Abstract

We investigated output-monitoring errors over speech based on findings in the research on the sense of agency. Several words were presented one-by-one, and we asked participants to say the word aloud, mouth the word, or imagine saying the word aloud. Later, participants were asked whether each word was said aloud. We found that the “said aloud” response was higher for generated words than that for observed words; it was decreased when the pitch of the feedback was lowered but still higher than when no feedback was received, and it was the same when no feedback was received and when feedback was replaced by another’s voice. Furthermore, we found that the “said aloud” response did not decrease even when the altered feedback was received with a short delay. These results were discussed according to the sense of agency and agency memory.