Does spatial ability help the learning of anatomy in a biomedical science course?

INTRODUCTION

A three-dimensional appreciation of the anatomy of the human body is the cornerstone of clinical anatomy. Eventually all health professionals will be called upon to apply their anatomical knowledge to diagnostic or therapeutic clinical procedures. Such procedures may involve, for example, the accurate assessment of a suspicious mass palpated during an examination of the abdomen, insertion of an endotracheal tube or central venous access line, and the drainage of fluid from pleural or peritoneal cavities. Each of these procedures demands a thorough understanding of three-dimensional anatomy. Wanzel et al. (2002), Hegarty et al. (2006), and Clem et al. (2010) have all shown that spatial ability is an important skill for specialties such as surgery and ultrasonography. Students, themselves, believe that knowledge of spatial anatomy is important in specific specialties (Langlois et al., 2009).

In addition, spatial ability has been previously found in some contexts to be associated with students' ability to learn anatomy and their examination performance. (Rochford, 1985; Garg et al., 2001). Guillot et al. (2007) found that a high ability in tests of mental rotation (MRT) predicted success in learning anatomy as measured by examination results. More recently Lufler et al. (2012) found that students' spatial ability was associated with performance on practical examinations in a gross anatomy course. Of particular relevance to educators is the finding by Fernandez et al. (2011) that aspects of spatial ability can be enhanced through experience. This raises the possibility of providing learning resources for students to enhance spatial understanding, so as to in turn facilitate their learning of anatomy.

With the advent of medical curricular innovations such as problem-based learning, system-based and integrated clinical curricula, anatomical education has been the subject of major change over the last two decades (Drake, 1998; Dangerfield et al., 2000; Drake et al., 2009). Despite the fact that many anatomists and surgeons are adamant that dissection is the best way to learn clinical anatomy (e.g., Korf et al., 2008) the teaching of a modern anatomy curricula involves less (often no) dissection, fewer lectures, and greater use of web-based and computer-based resources (Sugand et al., 2010).

If spatial ability is of greater importance for below-average students as Rochford suggests, then this has important implications educationally. Therefore, our aims in this study are twofold, to investigate whether, in a modern biomedical sciences curriculum: (a) there is an association between spatial ability and achievement at learning anatomy and (b) below-average students with low spatial ability are at a particular disadvantage.

The curriculum investigated in this study is an under-graduate second-year level human structure and function subject at the University of Melbourne, Australia. It is a feeder into graduate courses for the health professions, including medicine and physiotherapy. As such, the subject is an introduction to clinical anatomy and does not make the same demands of students as would, say, a surgical trainee program. The anatomy component of this integrated subject covers the general principles of anatomy and the normal structure of the musculoskeletal, cardiorespiratory, gastrointestinal, urinary, and reproductive systems. Consistent with many modern anatomy curricula the material is presented in 30 one-hour lectures, 4 two-hour laboratory sessions using human material, and 8 self-directed computer-based tutorials. There is no dissection and no practical examination. The theory examination at the end of semester consists of multiple choice questions (MCQ) and written questions, some of which involve labeling of photographed specimens and diagrams.

METHODS

The 3DAT online spatial ability test (Sutton and Williams, 2011) was developed to assess the spatial skills of students in design-based disciplines such as engineering, but reflects common contemporary approaches to the assessment of generic spatial ability. It was developed in accordance with psychometric test development standards (“testing the test”) that include measures of various forms of validity and reliability and item analysis (Sutton, 2011). Test retest reliability for the 3DAT is 0.85. The 3DAT comprises of several tests of specific types of mental manipulations of images, as summarized in Table 1 and examples of each type of subtest are given in Figure 1.

Each of these factors is measured by the corresponding subtest given in Table 1, with each subtest consisting of four items. We chose to assess spatial ability across these five spatial factors to identify whether any specific spatial factor was more closely associated with success at learning anatomy. An initial pilot study suggested we use four items in each of the five subtests in order to balance timetabling demands against acceptable reliability.

As stated above, participants were undergraduate students studying a biomedical science subject that incorporated human structure and function including the general principles of anatomy. The study aimed to correlate performance on the test of spatial ability (3DAT) with scores on the final written anatomy examination. Ethics approval was obtained for the study from the University of Melbourne Human Research Ethics Committee.

Participants completed the selected subtests of the 3DAT website under controlled conditions in a student computer laboratory. Trials were scheduled over a number of sessions to accommodate participant availability and the limited number of computer workstations. Each participant attended only one session.

