The scenario is simple, a man is blinded by two strokes which damage the visual processing region of the brain (the retina and the optic nerve are intact). A rather peculiar psychologist decides to show this blind individual a series of images.
When it came to ‘deadpan’ human faces he couldn’t make heads or tails.
When it came to shapes all he offered were wild guesses.
Pictures of animals that seem threatening also did not elicit any response.
But, the patient showed a non-trivial ability to distinguish human faces expressing various emotional states (he was 60% accurate, definitely not very good, but also non-trivial). There was heightened activity in the patient’s amygdala, which is known to respond to facial expressions and other non-verbal cues. Remember, this is a person whose visual center is shot, someone who is functionally blind, but modular independence of human facial recognition of emotion allowed this cognitive pathway to still transmit some information. In a BBC interview the lead researcher noted that the patient himself really had no idea what he was supposed to be seeing, and assumed that his guesses about emotional states were wildly inaccurate just as his previous ones had been, that is, his conscious reflective mind did not have direct access to this information! The information processing was occurring “under the hood.”
You can read a summary here, but the full paper will be out in Nature Neuroscience.
Massive modularity? I don’t know, but far more modular than you would think. Perhaps they are all a series of felicitous spandrels….
Posted by razib at 02:50 AM