Student or educator? Sign up with your school email and get 6 months of Pro free.
CoChat

The DOI matches its Crossref record; the references it cites are not checked yet.

Neural control of lexical tone production in human laryngeal motor cortex

Open access

Abstract

In tonal languages, which are spoken by nearly one-third of the world’s population, speakers precisely control the tension of vocal folds in the larynx to modulate pitch in order to distinguish words with completely different meanings. The specific pitch trajectories for a given tonal language are called lexical tones. Here, we used high-density direct cortical recordings to determine the neural basis of lexical tone production in native Mandarin-speaking participants. We found that instead of a tone category-selective coding, local populations in the bilateral laryngeal motor cortex (LMC) encode articulatory kinematic information to generate the pitch dynamics of lexical tones. Using a computational model of tone production, we discovered two distinct patterns of population activity in LMC commanding pitch rising and lowering. Finally, we showed that direct electrocortical stimulation of different local populations in LMC evoked pitch rising and lowering during tone production, respectively. Together, these results reveal the neural basis of vocal pitch control of lexical tones in tonal languages.

Overview

The claims and statistics of this paper haven’t been checked yet.

Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.

A few free checks a day · papers already verified open instantly · results are public

Integrity signals

Registry

Crossref · 10.1038/s41467-023-42175-9

Cites

Statistics

Figures

Fig. 1 Differential neural activity patterns are represented in single electrodes from LMC during lexical tone production. a Anatomical and physiological basis of vocal pitch control (copyright Jinxuan Wu, used with permission). When the subject is uttering syllables of four different Mandarin tones, the vocal cords are pushed together and vibrated by the passing air to produce the sound (voicing), cricothyroid stretches and increases the tension of vocal folds to raise the pitch, while thyroarytenoid contracts and decreases the tension of vocal folds to lower the pitch. b The flowchart of the lexical tone production task (copyright Zehao Zhao, used with permission). In each trial, the participants were presented with a randomly displayed tonal syllable as an auditory cue (0.44 s). Following a reaction time interval (1.0-1.2 s), the patients were instructed to repeat the word three times consecutively (0.4–0.6 s per word) with intervals of 0.4–0.6 s between repetitions. The next auditory cue was played after 1.0–1.5 s. Each patient completed 60-160 trials, resulting in 45–120 repetitions per lexical tone. c The average pitch dynamics (arbitrary units [a.u.]) for four lexical tones produced by all subjects. Upper panel: the average trajectory of pitch height (solid line: mean, shaded area: standard errors of the mean [s.e.m.] across all repetitions for each tone of all subjects); lower panel: the average trajectory of pitch change (solid line: mean, shaded area: s.e.m. across all repetitions for each tone of all subjects). The 0 ms time point represents the onset of the vowel. d The distribution of all tone discriminant electrodes. Different symbols represent electrodes from different subjects (S1-S8). Color bar indicates the degree of tone discrimination (F-statistics across 4 tones). e The average high-gamma responses for different lexical tone production (shaded area represents mean ± s.e.m. across repetitions of each tone) in two example electrodes (locations indicated by black arrows in (c). The 0 ms time point represents the onset of the vowel. Black dots indicate time points of significant difference across 4 tones (p < 0.05, one-way ANOVA, Bonferroni corrected). Source data are provided as a Source Data file.
Fig. 1 Differential neural activity patterns are represented in single electrodes from LMC during lexical tone production. a Anatomical and physiological basis of vocal pitch control (copyright Jinxuan Wu, used with permission). When the subject is uttering syllables of four different Mandarin tones, the vocal cords are pushed together and vibrated by the passing air to produce the sound (voicing), cricothyroid stretches and increases the tension of vocal folds to raise the pitch, while thyroarytenoid contracts and decreases the tension of vocal folds to lower the pitch. b The flowchart of the lexical tone production task (copyright Zehao Zhao, used with permission). In each trial, the participants were presented with a randomly displayed tonal syllable as an auditory cue (0.44 s). Following a reaction time interval (1.0-1.2 s), the patients were instructed to repeat the word three times consecutively (0.4–0.6 s per word) with intervals of 0.4–0.6 s between repetitions. The next auditory cue was played after 1.0–1.5 s. Each patient completed 60-160 trials, resulting in 45–120 repetitions per lexical tone. c The average pitch dynamics (arbitrary units [a.u.]) for four lexical tones produced by