The Attentional Blink is Immune to Masking-Induced Data Limits

Abstract

The attentional blink is the robust finding that processing a masked item (T1) hinders the subsequent identification of a backwards masked second item (T2), which follows soon after the first one. There has been some debate about the theoretically important relation between the difficulty of T1 processing and the ensuing blink. In Experiment 1 we manipulated the difficulty of T1 in such a way as to affect the quality of data without altering the amount of resources allocated to its identification. We found no relation between the accuracy of T1 identification and the blink. In Experiment 2, the same difficulty manipulation was applied to T2, and we observed an additive pattern with the blink. Together, this pattern of results indicates that a data-limited difficulty manipulation does not affect the blink, whether applied to T1 or T2. In Experiment 3 we used an individual differences methodology to show that performance in the traditional "stream"-like presentation (rapid serial visual presentation) was highly correlated with performance in our modified "target mask, target mask" paradigm, thus allowing for comparisons beyond the present methodology to much of the previous literature that has used the stream paradigm.

Object file continuity predicts attentional blink magnitude

“…This suggests that the new, composite texture pipe-pan mask did serve as a more effective mask for a pipe-pan target in streams consisting entirely of pipe-pan items, whether they were presented sequentially or not. This is supported by the fact that T1 performance was lower in Experiment 6 despite a far slower presentation rate having been used, although this reduced T1 accuracy may have resulted from resource-limiting factors, rather than from encoding-difficulty-related factors (McLaughlin, Shore, & Klein, 2001).…”

A conceptual and methodological framework for measuring and modulating the attentional blink

“…However, using performance at ceiling as a baseline for the posttarget deficit would be confounded with other effects on T2 performance. As McLaughlin et al (2001) pointed out, target performance is almost never at ceiling, even when it is not subject to dual-task impairments (see Fig. 5).…”

“…If T1 performance is too high, it is likely to overestimate the AB, just as ceiling performance will; if it is too low, it may abolish the AB entirely. However, if T1 performance is at a level similar to what could be expected for T2 performance in a control condition, it can be used effectively to estimate the AB 1 See Buchholz & Davies, 2007;Facoetti et al, 2008;Ferlazzo, Lucido, Di Nocera, Fagioli, & Sdoia, 2007;Ho, Mason, & Spence, 2007;Kihara et al, 2007;Lum et al, 2007;McLaughlin et al, 2001;McLean, Castles, Coltheart, & Stuart, 2010;McLean, Stuart, Visser, & Castles, 2009;Olivers & Nieuwenhuis, 2005Raymond & O'Brian, 2009;Raymond et al, 1992;Seiffert & Di Lollo, 1997;Visser et al, 2004;Visser & Ohan, 2007. In conclusion, various baselines can be used to capture the AB effect. The original control condition used by Raymond et al (1992)-that is, T2 performance when T1 is present but ignored-is the least problematic, as the parameters of T2 in the control condition only differ from those of T2 in the experimental condition by one factor: whether or not a previous target was attended to.…”

“…9 Two examples depicting interactions of condition or group with lag that are more difficult to interpret as modulations of the AB individual differences in the magnitude of the AB effect has been found (Dale & Arnell, in press;McLaughlin et al, 2001 (Colzato et al, 2010), working memory control Colzato et al, 2007), reading ability (McLean, Stuart, Visser, & Castles, 2009), global/local bias (Dale & Arnell, 2010), resting EEG (MacLean, Arnell, & Cote, 2012), and the ability to keep irrelevant material out of working memory (Arnell & Stubitz, 2010;Dux & Marois, 2009) have all been used to investigate individual differences in AB magnitude. In order to investigate individual differences in the AB using a correlational approach, it is necessary to capture the magnitude of the AB effect for each individual as a single value.…”

“…T1 performance has also been used as a baseline (Chua, 2005;McLaughlin et al, 2001;Shore et al, 2001). Shore et al suggested that using T1 performance as a baseline avoids the issue of any posttarget influence, and thus will not misestimate the AB.…”