Abstract
The brain is capable of producing visual experiences in the absence of retinal input (e.g. mental imagery and hallucinations). We present an unusual demonstration of this ability: a 34-year-old woman, NS, who developed persistent and complex vision-like experiences after losing her sight to a rare retinal degeneration disease. Unlike the hallucinations in Charles Bonnet Syndrome, these experiences are consistent with her environment. After receiving cues of the presence of an object, either through proprioception or haptic interaction, NS “sees” an iridescent representation of the object’s form. Unlike imagery, these representations are involuntary and persist as long as she understands the object to be within her line-of-sight. To characterize the neural correlates of these visual experiences, we used 3T fMRI to record BOLD signal while NS completed a task evoking visual perceptions. NS, and a sighted control, placed auditory-cued 3D objects on a plexiglass tray that held the object suspended in their field of view. Subjects then observed the object without touching or movement for 4 seconds: through veridical vision for control, and through “non-optic vision” evoked from prior haptic feedback for NS. This was followed by imagery runs, where subjects completed the motion of placing objects but imagined seeing them instead. A GLM analysis found significant (pFWE<0.05) clusters of activation during object viewing within the visual cortex and nearby temporal and parietal regions of both subjects. The presence of increased activity in visual areas of NS’s brain while “seeing” the objects is consistent with her reported vision-like experience. During imagery, activity was attenuated and more confined to the higher visual areas. This represents the first neuroimaging account of such advanced cross-modal non-optic sight in a blind patient, and further supports visual perception being a generative process that depends as much on top-down inference as on retinal input.