Abstract
We had observers perform an attentional gating task (Shih & Sperling, 2002) that required them to monitor a central 9.4 Hz RSVP stream for a cue digit. The identity of the digit instructed them to shift attention to one of two flanking streams and report the first letter they saw in that second stream. Attentional reaction time (ART) was defined as the interval between the cue and the reported letter. The average ART was 443 ms. We assume that observers attend to the cue stream, then make a voluntary attentional shift to the flanking stream. Are other attentional resources available during this shift? We assessed this by presenting a brief visual search task sometime during the gating task. Assuming a unitary spatial attention resource, one would predict that search performance would be severely impaired during the shift from one stream to the other. The search task required observers to report whether or not a T was present among distracting Ls in a 4-item array. The search array was presented at the same eccentricity as the flanking RSVP streams for 107 ms and then masked. The SOA between the cue in the central RSVP stream and the search array varied from −427 ms to 427 ms. There was no dual task cost on the attentional gating task. Search d′ was reduced in the dual task case relative to a single task control condition. Is this a simple dual-task cost or interference between attentional gating and search? The critical search trials are those that appear after the cue digit but before the reported letter in the flanking stream (0 < SOA < ART). We compared performance on these critical trials to control trials where the search task appeared either before the cue digit (SOA < 0) or after the reported flanking stream letter (SOA > ART). Search during the attentional shift (d′ = 1.20) was not impaired relative to control (d′ = 1.13). We conclude that it is possible to perform visual search while concurrently executing a volitional shift of attention.