Search Abstracts | Symposia | Slide Sessions | Poster Sessions
Neural modulation of pitch and formants during altered feedback
Poster Session B, Friday, September 12, 4:30 - 6:00 pm, Field House
Heather Kabakoff1, Amirhossein Khalilian-Gourtani1, Sarah Lewkowict2, Erika Jensen1, Caroline A. Niziolek3, Daniel Friedman1, Adeen Flinker1,4; 1New York University School of Medicine, 2New York University, 3University of Wisconsin-Madison, 4New York University School of Engineering
Precise and intelligible speech output depends on intact cortical-subcortical circuits controlling the ability to incorporate auditory feedback into ongoing motor commands. When auditory feedback is perturbed, a typical behavioral reaction is to compensate for the spectral perturbations (pitch/formants) by shifting frequencies in the opposite direction to the perturbation. There is substantial evidence from functional magnetic resonance imaging (fMRI) for early error detection in superior temporal gyrus (STG) and subsequent motor correction in sensorimotor regions in response to both pitch and formant perturbations (e.g., Niziolek & Guenther, 2013). However, fMRI lacks the temporal resolution to precisely characterize the dynamics of this circuit. One study employing electrocorticography has shown that pitch perturbations elicit feedback-related responses in posterior STG and near supramarginal gyrus (SMG) (Chang et al., 2013). However, precise neural mechanisms and dynamics of responses to pitch versus formant perturbations have not been investigated within the same individuals. In the present study, we used stereotactic electroencephalography (sEEG) to map within-subject neural responses to pitch (±200 cents) and first formant frequency (±250 mels) perturbations in patients with Epilepsy undergoing clinical monitoring. At bedside, patients sustained /ɛ/ in words while perturbations were applied in real time. We analyzed high gamma broadband neural signal (70–150 Hz), a robust marker correlated with BOLD and single unit responses, to characterize response to pitch and formant perturbations. From 875 electrodes spanning cortex across nine patients, unsupervised clustering revealed three distinct general patterns of spatiotemporal neural activity. A cluster localized to bilateral STG and left SMG showed early auditory error detection for both pitch and formant perturbation conditions, characterized by a brief increase in high gamma activity between 0-400 ms after perturbation onset. A second cluster centered in bilateral Heschl’s gyrus showed sustained neural activity (0-1500 ms) in both perturbation conditions, consistent with prolonged auditory feedback monitoring and suggesting involvement in transmitting error signals to motor-related regions. Interestingly, a third cluster in temporal pole and inferior frontal cortex showed sustained suppression across the trial, consistent with default mode network deactivation. To quantify perturbation-based effects, a one-way ANOVA for each time point grouped by condition (Bonferroni corrected for multiple comparisons, with ≥100ms of continuous significance) revealed a group of 28 electrodes in Heschl’s, superior temporal, and insular gyri significantly encoding differences between perturbation conditions (pitch/formant/none). Among significant electrodes, post-hoc analysis identified 24 of these electrodes with significantly higher magnitude response to pitch than formant perturbations (p<0.002), with peak difference at 156±64ms after perturbation onset. This is the first study employing the high spatiotemporal resolution of sEEG to investigate neural mechanisms of both laryngeal phonatory (pitch) and oral articulatory (formant) control in the same individuals. We provide evidence for the roles of STG and surrounding areas in detecting error signals and ongoing feedback monitoring in response to pitch and formant perturbations. Further, we identify neural regions showing magnitude-based differences distinguishing sensorimotor control of phonatory versus articulatory subsystems, a finding with implications toward development of neuromodulation therapies for targeted sensorimotor learning in these domains in patients with motor speech disorders.
Topic Areas: Multisensory or Sensorimotor Integration, Speech Motor Control