Effective visual short-term storage capacity and speed of encoding are affected by arousal

Effects of spatial cueing on visual attention have been thoroughly investigated during the last 30 odd years. Similar to spatial cueing, temporal cueing seems to afford a performance enhancement to an observer when he or she knows the point in time at which an event will occur (Coull & Nobre, 1998). Varying the statistical distribution of cue-stimulus onset asynchronies (foreperiods) is an effective way of manipulating the observer's temporal expectancies and, presumably, the observer's level of arousal. By use of this manipulation, Vangkilde and Bundesen (2009) found strong evidence that speed of encoding into visual short-term memory (VSTM) increased with the level of arousal in a single-stimulus identification experiment. Here we present a whole-report experiment corroborating the finding that speed of encoding stimulus items (letters) into VSTM increases with the level of arousal. However, by way of contrast, the maximum number of stimulus letters retained in VSTM appeared to decrease as the level of arousal was increased. A possible explanation for this finding is that, as the level of arousal was increased, selectivity deteriorated such that not only the items to be reported but also irrelevant material tended to be encoded into VSTM leaving less storage capacity for the items to be reported. The explanation predicts that performance in otherwise similar partial-report experiments will show that selection of targets rather than distractors becomes less efficient at high levels of arousal. Thus, in terms of the TVA model (Bundesen, 1990), partial report experiments should show that, at high levels of arousal, parameter a (the ratio of the attentional weight of a distractor to the attentional weight of a target) increases.