😳

If the presented images were part of the training dataset for the model used, you can junk this tweet's conclusion out of the gate

how the fuck do you even construct anything out of an fmri machine? its output is literary just the activity density in the brain. how the fuck do you know what the visual cortex is imagining visually from external observation of electric activity in parts of the brain, without taking about the model of transforming the said signals into visual anything. anyone got a link on the paper? lets check this stuff