cognitive test battery, and the fMRI tasks to be used in our third research year.
Year 3 Research Summary
The goal of the third year research was to investigate differences between students studying theater compared to music, and between performing arts students and non-performing arts students. We identified a group of students in each of these categories. Based upon our prior years’ research, we developed a verbal version of the Uses of Objects task that could be administered in an fMRI experiment, and developed a fine-grained version of our analogy generation task that could be conducted in a neuroimaging context. Based on our first two years of research, we predicted that we would observe differences between our groups of participants at both the behavioral and neural levels and could also determine whether differences between the groups were due to inherent genetic differences, or experience with a particular performing arts field. Postdoctoral fellow James Nelson, and graduate student Adam Green, were actively involved in this research. The third year research was conducted at three main levels:
Administering a cognitive test battery to the performing arts and non-performing arts students. The battery consisted of working memory tasks (backwards and forwards), analogy task, Uses of Objects task, language proficiency task, and a general background expertise questionnaire (based in Ericsson’s work and also similar to those used by Petitto and Jonides). We tested more than 60 students (3 groups of 20 students; performing arts theater, performing arts music, and non-performing arts students).
Obtaining buccal swabs (i.e., saliva) from this set of 60 students for DNA genotyping. The DNA genotyping focused on a small number of polymorphisms (small changes in sequence) of key genes that are markers of cognitive differences (DRD4, DRD5, MAOA, COMT, 5-HTTLPR, SNAP-25, 5HT1B). The goal of this part of the research was to determine if the performing arts students differ in terms of genes known to be related to the types of psychological traits—such as attentional control and temperament—that are hypothesized to be important in performing arts students.
Conducting an fMRI study of students that we have both genotyped and tested behaviorally on our test battery. We imaged 30 students (10 in each of our 3 groups). The two tasks that we used in the 2006-2007 year were our modified Uses of Objects task and our analogical reasoning task. We hypothesized that we will see differences in activation in key brain areas, such as the frontopolar cortex, and anterior cingulate cortex.
A total of 60 participants completed the behavioral session and had buccal swabs taken for genetic analysis. Twenty participants (10 male, 10 female) were in each of three subgroups—those with theater training (average age 20.9, range 18-25), those with music training (20.1, 18-22), and those with no or minimal performing arts training (20.6, 19-22). Performing arts students were all receiving professional training, and had at least 4 years of such training and 10 total years of experience; non-performing arts student were not currently receiving training or practicing and had not had professional training, at least from age 14 to the time of the study.
In a behavioral testing session, all participants completed the following tasks: Shipley Vocabulary (Shipley, 1940), WAIS-III Digit Span Forward and Digit Span Backward sub-tests (Wechsler 1997), an in-lab multiple choice Analogy Completion task, the Group Embedded Figures Task (O’Leary, et al. 1980), a 4-item written version of the Uses of Objects task (Hudson, 1967), and a questionnaire verifying and detailing their educational background. These tests were given in rotating order; the questionnaire was always given last. At the end of the session, participants gave a buccal swab sample for genetic analysis, described later.
Two planned t-tests were done for each test score: a comparison of performing arts (PA) students vs. non-PA students, and a comparison of music vs. theater PA students. While there were no significant differences between PA and non-PA students, music students had significantly better performance on the Backward Digit Span task than theater students, and theater students had a strong statistical trend towards higher divergence scores on the Uses of Objects task (number of non-standard uses listed for all objects). See Table 1. Participants with music training showed greater ability to manipulate items in working memory, while those with theater training showed a tendency towards more divergent thinking.
|Digit Span: Backward||N||Score||(se)||P.A. vs. Non P.A.||Music vs. Theater|
|Theater||20||7.95||.484|| || |
|Non P.A.||20||8.2||.574|| || |
Table 1. Structure of the Aesthetics Questionnaire
|Divergence Score||N||Score||(se)||P.A. vs. Non P.A.||Music vs. Theater|
|Theater||20||26.8||2.598|| || |
|Music||20||19.05||2.925||n.s.||T>M, n.s., p=.055|
|Non P.A.||20||24.5||2.220|| || |
DNA Genotyping Research
Buccal samples are being analyzed for genetic polymorphisms on a sample of genes thought to be associated with attention and cognition (e.g. Fosella, et al. 2002; Fan, et al. 2003) or to influence participation in a performing art (Bachner-Melman, et al., 2005). While analysis is still incomplete (23 participants have not been analyzed), initial results show interesting trends. First, for COMT position 158, the homozygous val/val genotype is prominent in our non-performing arts students (7/12), but relatively rare in performing arts students (3/25). COMT also may be related to scores of divergent thinking from the Uses of Objects task, as study participants who are heterozygous val/met show the lowest mean divergence score (n.s., p<.10). MAOA also shows trends (p<.10) for predicting performance on the Backwards Digit Span task for both Exon 14 and for promoter region polymorphisms.
