The research team recorded 552 online survey responses at a single institution and, after excluding ten pilot records and twenty-four responses submitted without consent, analyzed data from 518 medical students. The instrument contained four components: a twenty-item silhouette-recognition test in which students named structures shown only as outlines, a global difficulty rating, five items measuring anatomy-related spatial self-perception, and five exploratory items on learning preferences. The mean silhouette-recognition score was 12.10 out of 20, with a standard deviation of 3.33 and a range of 1 to 20, but individual item accuracy ranged from 25 percent to 96 percent.

That spread, the authors report, points to substantial variation in how recognizable individual structures are, with students repeatedly confusing visually similar structures. The psychometric properties of the new tool were mixed. The anatomy-related spatial self-perception measure showed strong internal consistency with a Cronbach's alpha of 0.899, and the five exploratory learning-preference items reached 0.821.

The silhouette test itself yielded a KR-20 of 0.666. When the researchers applied false-discovery-rate correction to control for multiple comparisons, only a handful of silhouette items retained statistically reliable differences across groups: the sphenoid bone, the internal carotid artery, the duodenum, and the larynx showed year-of-study differences, while the sphenoid bone also differed by age, all at p values of 0.05 or below. No silhouette item differed by sex.

Total scores, however, told a broader story. Recognition performance was significantly associated with year of study (η² = 0.035, p = 0.003) and, more strongly, with committee score (η² = 0.101, p < 0.001), indicating that students who performed better on formal anatomy assessments also tended to recognize silhouettes more accurately. Scores correlated positively with anatomy-related spatial self-perceptions (r = 0.312, p < 0.001) and negatively with perceived difficulty (r = −0.407, p < 0.001), but showed no association with sex or age.

The exploratory learning-preference items produced one of the study's more intriguing findings. Female students more strongly endorsed model- and cadaver-based learning (p = 0.034, V = 0.14) and peer-assisted learning (p = 0.011, V = 0.16) than their male counterparts, although the effect sizes were small. The authors frame these items as hypothesis-generating rather than as a basis for firm conclusions about how different groups learn anatomy best.

The study's conclusions are deliberately cautious. As a preliminary instrument-development effort, it identified recurrent confusion between visually similar structures and demonstrated that stimulus characteristics may contribute to item-level performance, a finding with implications for anyone designing anatomy assessments or visual learning materials. For now, the silhouette test stands as an early research instrument for measuring how students perceive the body's architecture in its barest form, with its ultimate value depending on the validation work still to come.

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

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

Article References: Ogut, E., Ulusahin, S., Soyer, N., & Efe, M. (2026). Development and preliminary evaluation of silhouette-based anatomy recognition performance in medical students: a cross sectional study. BMC Medical Education. https://doi.org/10.1186/s12909-026-10704-x

Image Credits: AI Generated

DOI: 10.1186/s12909-026-10704-x

Keywords: Development, preliminary, evaluation, silhouette-based, anatomy, recognition, performance, medical, students, cross, sectional, scientific research