You know what? I bet the differentiation in the exposure to light (via different pictures), combined with ever-so-slightly-still-electrolytically-active neuro-pathways, produced those results.
Dead salmon, though they may be, light-sensitive organs the eyes still are.
And provided that the salmon were probably preserved on ice, much of the carcasses probably remained fairly chemically stable, so while the electron pumps in the electron transport chains of the multi-polar neurons were deactivated, I bet the whole pipeline from salmon's visual cortex to its eyeball probably was still slightly capable of transmitting photo-sensitivity.
Then, it probably came down to the fMRI imaging software being so fine-tuned as to notice that differentiation, and amplify it into statistical significance!
If that is true, then fMRI studies of V1 in humans are totally screwed. You show them anything, and the unconscious effects overwhelm the stats software and you get 'hits'. As the article goes into, you better be damn careful with your controls and your p values if you want to actually do anything useful in fMRI. If you want to just publish papers, it seems like a great field, but there seems to be better ways to just make a paycheck.
Dead salmon, though they may be, light-sensitive organs the eyes still are.
And provided that the salmon were probably preserved on ice, much of the carcasses probably remained fairly chemically stable, so while the electron pumps in the electron transport chains of the multi-polar neurons were deactivated, I bet the whole pipeline from salmon's visual cortex to its eyeball probably was still slightly capable of transmitting photo-sensitivity.
Then, it probably came down to the fMRI imaging software being so fine-tuned as to notice that differentiation, and amplify it into statistical significance!