Task dependent modulation of voice fundamental frequency responses elicited by perturbations in pitch of auditory feedback during English speech and sustained vowels

A noninvasive behavioral technique, called the "pitch-shift paradigm", holds the potential to reveal much about the audio-vocal mechanisms of speech-motor control. This perturbation approach elicits compensatory corrections in voice fundamental frequency (F0) to perceived disparities between the pitch of auditory feedback and intended voice F0. This study examines the task-dependent role of auditory feedback in control of voice F0 in English speech and sustained vowels. Twenty-one native speakers of American English were asked to repeatedly inflect the pitch of either the first or second syllable of the English speech phrase "you know Nina," or produce non-word sustained vowels. Brief upward and downward perturbations in pitch of auditory feedback were introduced approximately 100 ms after vocal onset. Resultant changes in voice F0 due to perturbations in pitch were compared across vocal conditions and perturbation direction.;Results of this study indicate that auditory feedback is used in real-time to maintain voice F0 during normal English speech and non-word-sustained vowels. Incomplete compensatory corrections in voice F0 due to the perturbations in pitch of auditory feedback were prevalent during both dynamic speech (gain: ∼12.7%) and static sustained vowel phonations (gain: ∼19.3%), and appear to be regulated by the same underlying mechanism as indicated by similar response latency and magnitude. However, voice F0 can be modulated across syllable boundaries during dynamic speech in a task-dependent manner as demonstrated by the change in latency according to syllable pitch inflection. Thus, voice F0 responses help maintain the underlying suprasegmental meaning of a speech phrase by delaying the onset of the voice F0 response to coincide with the pitch-inflected syllable. Phase lags in the speech phrase (∼0.016 sec) were also prevalent when the direction and onset of a perturbation in pitch did not match the direction and onset of the pitch-inflected syllable. A linear addition of a steady-state voice F0 response from a sustained vowel with the dynamic voice F 0 contour of English speech did not account for the observed phase lags. In general, these data have important implications in the development of more complete neural models of normal and disordered voice F0 control.