I was recently overgeneralizing to a friend about the overall value I find in a lot of brain imaging research, and I just now had a moment to privately dance to music in a way that made me want to clarify what was I saying. A specific example of the problem would be the false positives in that dead fish fmri study. But that example is just to say there are real statistical problems when “rigourous” brain imaging is only connected to remedial cognitive science. The subsymbolic layer of thought is a blind spot for a lot of neuroscientists in my experience.

Of course there is lots of good brain imaging research. And certainly you don’t need a bleeding edge cognitive model of the data just to reduce the uncertainty of any leading systemic detections that might be discovered in localized brain regions. Like using Bayesian priors based on meta analysis of a literature, to say this or that brain region is associated with this or that function. Yes, that’s true. But even at this level you bring unmodeled priors to the table about what you’re trying to find with the research questions you’re asking. For example, doing instinguishable, though interesting, basic research, while a genocide in the Middle East conducted by our “ally” Israel, threatens to engulf the world in mass despotism that will certainly nullify the value of this interesting research. Almost as if it’s time for those concerned to revisit Chomsky1967 about the responsibility of intellectuals.

This is the song I was dancing to:

And Chomsky1967 spoon fed to you: