Neural Noise- Pitch centering (CNS correction of early pitch deviations toward an intended target) was examined in 24 individuals with laryngeal dystonia (LD) vs 29 controls (Frontiers in Human Neuroscience).1
- Provides a window into neural control of spontaneous speech and informs mechanisms relevant to future speech prostheses.1
- LD participants (20 adductor LD, 3 abductor LD, 1 both) and healthy controls were matched on age (LD mean 54.1 ± 10.5 y; controls 55.6 ± 14.6 y; P = 0.658), sex (P = 0.416), and race (P = 0.715); participants vocalized the vowel /a/ for 2.5 s over 74 trials inside a MEG scanner at UCSF.1
- Pitch was analyzed during the first 200 ms pre-perturbation window before any pitch shift was applied; trials were classified by initial pitch deviation tercile (lower, central, upper) to isolate centering magnitude as a function of starting deviation.1
- LD individuals showed significantly greater initial pitch deviations than controls (LS-mean difference = 6.36 cents, 95% CI [3.82, 8.89], P < 0.0001) and larger overall pitch movement magnitudes (difference = 8.63 cents, 95% CI [6.31, 10.95], P < 0.0001).1
- The primary finding: LD individuals demonstrated significantly enhanced pitch centering magnitude versus controls in centering trials (difference = 4.94 cents, 95% CI [2.74, 7.13], P < 0.0001), with a significant group × tercile interaction (P = 0.028), indicating the enhancement was especially pronounced for trials with larger initial deviations.1
- LD individuals also showed significantly greater anticentering magnitude (movement away from median pitch) in anticentering trials (difference = 5.68 cents, 95% CI [2.55, 8.81], P = 0.001), with a significant group × tercile interaction (P = 0.019), suggesting the speech motor system in LD is both over-correcting and misdirecting.1
- The proportion of centering vs. anticentering trials did not differ between groups (LD: 67.26% centering; controls: 67.4%), indicating the enhancement is in magnitude of correction rather than frequency of correction attempts.1
- Mechanistic interpretation: findings support disinhibition and hyperactivity in the laryngeal motor cortex (reduced intracortical inhibition, fMRI hyperactivation during phonation) plus abnormal auditory prediction generation—pre-phonatory MEG signals in superior/middle temporal gyri indicate disrupted auditory predictions even before voice onset.1
- The authors propose that elevated controller gain (stronger corrective responses per unit of deviation) explains enhanced centering, while elevated controller noise (increased variability in motor commands) explains the concurrent increase in anticentering, supporting a unified rather than dual control model.1
- First authorship is shared by Rabab Rangwala and Anantajit Subrahmanya (UCSF); co-authors include Kristina Simonyan, John F. Houde, and Srikantan S. Nagarajan; data and analysis pipeline are publicly available at OSF (https://osf.io/5umr4/).[^1]