Development and preliminary evaluation of silhouette-based anatomy recognition performance in medical students: a cross sectional study

We developed an anatomy silhouette recognition survey instrument and examined its association with academic variables and students’ self-perceptions of anatomy-related spatial competence. In this single-institution cross-sectional study, 552 online survey responses were recorded. After excluding 10 pilot records and 24 responses without consent, 518 medical students were included in the analysis. The instrument comprised a 20-item silhouette-recognition test, a global difficulty rating, five anatomy-related spatial self-perception items, and five exploratory learning preference items. We examined item responses, internal consistency, and factor structure using principal axis factoring with parallel analysis, group differences, and adjusted associations. The mean silhouette-recognition score was 12.10 (SD 3.33; range 1–20), and item accuracy ranged from 25% to 96%. Internal consistency was α = 0.899 for the anatomy-related spatial self-perception measure, α = 0.821 for the five exploratory learning-preference items, and KR-20 = 0.666 for the silhouette test. False-discovery-rate correction retained year-related differences for the sphenoid bone, internal carotid artery, duodenum, and larynx, and an age-related difference for the sphenoid bone ( p ≤ 0.05), whereas no silhouette item differed by sex. Female students more strongly endorsed model/cadaver-based ( p = 0.034, V = 0.14) and peer-assisted learning ( p = 0.011, V = 0.16). Total scores were associated with year of study (η² = 0.035, p = 0.003), committee score (η² = 0.101, p < 0.001), anatomy-related spatial self-perceptions ( r = 0.312, p < 0.001), and perceived difficulty ( r = − 0.407, p < 0.001), but not with sex or age. This preliminary instrument-development study identified substantial variation in item recognizability and recurrent confusion between visually similar structures, indicating that stimulus characteristics may contribute to item-level performance. These findings support the further refinement and validation of silhouette recognition as a complementary, group-level research approach in anatomy education before routine or individual-level use.