Sequential effects in the Simon task: conflict adaptation or feature integration?

Sequential effects--reduced behavioral interference after incongruent trials-provide a unique opportunity to examine the influence of previous experience on current attentional processing. We conducted a theoretical analysis of performance in the Simon paradigm to distinguish an attention account (conflict adaptation) from a working memory account (feature integration). Results supported the memory account. Feature unbinding contributed significantly to sequential effects, even when separated statistically from the effects of Simon conflict. Sequential effects to neutral stimuli were comparable to those found with Simon stimuli, suggesting at most a minor contribution from conflict adaptation. These patterns emerged in behavioral data (reaction time and accuracy) and in two event-related potential components: N2 and SP (a slow positivity 480-520 ms after stimulus onset). They suggest that sequential effects arise from the partial mismatch of S-R features in working memory.