Working memory and the attentional blink: fMRI investigations of the neural correlates of the working memory bottleneck

The attentional blink (AB) is a well-studied effect that highlights the limitation of temporal attention. The AB is characterised as the impaired detection of the second of two masked targets when the second target falls within ∼500ms of the first. Different theories have been proposed to explain the AB but all point to a bottleneck at the level of working memory (WM). And although there are now several published imaging studies of the AB, none have investigated the effect of a concurrent WM manipulation on the neural correlates. Here we present fMRI data of a WM-AB study, where successful completion of the AB task was dependent on successful WM performance. Participants viewed an array of geometric shapes (1 shape - ‘low load’ or 3 shapes - ‘high load’) prior to the start of the AB task that contained a further geometric shape in a stream of letters. The participants first target task was to determine whether the geometric shape in the RSVP matched one of those presented in the WM array; additionally, as a second target task, participants had to detect the presence of an ‘X’ that occurred either in the AB sensitive period (‘short lag’ - 200 ms) or outside it (‘long lag’ - 700 ms). The behavioural results show that during the AB period a high WM load impairs performance more than a low WM load; but outside the AB period, high and low WM load was comparable. The imaging results show several areas that have been previously implicated in WM tasks reveal a larger BOLD signal during AB trials (short lag) for high vs. low WM load. These results are explained in terms of shared capacity-limited resources between structures supporting attentional selection in the AB task and WM.