Transient Attentional Enhancement during the Attentional Blink: ERP correlates of the ST<span class='mathrm'><sup>2</sup></span> model

Chennu, Srivas and Craston, Patrick and Wyble, Brad and Bowman, H. (2008) Transient Attentional Enhancement during the Attentional Blink: ERP correlates of the ST<span class='mathrm'><sup>2</sup></span> model. In: FROM ASSOCIATIONS TO RULES: Connectionist Models of Behavior and Cognition. (Full text available)

PDF
Download (399kB)
[img]
Preview

Abstract

There has been considerable recent interest in the identification of neural correlates of the Attentional Blink (AB), and the development of neurally explicit computational models. A prominent example is the Simultaneous Type Serial Token (ST2) model, which suggests that when the visual system detects a task-relevant item, a spatially specific Transient Attentional Enhancement (TAE), called the blaster, is triggered. This paper reports on our investigations into EEG activity during the AB, and a hypothesized correlation between the blaster and the N2pc ERP component. Specifically, we demonstrate that the temporal firing pattern of the blaster in the model matches the N2pc component in human ERP recordings, for targets that are seen and missed inside and outside the attentional blink window. Such a correlation between a computational account of the AB and ERP data provides useful insights into the processes underlying selectivity in temporal attention.

Item Type: Conference or workshop item (UNSPECIFIED)
Uncontrolled keywords: Attentional Blink, Temporal Attention, Computational Modeling, ERP
Subjects: Q Science > QA Mathematics (inc Computing science) > QA 76 Software, computer programming,
Divisions: Faculties > Science Technology and Medical Studies > School of Computing > Computational Intelligence Group
Depositing User: Mark Wheadon
Date Deposited: 29 Mar 2010 12:09
Last Modified: 20 Apr 2012 11:33
Resource URI: http://kar.kent.ac.uk/id/eprint/23962 (The current URI for this page, for reference purposes)
  • Depositors only (login required):

Downloads

Downloads per month over past year