Two systems for spatial location during saccades

We investigated whether there may exist separate spatial representations for perception and action, taking advantage of a strong perceptual distortion that occurs with saccadic eye movements. Subjects were required to make 20í horizontal saccades, and report the position of a bar briefly displayed on a touch screen just prior to the saccade, either verbally or by pointing to a touch screen (with hand obscured from view). The screen was viewed through a liquid crystal shutter that could close 75 ms after bar display. With verbal reports, subjects reported large errors in localization, showing the characteristic compression of perceived positions towards the saccadic target (Ross et al, Nature, 1997). This occurred both with clear viewing and when the shutter closed after bar presentation. However, under the same conditions, subjects pointed accurately to the target position (with shutter or eyes closed at time of pointing). Interestingly, when subjects pointed to a screen in clear view (but with their hands concealed), the pointing showed the same compression as verbal reports. We conclude that there are separate representations of visual space, one plastic and subject to distortion, the other not. The plastic representation, used for conscious visual perception, seems to dominate when both are available.