Presentation
Search Abstracts | Symposia | Slide Sessions | Poster Sessions | Poster Slams
Real-time feedforward and feedback lateralization in speech motor control
Poster D33 in Poster Session D with Social Hour, Friday, October 7, 5:30 - 7:15 pm EDT, Millennium Hall
Also presenting in Poster Slam D, Friday, October 7, 5:15 - 5:30 pm EDT, Regency Ballroom
Francesco Mantegna1, Joan Orpella1, David Poeppel1,2; 1New York University, 2Ernst Struengmann Institute
Speaking requires the accurate coordination of multiple articulators at fast temporal rates. The ease with which we speak belies the complexity of the motor task. One foundational hypothesis is that speaking feels effortless because we use both feedforward and feedback mechanisms that allow us to anticipate the sensory consequences of speech and apply corrections as needed. The State Feedback Control (SFC) model suggests that these control mechanisms can be deployed not only when speech movements are executed but also in the absence of movement. In other words, feedforward and feedback mechanisms can be based on the external as well as the internal estimation of speech consequences. This internal estimation is crucial for many cognitive tasks such as motor preparation, internal rehearsal, and sensory prediction. Although feedback perturbation studies provide indirect evidence in the context of external estimation, direct evidence for both internal and external estimation is largely lacking. Speech imagery is an ideal tool for investigating speech motor control, assuming that it involves the same stages associated with overt speech except for motor execution. Here, we use speech imagery and magnetoencephalography to investigate the temporal dynamics underlying the internal estimation process. In particular, because of the distributed nature of the computational tasks (i.e., interregional communication between motor and sensory areas) we investigated how functional connectivity unfolds over time. Participants (N=45) imagined isolated syllables immediately after visual cue presentation. In line with the limb motor control literature, we observed that imagined speech was associated with alpha-beta power suppression with respect to the baseline, arguably reflecting internal motor execution. We considered this spectral signature as a reference, and we expected to find connectivity patterns reflecting feedforward and feedback control in the preceding and following time-windows, respectively. We measured functional connectivity across multiple regions in the peri-Sylvian language network in different frequency bands. We used a connectivity measure (weighted phase lag index) which accounts for the volume conductance problem inherent to electrophysiological measurements. Then, we identified network components that were consistent across subjects using non-negative matrix factorization. Our results reveal that alpha/beta power suppression was preceded by a left-lateralized network component in alpha and beta bands and was followed by a right-lateralized network component in the gamma band. The variance explained by these two network components in the current dataset was significantly higher than a control dataset obtained using surrogates having the same spectral properties but randomized phase relationships. The observed lateralization pattern is in line with an emerging trend in the literature suggesting that feedforward control is left-lateralized while feedback control is right-lateralized. Moreover, the spectral profiles of the observed connectivity patterns are consistent with the biophysical implementation proposed for the predictive coding model. Overall, our findings support the SFC model by showing that feedforward and feedback control mechanisms are used for internal estimation during speech imagery. Moreover, we provide, to our knowledge for the first time, a time-resolved characterization of the functional connectivity network underlying internal estimation processes.
Topic Areas: Speech Motor Control, Language Production