Abstract
Attending to a feature in one location can produce feature-specific modulation in a different location. This global feature-based attention effect has been commonly studied with two stimulus locations, one in each hemifield. With such a stimulus configuration, it is impossible to study the spatial profile for the spreading of feature-specific modulation.
We examined the spread of feature-based attention by measuring attentional modulation of motion aftereffect (MAE) at remote locations (Boynton, Ciaramitaro, & Arman, 2006). Observers viewed a stimulus composed of two overlapping dot fields that moved in opposite directions in the center of the screen (adapter). They attended to one of the dot fields by performing a two-interval forced choice task on its speed. Following the adapter, a single-direction dot field (test) appeared at one of six possible locations along the horizontal meridian (eccentricity: ±5, ±10, ±15 deg). The adapter and test did not spatially overlap. A nulling technique was used to measure the perceived stationary point (PSP). We also measured the baseline MAE effect for each location by presenting a single-direction adapter and test in that location. Attentional effect was quantified as a fraction of the baseline effect, to normalize against variation in adaptability across locations. Consistent with prior work, we found a global effect of feature-based attention (the PSP depended on the direction to which subjects attended to during adaptation). Importantly, the size of this effect was not spatially uniform, but showed a reduction with greater eccentricity.
These results show that the spreading of feature-specific modulation is not uniform across the visual field but has a measurable spatial gradient. Thus feature-based attention can be global, but is not completely location-independent.