The rest of the article discusses the different activation profiles of singing vs speaking but well done for finding this counterfactual where *repeating words/phrases* showed more overlap between speaking and singing. If you just search on google for 'stuttering when singing' you'll see its universally known to reduce/eliminate the effect, with citations going back to the 1960s on pubmed. Are you saying singing doesn't get rid of stutters? Are you a stutterer? I have never stuttered while singing songs despite a moderate stutter when speaking.