This research is an effort to best utilize the interactive anatomical images for instructional purposes based on cognitive load theory. Three studies explored the differential effects of three computer-based instructional strategies that use anatomical cross-sections to enhance the interpretation of radiological images. These strategies include: (1) cross-sectional images of the head that can be superimposed on radiological images, (2) transparent highlighting of anatomical structures in radiological images, and (3) cross-sectional images of the head with radiological images presented side-by-side. Data collected included: (1) time spent on instruction and on solving test questions, (2) mental effort during instruction and test, and (3) students' performance to identify anatomical structures in radiological images. Participants were 28 freshmen medical students (15 males and 13 females) and 208 biology students (190 females and 18 males). All studies used posttest-only control group design, and the collected data were analyzed by either "t" test or ANOVA. In self-directed computer-based environments, the strategies that used cross sections to improve students' ability to recognize anatomic structures in radiological images showed no significant positive effects. However, when increasing the complexity of the instructional materials, cross-sectional images imposed a higher cognitive load, as indicated by higher investment of mental effort. There is not enough evidence to claim that the simultaneous combination of cross sections and radiological images has no effect on the identification of anatomical structures in radiological images for novices. Further research that control for students' learning and cognitive style is needed to reach an informative conclusion. (Contains 4 tables and 4 figures.)