Direct estimates of processing delays in the attentional blink.

When two targets (T1 and T2) are presented in rapid succession, identification of T2 is often impaired (attentional blink: AB). The two-stage model accounts for the AB deficit by assuming that: (a) T2 is delayed in Stage 1 while Stage 2 is busy processing T1, and (b) T2 is vulnerable to masking while delayed. We report converging evidence for the model by evaluating these assumptions independently of the AB deficit itself. The results show that: (a) response times for T2 identification decreased as the lag between T1 and T2 was increased; (b) response times for T2 decreased across lags only if T1 was masked; and (c) accuracy of T2 identification increased as the stimulus-onset asynchrony between T2 and the trailing mask was increased.