In darkness, a brief flash stimulus presented near the time of a saccade is perceived in different positions from the actual stimulus position. This saccade-contingent mislocalisation was explained by the subtraction theory (Honda, 1989, Perception & Psychophysics, 45: 162-174). That is, mislocalisation is produced by the mismatch of the extraretinal eye position signal and the object image displacement on the retinal. However, the experimental results concerning the mislocalisation of a repeatedly flashing stimulus suggested that the attended position temporally proximate to the saccade is responsible for the time course of mislocalisation (Ebisawa and Fujiwara, submitted). Especially, when the subjects were asked to shift their attention to follow the apparent motion produced by the phantom flashes, the final phantom flash position shifted gradually along with the estimated time course of the attended position. In the present study, the effects of the flashing frequency and background stimuli upon the time course of mislocalisation of a flashing stimulus are investigated.

Three male students participated in the present experiment. Two experiments were conducted in complete darkness. Their heads were immobilised. Three green LED's were located in the centre of the visual field, 10 deg to the right, and 10 deg to the left as a fixation point (FP), saccade target (ST), and background stimulus (BS), respectively. A flashing stimulus (FS, red LED) was located in the centre, above the FP.

In Experiment 1, after extinction of the background stimuli (ST, FP, and BS), the subject made a saccade from the position of the extinguished FP to the position of the remembered ST. The FS was flashed repeatedly at one of four flashing frequencies (40 - 500 Hz, flash period: 1 ms). The FS was extinguished at various times after saccade onset. The subject was asked to report the finally perceived phantom flash position after observation of each trial. Generally, the FS was mislocalised to the same direction of the saccade. The degree of mislocalisation of the final flash of the FS decreased with the decrease of the flashing frequency. This result was considered as follows. The higher flashing frequency produced more phantom flashes, yielding the apparent smooth motion, and this guiding the attention toward the saccade goal.

Experiment 2 consisted of the dark and background conditions. In the background conditions, the stimuli (ST, FP, and BS) were presented throughout the saccade. The results showed that the time course of mislocalisation was a constant from onset until 200 ms after onset of the saccade in the background condition, whereas the localisation error increased gradually after onset until 50 ms after onset of the saccade in the dark condition. This means that the mislocalisation suddenly increased at saccade onset because of no mislocalisation before saccade onset (Ebisawa, submitted). These findings suggested that the background stimuli disturbed the smooth attention shift, producing the sudden attention shift close to saccade onset.