Department of Psychology, UC Santa Cruz
July 22, 2013 12:10pm
CMB Large Conference Room
267 Cousteau Place in East Davis
Space constancy is the perception that the visual world is stable despite all visual information arriving through retinas that move continually with respect to the world. Since the time of Helmholtz the dominant explanation for space constancy has been the efference copy or corollary discharge, a copy of the innervation to the eye muscles that sends an equal and opposite signal to the visual brain to compensate for the effects of eye movements. The goal was to achieve constancy of perceived visual direction and calibration of visually guided actions. Empirical work has shown, however, that the efference copy is too slow and has too low a gain to achieve either of these goals. We demonstrated this by pressing on the outer canthus of one eye while the other is covered. This evokes a sustained perception of offset of the visual world even though the eye muscles compensate for the offset, maintaining fixation on the target. The perceived deviation can be measured psychophysically by a nulling technique. At the same time the magnitude of the efference to the eye muscles can be measured by the deviation of the covered eye, using infrared recording. The result was that efference copy accounted for only about 5/8 of the required compensation. Proprioception, measured with an analogous method, added another 1}4 of the compensation. The deficit in gain accounts quantitatively for some previously unexplained illusions of visual direction, but is still not enough to underlie space constancy. Newer alternatives to the efference copy theory maintain that visual position information is not compensated, but is destroyed during each saccadic eye movement and replaced by information from the new fixation. In a ‘reference object’ theory the visual system first identifies a saccadic goal target, usually but not always the object to be on the fovea after a saccadic eye movement. During the saccade information about old image positions is destroyed, except for the reference object. If that object lands within a spatiotopic window after the period of saccadic suppression, the object is considered found and constancy of its position is assumed. Other objects in the new fixational view are localized relative to the reference object. It is only if the reference object is not found that efference copy, oculomotor proprioception and possibly other information sources are exploited to localize the new image, and space constancy often fails.