N400 and lexical decisions: automatic or controlled processing?

Objective: To investigate whether the N400 effect is sensitive to automatic or controlled processes. Methods: Two experiments were performed. In one experiment, directly related word pairs were used. In the other experiment, mediated-related word pairs were used. In order to reduce controlled processes, each experiment consisted of 3 tasks: Low- and high-proportion of related pairs, and single presentation lexical decision task. Results: In the first experiment, the amount of priming was equivalent for the 3 tasks. The N400 effect appeared in the high and low proportion of directly related words, but not in the single presentation task. In the second experiment, behavioral priming was also found in the 3 tasks. However, the N400 effect was observed only in the task with low proportion of related pairs. Conclusion: These results suggest that the N400 effect may be related to controlled processes.