A correct selection was scored as a 1 (one) whereas an incorrect selection was scored as a 0 (zero). Hence, the maximum possible score on the 3DAT test was 20.

The final written examination for the second-year anatomy subject was administered in November of 2010 and consisted of three sections, Section A comprised all MCQ, section B comprised short-answer questions most of which involved labeling of images and diagrams, Section C comprised conventional short answer questions requiring descriptive written responses.

RESULTS

There were 127 participants; one participant did not complete the final anatomy examination and another participant's score on the anatomy examination was an outlier (>3 standard deviations from the mean) so their results were removed from the analyses. The following results are summarized in Table 2. Of the remaining 125 participants, 82 (65.6%) were female and 43 (34.4%) male. Of the 2,540 attempts made by the cohort of participants in the 3DAT system, only 19 were unanswered (Attempt = 0). Spatial ability data were initially analyzed using the SPSS Descriptives procedure, which reported the mean combined score (i.e., the sum of correct selections) for all the 3DAT subtests as 13.80 (± SD = 3.40) out of a possible 20 points, with the distribution slightly negatively skewed and platykurtic. Reliability was assessed across all 20 items resulting in a Cronbach's alpha of 0.68. The Shapiro-Wilk test for normality returned a significant result (P = 0.002) indicating that the data were not normally distributed. Females had a lower mean combined score (13.60, ± SD 3.40) compared with males (14.20, ± SD 3.30), however this was not statistically significant (Mann-Whitney U Test, P = 0.40). The only statistically significant difference found between genders was response time (RT) in the MC subtest, where females displayed shorter RTs; (mean = 25.40 sec, ± SD 11.10), males (mean = 31.28 sec, ± SD 14.60; independent samples Mann-Whitney U Test, P = 0.04).

The participants' scores on the anatomy examination had a mean of 88.50 (± SD = 12.80) out of a maximum possible score of 120, and the Shapiro-Wilk test for normality returned a significant result (P < 0.01) indicating the data were not normally distributed. The difference in mean scores between males (mean = 87.00, ± SD 14.10) and females (mean = 89.30, ± SD 12.10) was not significant (independent samples Mann-Whitney U Test, P = 0.344).

As the data were not normally distributed Spearman's rank-order correlation coefficient was used for investigating associations. The correlations between the students' spatial ability scores and examination results are given in Table 3. With an r = 0.145 and P = 0.106 on the total scores, it would appear that spatial ability and success on the anatomy examination were not highly associated. There were no significant associations found between any of the individual 3DAT subtests and the examination score or between individual sections of the anatomy examination (i.e., sections A, B, and C) and the total 3DAT score.

As mentioned previously, Rochford's (1985) study was based on under-achieving students. We replicated this focus on under-achievers by selecting a subsample for correlation based upon anatomy examination score. The spatial ability scores and anatomy examination scores were analyzed for those participants whose anatomy examination score placed them in the 25th percentile (examination raw score < 81.8, equivalent to < 68%). The results are given in the Table 3. The correlation between the 3DAT total score and the examination score was lower for the 25th percentile in the comparison to the whole cohort and not significant. There were no significant associations found between any of the individual 3DAT subtests and the examination score for under-achievers.

DISCUSSION

The results of our investigations suggest that spatial ability is not highly associated with success at learning anatomy in a biomedical sciences subject, as measured by scores on a written summative examination. Our analysis also did not find any support for the hypothesis, as suggested by Rochford, that below-average students with low spatial ability are at a particular disadvantage when learning anatomy.

Our results are in contrast with some of the studies previously mentioned, in which spatial ability was found to be positively associated with success at learning anatomy. However, we are not alone in finding a low association in this domain. For example, Hegarty et al. (2009) found that while spatial ability was ‘somewhat predictive’ of specific practical skills in dentistry, it was not associated with learning anatomy in general (in dentistry). Lischka and Gittler (1997) reported a similarly weak association (r = 0.11) between spatial ability (measured by performance on a “three-dimensional cubes test”) and learning anatomy. These authors did in fact categorize anatomy questions according to the “predominant requirement” of each anatomy question; namely knowledge, spatial, radiology, skeleton, relations, and plasticity. The correlation with spatial ability for each category was of a similar order to the overall relationship, ranging from 0.09 for skeleton to 0.15 for spatial (all values were significant except for the radiology category).

