Abstract
In a real-world context, brief glimpses of moving forms, especially people, are both prevalent and salient. Yet in studies of search and selective attention, targets are typically static stimuli, such as characters, words, or pictures. In the present experiments, targets were briefly presented point-light walkers (PLWs), each consisting of a coherent set of moving dots that outline a person carrying out one of four actions (Serre & Giese 2007). The attentional blink (AB) is a deficit in reporting the second of two targets with an SOA of about 200 to 500 ms. The question being addressed was whether attention exhibits a blink for actions defined by coherent motion.
In a series of experiments, participants were shown a sequence of stimuli that included one or two PLWs. Subjects had to name the action of each target (climber, boxer, runner, walker). In experiment 1 target actions were presented for 120 ms and followed by 120 ms masking stimuli created by sampling dot vectors from the different actions and scrambling their starting positions. At an inter-target SOA of 240 ms, report of T2 was strongly impaired compared to report of T1 or an isolated target. In a second experiment in which one or two actions were presented for 166 ms amid a rapidly presented stream of motion masks the blink was found to recover gradually over 700 ms.
This research contributes to ongoing evaluation of a computational model of working memory and attention (eSTST; Wyble, Bowman & Nieuwenstein in press) which proposes that the AB reflects a mechanism for parsing visual input into episodic packets prior to entry into working memory. Using PLWs performing complex sequences of actions, it will be possible to test the model with stimuli that naturally convey information over time rather than with the use of artificially rapid static image presentation.