Dear Stephen and Antoine, thank you for the much appreciated feedback. I found this paper (http://bit.ly/14PwH8x) that describes how to build a MR-compatible stereoscope, and I was thinking to give it a try. The idea is simple but ingenious: instead of using mirrors it adopts prismatic lenses and a divider. I haven't talked to an optician yet, but I guess these lenses can be created with any form, and then one can easily use the MEG-compatible glasses. The only limit I can think of is that the prismatic distortion is fixed, so one can adjust the horizontal alignment to fuse the images only by changing the position of the stimuli on the screen. But if the distance between the subject and the screen is fixed, then this should be sufficient for inter-subject differences of eye distance. One added benefit of this option is that one should be able to use an eye-tracking system as well. The main drawback might be the cost of the lenses. I would like to hear your thoughts on this. Thanks! Matteo