Cognitive Brain Research 17 (2003) 177–187 www.elsevier.com / locate / cogbrainres Research report Event-related potential correlates of the attentional blink phenomenon a, a,b a,b * Cornelia Kranczioch , Stefan Debener , Andreas K. Engel a ¨ ¨ Cellular Neurobiology Group, Institute of Medicine, Research Center Julich, 52425 Julich, Germany b Institute of Neurophysiology, University Hospital Hamburg-Eppendorf , University of Hamburg, Martinistr. 52, 20246 Hamburg, Germany Accepted 21 February 2003 Abstract The attentional blink phenomenon results from a transitory impairment of attention that can occur during rapid serial stimulus presentation. A previous study on the physiological correlates of the attentional blink employing event-related potentials (ERPs) suggested that the P3 ERP component for target items presented during this impairment is completely suppressed. This has been taken to indicate that the target-related information does not reach working memory. To reevaluate this hypothesis, we compared ERPs evoked by detected and missed targets in the attentional blink paradigm. Eighteen subjects performed a rapid serial visual presentation (RSVP) task in which either one target (control condition) or two targets had to be detected. ERPs elicited by the second target were analyzed separately for trials in which the target had been detected and missed, respectively. As predicted, detected targets did elicit a P3 during and after the attentional blink period. No clear P3 was found for detected targets presented before the attentional blink, that is, at lag 1. In contrast, missed targets generally did not evoke a P3. Our results provide evidence that targets presented during the attentional blink period can reach working memory. Thus, these findings contribute to evaluating theories of the attentional blink phenomenon.  2003 Elsevier Science B.V. All rights reserved. Theme: Neural basis of behavior Topic: Cognition Keywords: P3; P2; ERP; Attentional blink; Temporal attention; RSVP 1. Introduction target (T2) is presented at various time lags following the first target (T1). In a dual task condition, in which subjects The attentional blink is a phenomenon that can be attend to both targets, T2 detection performance is im- described as a transitory impairment of attention that paired for presentations of T2 at 200–400 ms after T1. In occurs if multiple targets have to be processed in close contrast, T2 detection rate is less impaired if T2 is temporal proximity. Initially described in the mid-1980s presented at lag 1 (i.e. as the first item following T1) or [1,18,29], the phenomenon was named attentional blink by after a delay of about 500 ms or longer. However, if only Raymond et al. [17]. While the attentional blink phenom- the second of the two target items is to be detected, no enon has been addressed in a large number of behavioral attentional blink is observed. That is, in this single task studies (see Refs. [21,22] ), its physiological mechanisms control condition detection accuracy for T2 items is good are still largely unknown. at any lag (e.g. Ref. [17] ). Hence, the dual task characteris- Typically, the attentional blink is investigated in rapid tic is one necessity for the attentional blink to occur, serial visual presentation (RSVP) tasks, where stimuli are another is masking of the targets. For T1 it has been shown displayed at a single location with a frequency of about 10 that any mask will do, that is, T1 can be masked by per second. The RSVP stream consists of a series of integration, interruption, or meta-contrast. T2, in contrast, stimuli, of which two are defined as targets. The second has to be masked by interruption in order to obtain a lag-dependent attentional blink [4]. Different theories have been proposed to explain the *Corresponding author. Tel.: 149-246-161-4686; fax: 149-246-161- attentional blink phenomenon. Two important theories are 2820. E-mail address: c.kranczioch@fz-juelich.de (C. Kranczioch). the two-stage model and the interference model. The two- 0926-6410 / 03 / $ – see front matter  2003 Elsevier Science B.V. All rights reserved. doi:10.1016 / S0926-6410(03)00092-2