Divergent and Analogical Thinking
Nine non-PA students, eight music students, and 11 theater students underwent an fMRI session, which alternated two blocks each of an analogy verification task and a simplified Uses of Objects task. Seven participants from each group were entered into the final analysis of the Uses of Objects task (others were eliminated due to excessive movement, or due to artifacts in the MRI data). In this version of the Uses of Objects task, students were given a pair of words, and evaluated whether the first-named object could be used as the second-named object (e.g., Hose, Rope), and rated the object on a 1 to 3 scale (1= “Good use”, 2 = “Fair Use”, 3 = “Poor Use”). Pairs of words were chosen, based on pilot data, to provide a range of “good” to “poor” uses. Sixty pairs of objects were evaluated, with 4 seconds allotted for each judgment. Thirty fixation trials of 4 seconds length were randomly inserted (with no more than three fixation trials in a row).
For the analogy verification task, participants completed 60 trials, split evenly between valid and invalid analogies, and cross-domain versus within-domain analogies. Six seconds were allotted for the participant to rate each four-item analogy For this analogy task, at the time of writing, only six individuals had been fully analyzed, an insufficient number for meaningful comparison based on group membership or performance.
However, the analyses for the Uses of Objects task are complete. The results comparing the Uses task to fixation trials across all participants are shown in figure 2. Red regions are significant at a False Detection Rate of p<.05 and include bilateral PFC, left parietal, left inferior temporal brain regions, and cerebellar activations. The yellow areas were significant at the much more stringent threshold of Family-wise Error p<.05. Prominent is a peak activation in left dorsolateral prefrontal cortex in Brodmann’s Areas 9 and 46 (peak voxel at MNI coordinates x=-56, Y=12, Z=28, 2.36 cubic centimeters volume).
Whole brain contrasts reveal no differences among differing performing arts backgrounds, but a median split of high vs. low divergence scores shows bilateral activation (uncorrected p<.001, >5 contiguous voxels), with right hemisphere activation dominated by an activation in inferior parietal cortex near the right temporal-parietal junction in Brodmann’s areas 40 and 43 (peak voxel at MNI 50 -24 16, 1.54 cubic cm volume), and left hemisphere activation scattered among several smaller activations surrounding the left temporal parietal junction.
Further ROI (region of interest) analyses were done on the left DLPFC area identified in the whole brain Task vs. Fixation contrast, and also an area of anterior frontal cortex identified by Green, et al., (2006) that is active during abstract analogical reasoning (MNI:-8, 60, 31; 10mm radius sphere). However, no significant group differences were found in these frontal ROIs, nor were differences found between high and low divergent subjects (figure 3). This difference in results may be due to the fact that we switched from a version of the task where students were asked to generate Uses of Objects to a version of the task where they assessed how well a particular use fits an object.
Year 3 Research Summary and Discussion of Results
The results of our behavioral tasks indicate significant working memory differences between the music students and our other two groups of students (theater and controls). In addition, our DNA genotyping results may indicate that there are genetic differences (in MAOA and COMT) between our performing arts students and the controls. We have already collected DNA from a further set of participants that would allow us to determine whether these differences are statistically significant. Unlike our year 1 results, our year 2 results did not reveal differences between students’ brain activity for the Uses of Objects task. We suspect that the differences in our task between the two years are responsible for this outcome. Our analyses of the fMRI data did not reveal differences between the performing arts students and controls. However, our fMRI analyses did reveal significant differences between students who scored high on our test of creativity and those who scored low.