all subjects. Upper panel: the average trajectory of pitch height (solid line: mean, shaded area: standard errors of the mean [s.e.m.] across all repetitions for each tone of all subjects); lower panel: the average trajectory of pitch change (solid line: mean, shaded area: s.e.m. across all repetitions for each tone of all subjects). The 0 ms time point represents the onset of the vowel. d The distribution of all tone discriminant electrodes. Different symbols represent electrodes from different subjects (S1-S8). Color bar indicates the degree of tone discrimination (F-statistics across 4 tones). e The average high-gamma responses for different lexical tone production (shaded area represents mean ± s.e.m. across repetitions of each tone) in two example electrodes (locations indicated by black arrows in (c). The 0 ms time point represents the onset of the vowel. Black dots indicate time points of significant difference across 4 tones (p < 0.05, one-way ANOVA, Bonferroni corrected). Source data are provided as a Source Data file.CC BY · PubMed Central
Fig. 2 Tone discriminability from the neural activity is correlated with pitch height and pitch change representation. a–e, g, h Scatter plot of single electrode encoding properties across speech-selective electrodes from all participants. Each dot represents a single electrode (n = 448). The colored dots indicate electrodes that had significant encoding of x-axis (blue, p < 0.005, one-sided permutation test), y-axis (red, p < 0.05, one-way ANOVA, Bonferroni corrected) encoding features or both (magenta), while the black dots indicate non-significant electrodes. The r and P values are computed among the significant electrodes (colored dots) using Pearson’s correlation (two-sided) between the x- and y-axis. a Scatterplot of the unique variance explained by pitch features (R2 of pitch change and pitch height) and tone discriminability (F-statistics across 4 tones). b Scatterplot of the unique variance explained by intensity (R2) and tone discriminability (F-statistics). c Scatterplot of the unique variance explained by syllable onsets (R2) and tone discriminability (F-statistics). d Scatterplot of the unique variance explained by binary pitch feature (R2) and tone discriminability (F-statistics). e Scatterplot of the unique variance explained by tone category (R2) and tone discriminability (F-statistics). f Scatterplot of the unique variance explained by pitch change (R2) and pitch height (R2). The colored dots indicate electrodes that had significant encoding of either pitch change or pitch height (p < 0.005, one-sided permutation test), with color indicating the proportion of variance explained by pitch height relative to the total variance explained by pitch change and pitch height. g Scatterplot of the unique variance explained by pitch height (R2) and tone discriminability (F-statistics). h Scatterplot of the unique variance explained by pitch change (R2) and tone discriminability (F-statistics). Source data are provided as a Source Data file.
Fig. 2 Tone discriminability from the neural activity is correlated with pitch height and pitch change representation. a–e, g, h Scatter plot of single electrode encoding properties across speech-selective electrodes from all participants. Each dot represents a single electrode (n = 448). The colored dots indicate electrodes that had significant encoding of x-axis (blue, p < 0.005, one-sided permutation test), y-axis (red, p < 0.05, one-way ANOVA, Bonferroni corrected) encoding features or both (magenta), while the black dots indicate non-significant electrodes. The r and P values are computed among the significant electrodes (colored dots) using Pearson’s correlation (two-sided) between the x- and y-axis. a Scatterplot of the unique variance explained by pitch features (R2 of pitch change and pitch height) and tone discriminability (F-statistics across 4 tones). b Scatterplot of the unique variance explained by intensity (R2) and tone discriminability (F-statistics). c Scatterplot of the unique variance explained by syllable onsets (R2) and tone discriminability (F-statistics). d Scatterplot of the unique variance explained by binary pitch feature (R2) and tone discriminability (F-statistics). e Scatterplot of the unique variance explained by tone category (R2) and tone discriminability (F-statistics). f Scatterplot of the unique variance explained by pitch change (R2) and pitch height (R2). The colored dots indicate electrodes that had significant encoding of either pitch change or pitch height (p < 0.005, one-sided permutation test), with color indicating the proportion of variance explained by pitch height relative to the total variance explained by pitch change and pitch height. g Scatterplot of the unique variance explained by pitch height (R2) and tone discriminability (F-statistics). h Scatterplot of the unique variance explained by pitch change (R2) and tone discriminability (F-statistics). Source data are provided as a Source Data file.CC BY · PubMed Central