Together with our own results, these two studies paint a different picture of the relationship between spatial ability and learning anatomy from Rochford's, where two out of three “spatially oriented” anatomy examinations showed significant correlations with spatial ability (0.32 and 0.39, respectively; Rochford, 1985). The difference may well lie in the notion of spatial orientation of anatomy questions. As discussed under the limitations section, a key difference between our study and Rochford's is that we elected, ultimately, not to divide anatomy questions into “spatial” and “nonspatial” types (Rochford, 1985). We acknowledge that the effect of this decision would most likely have been to weaken any underlying association that may exist between spatial ability and learning anatomy. However, the study of Lischka and Gittler mitigates to some degree against this explanation (Lischka and Gittler, 1997). Further research in the way students respond to anatomy questions, including their use of spatial strategies or otherwise, would be a welcome addition to the literature on the role of spatial ability in learning anatomy.

It is of course possible that the low correlations in our study might in fact represent the reality that contemporary methods of Anatomy education, which rely more on models and digital images than cadaveric specimens, enable students of low spatial ability to successfully learn anatomy regardless of their spatial ability. This interpretation is possibly supported by the recent study of Lufler et al. (2012), who reported in this journal that spatial ability, as measured by a MRT, is in fact predictive of success in summative practical examinations in medical gross anatomy. From the description of the course and examinations it appears that their curriculum relied more heavily on dissection of cadavers and identification of structures in specimens than the bioscience subject of our study. Under such conditions, and also maybe in future curricula that include the interactive 3D simulations recently described by Trelease (2008) that require the user/learner to interact with simulated 3D objects to obtain information, the advantage conferred by good spatial ability seems entirely plausible and ought to be harnessed for pedagogical purposes, where appropriate. However, we would argue that care need to be taken in the choice of educational interventions and consequent demands that any spatially focused intervention places upon the student. This is especially relevant in light of recent work by Vorstenbosch et al. (2013) who, demonstrating a reciprocal advantageous effect between the study of anatomy and practice for spatial cognition, conclude that the study of anatomy is in itself beneficial for the students' spatial ability and no further intervention may therefore be required.

LIMITATIONS

Our intention was to categorize the questions in Section A of the final anatomy examination by degree of spatial content, to compare our results with Rochford's. However, despite agreeing on criteria and discussing anomalous decisions during the pilot, the authors felt unable to make this categorization consistently and/or confidently. Our experience of this process suggested that considerable speculation was involved in trying to determine whether a question was inherently “spatial” or otherwise, or which strategies a student might predominantly use to answer a question. We concluded that this depended considerably on the way the content had been presented to, or learned by, the students. Moreover, as Kyllonen et al. (1984) have shown, individuals frequently use different strategies when faced with the same spatial tasks. We therefore felt it was desirable to minimize one potential source of confounding and leave as tacit the particular strategy used by students when answering anatomy questions.

The 3-DAT is relatively new and, although its validity was established during its development (Sutton, 2011), it has not had the revalidation that more established tests such as Vandenberg and Kuse (1978) have enjoyed from repeated use by a wide variety of researchers over an extended period. As such, the results and conclusions of this study must be viewed in this light.

CONCLUSION

In the context of this biomedical science subject it appears that while spatial ability may be a significant factor at more advanced levels of health sciences education, such as graduate medical programs, training of surgeons or in other contexts such as dental science or ultrasonography, it may not be so important in undergraduate contexts where the predominant modes of anatomy education are less experiential than traditional cadaveric dissection. In such contexts, our study seems to suggest that students of lower spatial ability are not particularly disadvantaged when it comes to learning anatomy, and are perhaps best counseled to focus on learning anatomy content rather than using that time to improve their spatial ability.

NOTES ON CONTRIBUTORS

KEVIN SWEENEY, B. Pharm., Grad. Dip. (Data Proc.), M.I.T., is an honorary fellow in the Medical Education Unit of the Melbourne Medical School, Faculty of Medicine, Dentistry and Health Sciences, University of Melbourne, Melbourne, Australia. His research interest is in spatial thinking in health sciences education.

JENNIFER A. HAYES, M.B.B.S., is an associate professor in the Department of Anatomy and Neuroscience, Faculty of Medicine, Dentistry and Health Sciences, University of Melbourne, Melbourne, Australia. She teaches undergraduate and postgraduate students and coordinates the medical anatomy program.

NEVILLE CHIAVAROLI, B. App. Sc. (Phty), B. A. Hons., M. Phil., M. Ed., is a senior lecturer in the Medical Education Unit of the Melbourne Medical School, Faculty of Medicine, Dentistry and Health Sciences, University of Melbourne, Melbourne, Australia. His research interest is in assessment practices in medical education and the medical humanities.