Figure 1.
An illustration of temporal learning via anticipatory drops in the response threshold.
The threshold drops earlier in the mostly congruent condition (top panel), benefiting congruent trials. The threshold drops later in the mostly incongruent condition (bottom panel), benefiting incongruent trials. Vertical tick marks on the normal threshold represent retrieved response times.
Figure 2.
The structure of the Parallel Episodic Processing (PEP) model.
Input nodes are stimulated first. Words and colours compete in Identity nodes, before passing activation on to Response nodes. Words also activate Episode nodes, which then activate the associated Response nodes. New to the model, Episode nodes also affect the response deadline dynamically over the course of a trial.
Table 1.
Frequencies of critical items, and mostly congruent and mostly incongruent filler items in the list-level PC manipulation.
Figure 3.
Analysis 1 data for congruency and proportion congruency.
Model-simulated (A^rpar; cycle times and (B) error percentages. For comparison, the original experimental (C) response times and (D) error percentages adapted from Hutchison (2011).
Table 2.
Analysis 2 coefficients, standard errors, t values, and p values for congruency x proportion congruency x previous RT mixed model on inverse RTs.
Figure 4.
Experiment 1 data for contrast and proportion easy.
Mean (A) response times and (B) error percentages.
Table 3.
Experiment 1 coefficients, standard errors, t values, and p values for contrast x proportion easy x previous RT mixed model on inverse RTs.
Table 4.
Analysis 3 coefficients, standard errors, t values, and p values for a congruency x proportion congruency x previous RT mixed model on inverse RTs.