Medical students have difficulties in interpreting two-dimensional (2D) topographic anatomy on sectional images. Hands-on and no hands-on training in ultrasound imaging facilitate learning topographic anatomy. Hands-on training is linked with active search for patterns of anatomical structures and might train pattern recognition for image interpretation better although the added value on learning outcomes is unclear. This study explores first year medical students' knowledge in topographic anatomy of the upper abdomen after attending hands-on or no hands-on training in ultrasound in a randomized trial. While students in the hands-on ultrasound group (N = 21) generated and interpreted standardized planes of ultrasound imaging, students in the no hands-on seminar group (N = 22) interpreted provided ultrasound images by correlation to three-dimensional (3D) anatomical prosections. Afterwards knowledge in topographic anatomy was measured repetitively by text and ultrasound image-based multiple choice (MC) examinations. As surrogate for pattern recognition, students rated whether answers were known after reflection or instantly. While intrinsic motivation was higher in the ultrasound group, no differences in the MC-examination score were found between ultrasound and seminar group instantly (66.5 ±10.9% vs. 64.5% ±11.0%, P = 0.551) or six weeks (62.9% ±12.3% vs. 61.5% ±11.0%, P = 0.718) after training. In both groups scores in text-based questions declined (P < 0.001) while scores in image-based questions remained stable (P = 0.895) with time. After six weeks more image-based questions were instantly known in the hands-on ultrasound compared to seminar-group (28% ±17.3% vs. 16% ±13.5%, P = 0.047). Hands-on ultrasound-training is linked with faster interpreting of ultrasound images without loss in accuracy. The added value of hands-on training might be facilitation of pattern recognition.