A Neural Marker of Speech Intention: Evidence From Contingent Negative Variation

Abstract

Purpose This study investigated whether changes in brain activity preceding spoken words can be used as a neural marker of speech intention. Specifically, changes in the contingent negative variation (CNV) were examined prior to speech production in three different study designs to determine a method that maximizes signal detection in a speaking task. Method Electroencephalography data were collected in three different protocols to elicit the CNV in a spoken word task that varied the timing and type of linguistic information. The first protocol provided participants with the word to be spoken before the instruction of whether or not to speak, the second provided both the word and the instruction to speak, and the third provided the instruction to speak before the word. Participants ( N = 18) were split into three groups (one for each protocol) and were instructed to either speak (Go) or refrain from speaking (NoGo) each word according to task instructions. The CNV was measured by analyzing the difference in slope between Go and NoGo trials. Results Statistically significant effects of hemispheric laterality on the CNV slope confirm the third protocol where the participants know they will speak in advance of the word, as the paradigm that reliably elicits a CNV response related to speech intention. Conclusions The maximal CNV response when the instruction is known before the word indicates the neural processing measured in this protocol may reflect a generalized speech intention process in which the speech-language systems become prepared to speak and then execute production once the word information is provided. Further analysis of the optimal protocol identified in this study requires additional experimental investigation to confirm its role in eliciting an objective marker of speech intention. Supplemental Material https://doi.org/10.23641/asha.14111468