Endovascular surgical procedures require visual-spatial coordination in workspaces with restricted motions and temporally limited imaging. The development of the skills needed for these procedures can be facilitated by 3D simulator-based training. Cerebral angiography (CA) has lagged behind in this training approach due to the lack of validated, realistic training models, relying strictly on clinical case exposure frequency ("number of hours logged") as a means of assessing proficiency. The ANGIO Mentor visual-haptic simulator is regarded as an effective training tool, however, this simulator has not been tested thoroughly in its ability to train interventional skills. In particular, the details of the aneurysm coiling process during simulation-based training have not been assessed. In this study, 12 novice medical students were given simulation-based diagnostic CA training until a procedural plateau in performance, established in our previous work. Subsequently, they were trained using video tutorials and written instructions to identify, measure and intervene with cerebral aneurysms using endovascular coils. Over the span of 6 sessions, participants were assessed on their procedural task time, coiling quantity and quality, and perforation rates. Prior to commencing the study, participant spatial ability was assessed using a mental rotation test (MRT) and used as a comparative baseline for the performance analysis. We found that all individuals were able to perform the procedure faster after 6 sessions, reducing their average time from 42 to 24 minutes. Coil success rate improved over from 82% to 88% and coil packing rate remained consistent at 30% throughout testing. High perforation rate seen at the start of the study showed a trend of decreasing over the latter sessions, however, over half of aneurysms were still being perforated by the novice participants. No change in aneurysm coiling quality was found, with a slight decrease in number of parent artery coil protrusions. High MRT individuals were better able to establish necessary tools prior to coiling, however, no other MRT-specific changes were seen. This work identifies the utility of simulation-based CA training in identifying the particular difficulties trainees experience in learning procedural skills, including prevention of perforations, proper positioning and success of coils within the aneurysm.