Fig. 3 Two tuning commands exist in the LMC to generate different tone control of vocal pitch. a The application of Fujisaki model. First row: example syllable waveforms of different tones; second row: original pitch trajectories; third row: positive (red) and negative (blue) tone commands extracted using Fujisaki model; fourth row: reconstructed pitch trajectories using Fujisaki model (black curve), compared to original trajectories. b Scatter plot of single electrode encoding of tone command features (R2) versus tone discrimination (F-statistics) across speech-selective electrodes from all participants. Red dots indicate electrodes that had significant tone discrimination only (p < 0.05, one-way ANOVA, Bonferroni corrected); Blue dots indicate electrodes that had significant encoding of tone command only (p < 0.005, one-sided permutation test); Magenta dots indicate electrodes that had significant encoding of both tone discrimination and tone command. The r and P values are computed among significant electrodes (colored dots) using Pearson’s correlation. c Left: average neural tuning curve for negative tuning electrodes (dashed line: mean high-gamma response, shaded area: s.e.m.). Solid curve corresponds to an example electrode e1. Right: average high-gamma activity for different tones (shaded area represents mean ± s.e.m. across repetitions of each tone) in electrode e1. d Left: average neural tuning curve for positive tuning electrodes (dashed line: mean high-gamma response, shaded area: s.e.m.). Solid curve corresponds to an example electrode e2. Right: average high-gamma activity for different tones (shaded area represents mean ± s.e.m. across repetitions of each tone) in electrode e2. In (c, d), 0-ms time point represents the vowel onset. Black dots indicate time points of significant difference across 4 tones (p < 0.05, one-way ANOVA, Bonferroni corrected). e Distribution of positive (red) and negative (blue) tuning electrodes on the cortical surface. Different symbols represent significant tone tuning electrodes from different subjects. Darker color indicates stronger tuning. The example electrodes e1 and e2 are marked by black arrows. Blue dashed line represents the boundary line separating the dLMC and vLMC clusters (indicated by gray dashed circles). See also Fig. S5. Source data are provided as a Source Data file.
Fig. 3 Two tuning commands exist in the LMC to generate different tone control of vocal pitch. a The application of Fujisaki model. First row: example syllable waveforms of different tones; second row: original pitch trajectories; third row: positive (red) and negative (blue) tone commands extracted using Fujisaki model; fourth row: reconstructed pitch trajectories using Fujisaki model (black curve), compared to original trajectories. b Scatter plot of single electrode encoding of tone command features (R2) versus tone discrimination (F-statistics) across speech-selective electrodes from all participants. Red dots indicate electrodes that had significant tone discrimination only (p < 0.05, one-way ANOVA, Bonferroni corrected); Blue dots indicate electrodes that had significant encoding of tone command only (p < 0.005, one-sided permutation test); Magenta dots indicate electrodes that had significant encoding of both tone discrimination and tone command. The r and P values are computed among significant electrodes (colored dots) using Pearson’s correlation. c Left: average neural tuning curve for negative tuning electrodes (dashed line: mean high-gamma response, shaded area: s.e.m.). Solid curve corresponds to an example electrode e1. Right: average high-gamma activity for different tones (shaded area represents mean ± s.e.m. across repetitions of each tone) in electrode e1. d Left: average neural tuning curve for positive tuning electrodes (dashed line: mean high-gamma response, shaded area: s.e.m.). Solid curve corresponds to an example electrode e2. Right: average high-gamma activity for different tones (shaded area represents mean ± s.e.m. across repetitions of each tone) in electrode e2. In (c, d), 0-ms time point represents the vowel onset. Black dots indicate time points of significant difference across 4 tones (p < 0.05, one-way ANOVA, Bonferroni corrected). e Distribution of positive (red) and negative (blue) tuning electrodes on the cortical surface. Different symbols represent significant tone tuning electrodes from different subjects. Darker color indicates stronger tuning. The example electrodes e1 and e2 are marked by black arrows. Blue dashed line represents the boundary line separating the dLMC and vLMC clusters (indicated by gray dashed circles). See also Fig. S5. Source data are provided as a Source Data file.CC BY · PubMed Central
Fig. 4 Tone can be decoded from distributed neural activity. a Time course of mean pair-wise tone classification accuracy using a sliding time window of 100 ms (solid line: mean, shaded area: s.e.m. across subjects S1-S8) and population neural activity in all speech-responsive electrodes in bilateral LMC. The 0 ms time point represents the onset of the vowel. Horizontal dashed line: p = 0.05 threshold, two-sided t-test for independent samples, Bonferroni corrected; vertical dashed line: peak accuracy time. b Averaged classifier weights at peak accuracy time on each electrode. Each dot represents one electrode (different symbols represent electrodes from different subjects), with a darker color indicating a larger absolute classifier weight. Source data are provided as a Source Data file.
