The Simultaneous Type/Serial Token model (Bowman & Wyble 2007) is a neural network simulation of working memory encoding.  The model suggests that the attentional blink phenomenon (Raymond, Shapiro & Arnell 1992) is actually a cognitive strategy that facilitates the encoding of episodically distinct visual representations.

The eSTST model  (Wyble, Bowman & Nieuwenstein in Press)  is an elaboration of this account with a few modifications:


1.  Attention is still mediated by transient attentional enhancement, but the duration of each attentional episode can vary according to the sequence of targets

2.  In STST, multiple types could be assigned to a single token during lag-1 sparing. In eSTST, tokens are rigidly defined: only one type can be assigned a token.  



Presentation slides from ECVP 2008 (Utrecht)

Keynote
Powerpoint (exported from Keynote, may be buggy!)



Computational Model:

Matlab Code   (updated July 16, 2008)

This package contains a graphical interface which allows simulation at the levels of single trials or entire experiments.