I think you're confusing "dangerous" with "working as designed" - the whole point of the experiment was to have the subjects "learn" without explicitly having t

I think you're confusing "dangerous" with "working as designed" - the whole point of the experiment was to have the subjects "learn" without explicitly having to do anything. Plus, the article also states that "researchers have said that their method has only been shown to work with visual perceptual learning". Also, I believe fMRI induction is quite safe (on a physical level), as the actual magnetic field strengths are controlled very precisely to be as weak as required. At most, if you think of parts of the brain as neural networks, the fMRI stimulation might only act to modify the NN weights.