Scientific Experiment Linking Eye Movements and Dreams

While adults who have been blind all their lives report no visual dreams and have no measurable rapid eye movements (using EOG measures), they do have periods of low-voltage EEG with reduced muscle tone, which is consistent with REM sleep (Berger et al., 1962), and if mechanical sensors are used to assess eye movement, it appears that their eyes do indeed wiggle about during these sleep phases (Gross et al., 1965). The lack of evidence of movement using standard EOG measures can be attributed to retinal degeneration in these subjects. This commonly results in the loss of the corneo-retinal potential, a standing potential difference to the order of one tenth of a volt across the back of the eyeball, which contributes largely to the measurable changes in potential on the face resulting from eye movements in normally sighted people.

Attempts to map eye movements directly onto reported dreams (in terms of direction of gaze and so on) have been dogged by the problem of measuring the position of the eyes using standard EEG equipment. (EEG amplifiers record changes in potential, but because they automatically filter out slow changes they are unable to provide information on the actual position of the eyes. They are not direct coupled - DC - amplifiers.) When Jacobs et al. (1972) made recordings using DC amplifiers it became apparent that during REM sleep the eyes of the subjects were almost constantly in motion, making slow, roving movements that could not be picked up by the standard (AC) amplifying equipment. Bursts of fast eye movement were superimposed on these slow movements. This study also found that fast eye movements were associated with passages of vivid dreaming that involved tracking moving images, but the measured direction of eye movement did not correspond very well with the reported direction of movement of the objects in the dream.

Waking saccadic (scanning) eye movements are faster when the eye has further to travel, with the effect that an eye movement over five degrees and one over 10 degrees are both completed in about 60 milliseconds. These movements are preprogrammed 'ballistic' responses whose velocities are determined by the anticipated distance of travel, ensuring that the eye arrives at the right place after a fixed period of time. Aserinsky et al. (1985) discovered that the fast eye movements of REM sleep tend to be consistently slower than waking saccadic movements. In addition there is no difference in acceleration or final velocity between movements that are long and movements that are short. They checked on the possibility that saccadic movements made with the eyes closed, or in the dark, might be slowed, and that velocity may then be independent of length of saccade, and found that they were not. That is, waking eye movements with the eyes closed, or in the dark, are the same as with the eyes open and in the light. The pattern of velocity of eye movements in REM sleep is quite unlike that of waking eye movements, and seems inconsistent with the idea that they reflect scanning responses.