Fig. 4 Tone can be decoded from distributed neural activity. a Time course of mean pair-wise tone classification accuracy using a sliding time window of 100 ms (solid line: mean, shaded area: s.e.m. across subjects S1-S8) and population neural activity in all speech-responsive electrodes in bilateral LMC. The 0 ms time point represents the onset of the vowel. Horizontal dashed line: p = 0.05 threshold, two-sided t-test for independent samples, Bonferroni corrected; vertical dashed line: peak accuracy time. b Averaged classifier weights at peak accuracy time on each electrode. Each dot represents one electrode (different symbols represent electrodes from different subjects), with a darker color indicating a larger absolute classifier weight. Source data are provided as a Source Data file.CC BY · PubMed Central
Fig. 5 Direct electrical cortical stimulation of dLMC elicits pitch rising and lowering. a Acoustic waveform and absolute pitch diagrams of two syllable production paradigms. The pitch of normal /mā/ decreases slowly (control, marked in gray), and the pitch rising response is defined as the direct electrical stimulation (DES, black dotted line) evoked increase in pitch (marked in red) when the patient is pronouncing /mā/. The typical /má/ pitch drops first and then rises rapidly (control, marked in gray). The pitch lowering response is defined as the DES (black dotted line) evoked decreases in absolute pitch (marked in blue) relative to the control group. b–d In a typical patient, the pitch rising and pitch lowering responses were both evoked in the patient’s left dLMC. b Cortical representation of the pitch rising (the red Label 1) and pitch lowering (the blue Label 2) sites in this typical case. c Average pitch contours of the electrical stimulation-induced pitch rising response at T1 (in red) compared to the average pitch contours of normal /mā/ (in gray). The horizontal dotted line and shaded area represent the mean ± standard error of the mean (s.e.m.) across repetitions. The vertical dotted line represents the average start time of electrical stimulation (mean ± s.e.m.: 452 ± 49 ms). d Average pitch contours of the electrical stimulation-induced pitch lowering response at T2 (in blue) compared to the average pitch contours of normal /má/ (in gray). The horizontal dotted line and shaded area represent the mean ± s.e.m. across repetitions. The vertical dotted line represents the average start time of electrical stimulation (mean ± s.e.m.: 203 ± 19 ms). The red solid line indicates time points survived p < 0.05, two-sided paired t test, FDR correction). e, f The group-level cortical distributions of pitch rising (red), pitch lowering (blue), speech arrest (purple) and other sites (coded by distinct colors) on the (e) left and (f) right hemispheres. Source data are provided as a Source Data file.
Fig. 5 Direct electrical cortical stimulation of dLMC elicits pitch rising and lowering. a Acoustic waveform and absolute pitch diagrams of two syllable production paradigms. The pitch of normal /mā/ decreases slowly (control, marked in gray), and the pitch rising response is defined as the direct electrical stimulation (DES, black dotted line) evoked increase in pitch (marked in red) when the patient is pronouncing /mā/. The typical /má/ pitch drops first and then rises rapidly (control, marked in gray). The pitch lowering response is defined as the DES (black dotted line) evoked decreases in absolute pitch (marked in blue) relative to the control group. b–d In a typical patient, the pitch rising and pitch lowering responses were both evoked in the patient’s left dLMC. b Cortical representation of the pitch rising (the red Label 1) and pitch lowering (the blue Label 2) sites in this typical case. c Average pitch contours of the electrical stimulation-induced pitch rising response at T1 (in red) compared to the average pitch contours of normal /mā/ (in gray). The horizontal dotted line and shaded area represent the mean ± standard error of the mean (s.e.m.) across repetitions. The vertical dotted line represents the average start time of electrical stimulation (mean ± s.e.m.: 452 ± 49 ms). d Average pitch contours of the electrical stimulation-induced pitch lowering response at T2 (in blue) compared to the average pitch contours of normal /má/ (in gray). The horizontal dotted line and shaded area represent the mean ± s.e.m. across repetitions. The vertical dotted line represents the average start time of electrical stimulation (mean ± s.e.m.: 203 ± 19 ms). The red solid line indicates time points survived p < 0.05, two-sided paired t test, FDR correction). e, f The group-level cortical distributions of pitch rising (red), pitch lowering (blue), speech arrest (purple) and other sites (coded by distinct colors) on the (e) left and (f) right hemispheres. Source data are provided as a Source Data file.CC BY · PubMed Central

