When a target is flashed before, during, or after a saccade its perceived location depends on the timing of the flash (Jordan & Hershberger, 1994, Schlag & Schlag-Rey, 1995), but also on the retinal eccentricity of the flash (O'Regan, 1984) and on the visual frame of reference (Honda, 1999). These studies showed how saccades affect the metrics of the visual space, but they also revealed that these metrics depend on the content of the actual visual scene as well. We investigated whether the localisation of brief flashes depends 1) on symmetry of the visual stimulation around the point of fixation or 2) on the centre of spatial attention which can be shifted even without a subsequent saccade.

Six healthy subjects, whose eye movements were recorded with an infrared system (IRIS, Skalar, Delft, Netherlands), participated in the experiments. Subjects were seated in darkness and fixated always a central target spot. An additional single reference target was presented for 800 ms on the right or on the left side on the horizontal meridian at eccentricities that varied from trial to trial (9 ±1 deg). The central fixation spot was blanked 400 ms after the disappearance of the reference target and 100 ms later a test flash (duration 30 ms) appeared on the horizontal meridian in close neighbourhood to the previous reference target. In a forced choice alternative task (200 trials per session) the subjects had to decide whether the test flash appeared right or left from the reference.

1) In this first experiment 5 of the 6 subjects overestimated of the eccentricity of the test flash with respect to the reference position. The average mislocalisation was 0.5 ±0.7 deg (N = 6).

2) In a second series of experiments (non-cued) two reference targets were presented symmetrically around the fixation spot. The subjects did not know on which side the test flash would be presented next and consequently they had to distribute their spatial attention symmetrically to the left and to the right. In this experiment subjects showed only an insignificant tendency (p < 0.26) to overestimate the eccentricity of the test flash with respect to the reference position by 0.45 ±0.9 deg (N = 6).

3) In the third experiment (cued) the subjects were instructed to memorise only one of the two reference targets which was cued by the colour of the fixation spot. Shifting the visual attention in such a way towards the reference target, did not affect the localisation error (0.61 ±0.66 deg; N = 6). The small increase of 0.16 deg with respect to the non-cued experiment did not reach significance (p < 0.8). In all experiments the accuracy of the localisation ranged between 0.6 and 3 deg.

These results suggest that large mislocalisations of more than 3 deg as they are observed before saccades are not induced by asymmetries of the visual stimulation (1, 2) or a shift of the visual attention towards the saccade target alone (2, 3).