Abstract Purpose Simulators are increasingly used in the training of endovascular procedures; however, for the use of the Mentice vascular interventional system trainer (VIST) simulator in neuroradiology, the validity of the method has not yet been proven. The study was carried out to test the construct validity of such a simulator by demonstrating differences between beginner and expert neurointerventionalists and to evaluate whether a training effect can be demonstrated in repeated cases for different levels of experience. Methods In this study 4 experts and 6 beginners performed 10 diagnostic angiographies on the VIST simulator (Mentice AB, Gothenburg, Sweden). Of the cases four were non-recurring, whereas three were repeated once and ten subjects performed all tasks. Additionally, another expert performed only five non-recurring cases. The simulator recorded total time, fluoroscopy time, amount of contrast medium and number of material changes. Furthermore, gaze direction and heart rate were recorded, and subjects completed a questionnaire on workload. Results Beginners and experts showed significant differences in total duration time, fluoroscopy time and amount of contrast agent (all p < 0.05). Gaze direction, dwell time and heart rate were similar between both groups. Only beginners improved during training with respect to total duration time, fluoroscopy time and amount of contrast agent. If a case was previously known to them, the total duration and fluoroscopy time were significantly shortened (p < 0.001). Conclusion This study demonstrated both the construct validity of a diagnostic neuroangiography simulator as well as a significant training effect for beginners. Therefore, in particular beginner neurointerventionalists should use such simulation tools more extensively in their initial training.