Blindsight is an intriguing phenomenon that really should lead us to question a lot of what we ‘know’ about consciousness. Someone with blindsight will get in an elevator and insist they can’t see the buttons. But after being pushed and cajoled to press their floor’s button, they will consistently succeed every time. A more classic experiment goes something like this: give a blind or blindsighted person an envelope and ask them to fit it through a slot which is either horizontal or vertical. A blind person has trouble and never does above chance. A blindsighted person does perfectly, at exactly the same level as a seeing person. So they cannot consciously ‘see’, but the visual information seems to all be in there somewhere.
Schmid et al test the hypothesis that the LGN has a causal role in blindsight. Blindsight is often (always?) caused by damage to V1, so they must show that the LGN has a role in V1-independent visual processing. Many pathways have been proposed to be crucial to blindsight; most prominent are the superior colliculus and a secondary thalamic nucleus, the pulvinar, which projects to extrastriate cortex.
Schmid et al lesioned two adult macaques in V1. When high-contrast stimuli were presented in the scotoma, the monkeys could accurately saccade to the location; low-contrast stimuli elicited no response, as if the monkey were unable to perceive it. fMRI scans showed that extrastriate responses in V1-lesion hemispheres showed activity that are consistent with human blindsight data, about 20% of normal. By inactivating the (posterior) LGN with injections of the GABAA agonist THIP, scans now showed noactivity and removed the animals’ ability to detect high-contrast stimuli. This indicates that extrastriate activation routes through LGN rather than some other pathway. The figure above shows how well the two monkeys perform when the stimulus is presented in the scotoma or contralaterally; the left two are for the normal condition, the right two are for inactivated LGN.
What have we seen? That visual activation of extrastriate areas requires LGN. It has been shown elsewhere that superior colliculus is actively involved in blindsight; interestingly, V1-bypassing neurons receive input from the superior colliculus. As always, it is not one or the other that is important, but both regions that are required in concert to support V1-independent vision. It would have been nice to see an example of blindsight in this blindsight paper (where was the unconscious detection of the stimuli?), but otherwise it is a nice demonstration of the confluence of various techniques to produce an interesting result.
Schmid M, Mrowka S, Turchi J, Saunders R, Wilke M, Peters A, Ye F, Leopold D (2010). Blindsight depends on the lateral geniculate nucleus. Nature. DOI: 10.1038/nature09179