OBJECTIVES: Virtual reality (VR) simulation has been suggested to objectively assess endovascular skills. The aim of this study was to determine the impact of cognitive training on technical performance of inexperienced subjects on a commercially available VR simulator (VIST, Vascular Intervention Simulation Trainer, Mentice, Gothenburg, Sweden). METHODS: Forty-seven subjects treated an identical virtual iliac artery stenosis endovascularly. Surgical trainees without endovascular experience were allocated to two training protocols: group A(1) (n = 10) received a 45 minute didactic session followed by an expert demonstration of the procedure that included error-based learning, whereas group A(2) (n = 10) was only given a demonstration of an iliac dilation and stent procedure. All trainees performed the intervention immediately following the expert demonstration. Twenty-seven endovascular physicians were recruited (>100 endovascular interventions). Performance was assessed using the quantitative (procedure and fluoroscopy time) and qualitative (stent/vessel ratio and residual stenosis) assessment parameters recorded by the simulator. RESULTS: The end-product (qualitative metrics) in the cognitive-skills group A(1) was similar to those of the endovascular physicians, though A(2) performed significantly worse than the physicians (group B): stent/vessel ratio (A(1) 0.89 vs B 0.96, P = .960; A(2) 0.66 vs B 0.96, P = .001) and residual stenosis (A(1) 11 vs B 4%, P = .511; A(2) 35 vs B 4%, P < .001). Group A(1) took longer to perform the procedure (A(1) 982 vs B 441 seconds, P < .001), with greater use of fluoroscopy than group B (A(1) 609 vs B 189 seconds, P < .001) whereas group A(2) performed the intervention as quickly as group B (A(2) 358 vs B 441 seconds, P = .192) but used less fluoroscopy (A(2) 120 vs 189 seconds, P = .002). CONCLUSION: Cognitive-skills training significantly improves the quality of end-product on a VR endovascular simulator, and is fundamental prior to assessment of inexperienced subjects.