Background: Advanced medical simulators have been introduced to facilitate surgical and endoscopic training and thereby improving patient safety. Residents trained in a laparoscopic simulator, MIST, do laparoscopic cholecystectomy better and faster than a control group without simulator training. Little has been reported whether factors like gender, computer experience and visio-spatial tests can predict the performance in a medical simulator. Our aim was to investigate whether such factors may influence the performance of simulated gastrocopy. Methods: Seventeen medical students were asked about computer and gaming experiences. Prior to the virtual endoscopy they performed the visio-spatial test Picsor, which aquires the ability of the tested person to create a 3-D image from a 2-D presentation. Each student performed one gastroscopy (level 1 case 1) in the GI-mentor II, Simbionix ® and several varables related to performance were registred. Results: Percent of time spent with clear view in the endoscope correlated well with the performance in the Picsor test (r=0,56, p<0.001). Efficiency of screening also correlated with Picsor (r=0,23, p<0.05). In students with computer gaming experience the efficiency of screening increased (33.6 ± 3.1 % vs. 22.6 ± 2.8 %, p<0.05) and the duration of the examination decreased by 1.5 min (p<0.05). A similar trend was seen in men compared to women. Conclusion: The visio-spatial test Picsor predicts the results in the endoscopic simulator GI-mentor II. 2-D image experience, as in computer games, also seems to influence the outcome.