You must log in or register to comment.
This explanation is a really terrible explanation - it doesn’t even explain the point of the study or what multiple comparisons are in statistics…
This is what the salmon study was saying - if you did a whole brain analysis with fmri without an a priori prediction you need to adopt a radically more stringent threshold for significance
The salmon in the MRI was a bowl with a million jellybeans in it
The explanation was pretty good without going into the statistical details which is perfectly OK for most people on social media.
wow amazing. dead salmon can read human emotions :O
so, as far as i understand it:
- either dead fish are geniuses
- or every single MRI result we ever got is bs and thus thousands of papers using them as a main component are invalidated
do i get that right?
if so, mygosh, i gotta get myself a dead fish to see if im faking depression or not.
maybe its an LM type situation: it doesnt matter if its fake “roleplaying” or not, the actions are what matters.
Not all MRI studies, just a large number of particularly lazy ones. Unfortunately lazy journalists pick up the lazy science and so these studies tend to be really popular in lazy clickbait journalism.
Basically there’s still good science, as well as a parallel economy of bullshit from people who have long ago stopped giving a shit.
do i get that right?
Yes.
i gotta get myself a dead fish to see if im faking depression or not.
No, no, not like that.
Study for those interested: https://www.sciencedirect.com/science/article/abs/pii/S1053811909712029
Diogenes, holding up a salmon.
Behold: a salMAN.
for people who want more than a social media screenshot about this:
- https://en.wikipedia.org/wiki/Multiple_comparisons_problem
- https://www.scientificamerican.com/blog/scicurious-brain/ignobel-prize-in-neuroscience-the-dead-salmon-study/
- the actual paper: https://teenspecies.github.io/pdfs/NeuralCorrelates.pdf
(ht to @bjorney@lemmy.ca who already linked the inevitable relevant xkcd…)