Summary Comments and Future Directions
The research conducted by the Dunbar laboratory was concerned with investigating whether there were differences between performing arts students (music and theater students) and non performing arts students. Four main routes were taken to investigate this question. First, we used standardized behavioral tasks such as digit span and the group embedded figures task. Second, we used behavioral tasks that tap the abstract reasoning processes thought to be stimulated by a performing arts education (generation of object uses and generation of analogies). Third, we conducted neuroimaging, using fMRI of students as they performed our behavioral tasks. Fourth, we conducted genetic analyses (DNA genotyping) of our different groups of participants. Overall, we found some differences at both the behavioral and genetic level. At the neural level we found differences between the performing arts students and non-performing arts students in our first year.
In particular we found differences in left hemisphere frontal lobe activation that are consistent with the hypothesis that the performing arts students are more likely to be engaged in the symbolic retrieval than non-performing arts students. When we modified our Uses of Objects task for our year three research, we found no differences in brain activation between the performing arts students and the non-performing arts students. This indicates that it is in the generation of novel ideas, and not the responding to novel ideas, that is the key difference between the two groups.
We plan to analyze the remaining genetic data collected to more definitively determine whether there are differences among our groups at the genetic level. A second avenue would be to increase our sample size, as our sub-group sample sizes are still quite small. A third avenue is to administer the same tasks to a more diverse population in which participants have a wider range of SAT scores and a more diverse socio-economic background. A fourth approach would be to conduct these studies of tasks on a younger population of participants, to determine whether there is a “sensitive period” for the acquisition of the mental processes involved in the performing arts. Finally, we are in the process of developing an epigenetic approach that will allow us to further deliniate potential differences between the performing arts and non-performing arts students.
back to top
Bachner-Melman, R., Dina, C., Zohar, A.H., Constantini, N., Lerer, et al. (2005). AVPR1a and SLC6A4 Gene Polymorphisms Are Associated with Creative Dance Performance. PLOS Genetics, 1(3): e42.
Carlsson, I., Wendt, P., & Risberg, J. (2000). On the neurobiology of creativity:Differences in frontal lobe activity between high and low creative subjects. Neuropychologia, 38, 873-885.
Dunbar, K. N., & Nelson, J. K. (In preparation). Brain based mechanisms underlying the development of expertise in Science.
Dunbar, K.N. & Petitto, L.A. (In preparation). Molecular epigenisis: A new microarray based method for determining the interactions of genes and environment.
Fan, J., Fosella, J., Sommer ,T., Wu Y., Posner, M.I. (2003). Mapping the genetic variation of executive attention onto brain activity. Proceedings of the National Academy of Science, USA, 100, 7406-7411.
Fosella, J., Sommer, T., Fan, J., Wu, Y., Swanson, J.M., Pfaff, D.W., Posner, M.I. (2002). Assessing the molecular genetics of attention networks. BMC Neuroscience, 3:14.
Guilford, J.P. (1967). The Nature of Human Intelligence. New York: McGraw-Hill.
Hudson, L. (1967). Contrary Imaginations, Penguin Books, Harmondsworth, UK.
Green, A., Fugelsang, J., Kraemer, D.J.M., Shamosh, N., & Dunbar, K. (2006). Frontopolar cortex mediates abstract integration in analogy. Brain Research, 1096, 125-137.
Green, A., Kraemer, D.J.M., Fugelsang, J., & Dunbar, K. Frontopolar cortex becomes more active for more abstract analogies. Poster presented at the Cognitive Neuroscience Society (2006).
Green, A., Kraemer, D.J.M, Fugelsang, J., Dunbar, K. Frontopolar cortex and creative thinking: An fMRI study of the Uses-of-Objects task. Poster presented at the Cognitive Neuroscience Society (2007).
Hudson, L. (1967). Contrary Imaginations: a psychological study of the English Schoolboy Harmondsworth: Penguin.
Ishai A, Ungerleider, L.G., Martin, A., & Haxby JV. (2000). Representation of objects in the human occipital and temporal cortex. Journal of Cognitive Neuroscience, 12, 35-51.
Martin, A., Chao, L.L. (2001). Semantic memory and the brain: structure and processes. Current Opinions in Neurobiology, 11, 194-201.
O’Leary, M.R., Calsyn, D.A., Fauria, T. (1980) The Group Embedded Figures Test: a measure of cognitive style or cognitive impairment. Journal of Personality Assessment, 44, 532-537.
Shipley, W C (1940). A self-administering scale for measuring intellectual impairment and deterioration. The Journal of Psychology 9, 371-377.
Wechsler, D.A. (1997). Wechsler Adult Intelligence Scale (3rd ed.). San Antonio, TX: The Psychological Corporation.
back to top