The dead salmon study showed the need to correct fMRI readings, if I recall correctly. I think they showed the dead salmon some pictures instead of playing music.
It is possible to misuse any technology to lie or tell a story other than the truth. I worked in a pioneering fMRI lab back in the early 00s, and even then we were very clearly stating the limitations of the technology. Itβs popular media that is uncomfortable with ambiguity that place unrealistic and magical powers of observation on fMRI.
See e.g. https://www.liebertpub.com/doi/abs/10.1089/ham.2011.1073