The effect of explicit knowledge on sequence learning: A graded account

Abstract

In this paper, we study the effect of conscious knowledge on implicit sequence learning. To do so, in three sequence learning experiments, we manipulated (1) the extent to which instructions were intentional vs. incidental - intentional participants were informed of the existence of sequential regularities, and (2) the amount of explicit knowledge given to participants about the stimulus material. Results indicated that explicit knowledge improves sequence learning, as indexed by an increase in reaction times when the training sequence is unexpectedly replaced by another one. To enable us to differentiate between implicit and explicit learning, we applied the process dissociation procedure in a subsequent free generation task. Results indicated that both reaction time and generation results were influenced by different levels of explicit knowledge. However, we failed to find any evidence for an automatic influence on generation performance. We also report on simulation studies using the simple recurrent network, and show that the model can account for the effects of explicit knowledge on both reaction time and generation performance. Because the model uses a single pathway to process information, these simulation results are suggestive that dissociations between implicit and explicit learning might result from continuous, gradual changes in a single dimension rather than from the involvement of different brain networks.