Historically, learning anatomical specimens was limited to studying cadaveric materials, and by extension, specimens which are “life‐sized”. Recent technological advancements in 3D scanning and printing now allow for the production of inexpensive, durable anatomical replicas at virtually any size. This, however, creates a dilemma: what is the most effective model size to learn from? The goal of this project is to discover the appropriate size of an object to learn nominal anatomy and thus provide a critical step in improving anatomic education.We hypothesize that there is a curvilinear relationship between model size and learning, where a model too small or too big would not be the most conducive to learning and an ideal intermediate size can be determined. In this study, undergraduate students (n = 351) without prior anatomical training learned from four bones of varying normal anatomical size and features and were assessed on their ability to identify various landmarks. Thoracic vertebra (VE), hemipelvis (HE), sphenoid (SP), and scapula (SC) was 3D‐printed at four different scalar sizes. The VE and HE models were printed at 50%, 100%, 200%, and 400% scale, while SP and SC models were printed at 50%, 100%, 200%, and 300% scale. Each participant was randomly assigned to a group of two bone models (VE/HE or SP/SC) of a certain size, and randomized across the order in which they learned the models. They were then tested on the respective real bone specimens, followed by a qualitative survey reporting their experience with the 3D‐printed models, a Mental Rotations Test (MRT), and an Operation Span Test (OSPAN). Data collection for the 50% SP/SC group is still ongoing.Multiple regression suggested significant effects of Model Type, Model Size, MRT and OSPAN, (F(9, 596) = 17.96, p = 0.000, R2 = 0.2133). The most significant predictor of test score was MRT, which suggested a 10% increase in MRT score is associated with a ~3% increase in test score. The score variability independently accounted for by MRT and OSPAN was 14.6%, while the variability independently accounted for by model size and type was 7.8%. This means, that while test scores are primarily driven by participants’ mental rotation ability, model size remains an important feature that can be manipulated to improve learning. 3D printing allows for this in a cost‐effective way.Support or Funding InformationThis study was funded by the Education Program in Anatomy at McMaster University. Many thanks to the University of Buffalo for 3D printing the 400% VE and HE models.
Read full abstract