To highlight relevant information in dialogues, both wh-question context and pitch accent in answers can be used, such that focused information gains more attention and is processed more elaborately. To evaluate the relative influence of context and pitch accent on the depth of semantic processing, we measured event-related potentials (ERPs) to auditorily presented wh-question-answer pairs. A semantically incongruent word in the answer occurred either in focus or in non-focus position as determined by the context, and this word was either accented or unaccented. Semantic incongruency elicited different N400 effects in different conditions. The largest N400 effect was found when the question-marked focus was accented, while the other three conditions elicited smaller N400 effects. The results suggest that context and accentuation interact. Thus accented focused words were processed more deeply compared to conditions where focus and accentuation mismatched, or when the new information had no marking. In addition, there seems to be sex differences in the depth of semantic processing. For males, a significant N400 effect was observed only when the question-marked focus was accented, reduced N400 effects were found in the other dialogues. In contrast, females produced similar N400 effects in all the conditions. These results suggest that regardless of external cues, females tend to engage in more elaborate semantic processing compared to males.