Abstract
How does dividing attention affect visual sensitivity? Cueing a location improves contrast sensitivity (compared to uncued locations); but in life, a pertinent stimulus may appear in more than one place. If attention draws on limited processing resources, the straightforward prediction is that as the number of attended locations increases, contrast sensitivity should decline. However, spatial attention is known to be remarkably flexible, and manipulations of the number of cued locations have led to inconsistent results. Here we investigated the relation between contrast sensitivity and the number of attended locations.
In Experiment 1, participants reported the orientation (vertical/horizontal) of a target grating on each trial. Targets had varying contrasts, could appear in one of four locations arranged in a square around fixation, and were preceded by one to four briefly-presented peripheral cues directing attention to possible target locations. Cues were always valid (targets appeared only at one of the cued locations). Results showed that sensitivity decreased steadily as number of cues increased. This reduced sensitivity demonstrates the limits of divided attention at attended locations (‘bright’ attention).
To investigate the effect of attention at uncued locations (‘dark’ attention) in Experiment 2, either one or three locations were cued but cues had no predictive validity for target location. If dividing attention reduces sensitivity overall, sensitivity should decline with the number of cues at both attended and unattended locations. But if it is the efficiency of resource distribution that is impaired by divided attention, this leads to a counterintuitive prediction: sensitivity at unattended locations should improve as the number of attended locations increases. Indeed, more cues again led to reduced sensitivity at attended locations, but to a symmetrical improvement at unattended ones. This finding demonstrates the existence of ‘dark’ attention (attentional operations at unattended locations) and resolves previous inconsistencies regarding divided attention.
Supported by grant NIH R01 EY016200 to MC.