Attention shift not memory averaging reduces foveal bias.

Two experiments examined which of two mechanisms, attention shift or memory averaging, reduces foveal bias. The target stimulus was a black dot presented for 80 ms while observers maintained fixation. The two main conditions were 'with' and 'without' vertical and horizontal bars as landmarks, which were placed on more eccentric positions than the target stimulus. To induce attention, the landmark was flashed on for 80 ms (Experiment 1) or disappeared (Experiment 2) with a stimulus onset asynchrony of 0, 106.4, or 212.8 ms in both experiments. As a control, non-flashed and non-disappeared landmark conditions were employed. The observers' task was to point to the remembered location of the target with a mouse cursor. The results showed that the magnitudes of foveal bias were significantly lower in the flashed and disappeared landmark conditions than in the without landmark condition. Furthermore, the magnitudes in the flashed and disappeared landmark conditions did not differ from their respective control conditions. The latter finding in the disappeared landmark conditions provides evidence for 'attention shift' against 'memory averaging' as the mechanism reducing foveal bias.