Predictive Timing in Speech Perception and Production
PI: Philip Hoole and Simone Falk
Collaborator: Simone Dalla Bella (Montreal)
2017-2023
contact: Phil Hoole – hoole@phonetik.uni-muenchen.de
Project Description
The production of speech is one of the fastest and most complex motor achievements humans are capable of. Fluent articulation is linked to successful sensorimotor integration guided by predictions about the sensory consequences of actions. These predictions include information both about the expected sensory consequences of a speaker s own actions and the time at which various landmarks in the speech signal should occur. The temporal anticipation and precise motor timing are what we call predictive timing. To date predictive timing in speech is poorly understood. The present proposal addresses fundamental issues in how predictive timing is used to couple action and perception in speech. Our first aim is to investigate whether there is a common cognitive basis to predictive timing in speech and non-verbal behavior such as music, and the extent to which individuals vary in their predictive timing abilities across these domains. To this end, individual predictive timing profiles (i.e., event-based and duration-based timing mechanisms) in perception and production with verbal and non-verbal auditory stimuli will be established. These profiles will be related to the behavior of speakers when predictive timing in speech is disrupted. This is investigated by focusing on predictive timing deficits in stuttering and adaptive behavior to auditory feedback manipulation. With respect to stuttering, our second aim is to develop a unified model of predictive timing in speech production of typically fluent individuals and persons who stutter. Predictive timing abilities and potential deficits are examined in groups of stuttering children and adolescents, using tasks in which fluent speech is synchronized with a periodic external rhythm (i.e., paced speech). The third aim of the proposal is concerned with testing predictive timing in unpaced speech when temporal information in auditory feedback is perturbed: specifically on how adaptation to perturbed feedback is constrained by syllable structure in fluent and disfluent speakers. The proposal will open up new horizons for understanding the role of predictive timing in speech perception and production. Results will show whether predictive timing is deficient in stuttering and which timing mechanisms are involved. Furthermore, the proposal will shed light on the structures (i.e., on the syllabic level) at which predictive timing operates. The proposal provides the foundations for integrating models of the temporal control of speech as developed in the articulatory phonology framework with those that have focused on forward prediction
and the role of feedback in speech production such as the DIVA model.