Is there any real virtue of virtual reality?: the minor role of multiple orientations in learning anatomy from computers.

Abstract

Accurate three-dimensional (3D) computer anatomical models have been proliferating as a consequence of the availability of digital resources derived either from mapping the surface of real objects (such as the human skeleton) or by tracing out regions of interest from images of serial slices of the human body. Such models would appear to have obvious educational advantages over the standard three-view book presentation, since the learner can control the position of the object just as if it were a real 3D object in his or her hands. Surprisingly, these advantages may be more imagined than real. Implicitly, these models assume that the learner can accurately assimilate and remember spatial information from multiple viewpoints. However, some evidence suggests that humans actually synthesize spatial information presented in oblique orientations by first rotating back to a standard (top, side, front) view, then learning the visual information. 1 If so, presenting information in multiple orientations may place a heavy load on the individual’s ability to rotate the figure. This load will be seen most acutely in those with relatively poorer spatial ability, since mental rotation is a critical element of spatial ability. 2 Consequently, studying projections of a 3D object in an orientation in which an object is best visualized and least obscured (called canonical or key viewpoints) may be all that is necessary. Once learned, the object in memory can be mentally rotated to enable recognition of new objects. This hypothesis was borne out in the first of a series of studies, 3 which contrasted one group where wrist-bone anatomy was presented in multiple views rotating at 157 increments every 20 seconds (the multiple-view group) with a second group that saw only the palmar and dorsal, posterior/anterior (p/a) views for the same time (the key-view group). The test was based on presentations of a skin-covered wrist in rotated positions, with an arrow pointing to a particular bone that was to be identified. All the rotated positions had been seen by the multiple-view group during learning. The multiple-view presentation had a small benefit for learners with good spatial ability, but it substantially handicapped learners with poor spatial ability. Anecdotal evidence from this study showed that participants in the multiple-view group were using a strategy of memorizing key views; 88% reported that they first mentally rotated the image to a standard position, then memorized the names. To confirm the self-reported strategy of mental rotation and memorizing key views empirically, a second study 4 allowed students in one group to control the presentation through multiple views, while those in the control group were restricted to controlling the object in either of the p/a views. Consistent with the post-hoc observations of the first study, learners spent most of their time examining the anterior and posterior views, with small variation around these key views. In this study, there was a small advantage for the multiple-view presentation, although the strategy adopted by the group amounted to spending most of the time on key views, with a significant but small variation around the 07 and 1807 presentations. Although this might be interpreted as evidence of the superiority of active learning, an alternative explanation is that, consistent with theory, participants extracted most information from the key views, but deliberately induced a small ‘‘wiggle’’ around the key view to gain some sense of the third dimension. Both these studies are notable in that they controlled for many of the confounding variables that plague studies comparing media. All presentations and tests were done on the computer for both groups, time of presentation was controlled, adjustment was made for spatial ability, and the test was consistent with the instruction. They lead to two conclusions: (1) When presented in a fixed sequence, multiple views had a small advantage over p/a views for high-spatial-ability students, but a disadvantage for low-ability students. This conclusion is consistent with the self-reports that, when presented with an oblique orientation, students mentally rotate back to the canonical form, with a consequent increase in cognitive load and reduction of performance. (2) When learners can control the presentation, they consistently perform better than do students who have only p/a views, but actually spend most of the time at or near the standard orientations. What is unclear from the superficially opposite conclusions of the two studies is whether the superiority of the multiple-views/ active-control group in the second study derives from active learning, or whether students actually gained information from ‘‘wiggling’’ the views around the canonical presentations to detect the depths of the objects, while not losing the straight-on presentation. The present study addressed this question by permitting both groups to have active control over the orientation, but in one group, it was restricted to 6107 around the key views; in the other, it is unconstrained. Method