Introduction. In a texture segmentation task in which performance peaks at mid-peripheral locations and drops at both central and peripheral locations, attention improves performance at peripheral locations but it hinders it at central locations. This indicates that attending to the target location enhances spatial

resolution - improving performance where the resolution is too low (periphery) but impairing it where resolution is already too high (central locations) for the task (Yeshurun & Carrasco 1998, 2000).

Goal. To evaluate the hypothesis that attention enhances spatial resolution by increasing sensitivity to high spatial frequencies, we employed a backward band-masking paradigm while manipulating

transient covert attention by precueing the target location.

Method. In a 2IFC task, observers were presented with a precue (60 ms) that indicated either both target location and display onset (peripheral cue) or just display onset (neutral cue). This was followed by a 60 ms ISI and then the texture pattern was presented for approximately 50 ms. A target patch appeared at varying eccentricities along the horizontal meridian. Immediately after the display presentation a post-mask appeared for 100 ms; there were 3 types of masks: a) a wide-spectrum frequency post-mask, b) a low-pass mask, or c) a high-pass mask.

Results and Conclusion. The high-pass mask diminished the central performance drop for both the neutral and the peripheral precue conditions. In contrast, the low-pass mask augmented the central performance drop in the peripheral precue condition. These results are consistent with the finding that selective adaptation to high spatial frequencies diminishes the central performance drop and eliminates the attentional impairment (Carrasco & Loula, 2000). Both studies provide evidence for the spatial nature of the central performance drop and indicate that attention enhances spatial resolution by shifting sensitivity to higher spatial frequencies.