Verification

Claims the paper makes

The claims and statistics of this paper haven’t been checked yet.

Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.

A few free checks a day · papers already verified open instantly · results are public

Claims other papers make about this one

Each is checked against this paper’s own text: does it say what the citing paper says it does?

Self-check

Percentages, sums and reported differences re-computed from the paper’s own numbers.

The self-check needs the paper’s full text.

Citations

The citation lists are queued

They are fetched later and appear here then.

References 40 · Cited by 26

Research works printed but not linked: 5

  1. Chao, Y. R. A grammar of spoken Chinese (University of California Press, 1965).
  2. Howie, J. M. & Howie, J. M. Acoustical Studies of Mandarin Vowels and Tones Vol. 18 (Cambridge University Press, 1976).
  3. Fujisaki, H., Ohno, S. & Gu, W. Physiological and physical mechanisms for fundamental frequency control in some tone languages and a command-response model for generation of their F0 contours. In International Symposium on Tonal Aspects of Languages: With Emphasis on Tone Languages (2004).
  4. Boersma, P. Accurate short-term analysis of the fundamental frequency and the harmonics-to-noise ratio of a sampled sound. In Proceedings of the Institute of Phonetic Sciences 17, 97–110 (Amsterdam, 1993).
  5. Mixdorff, H. FujiParaEditor. TFH Berlin University of Applied Sciences. http://public.bht-berlin.de/~hmixdorff/thesis/fujisaki.html (2009).

Metadata

Authors
Lu, J., Li, Y., Zhao, Z., Liu, Y., Zhu, Y., Mao, Y., Wu, J., & Chang, E. F.
Affiliations
  • Shanghai Medical College of Fudan University (CN)
  • University of California, San Francisco (US)
Corresponding author
  • Mao, Ying
  • Wu, Jinsong
  • Chang, Edward F.
Venue
Nature Communications
Published
2023-10-30
Volume
14
Issue
1
Pages
6917-6917
ISSN
2041-1723 (ISSN-L)
OpenAlex
W4388015458
Cited by
26
References
40
Citations by year
2026: 6 · 2025: 16 · 2024: 4
FWCI
6.70
Type
journal article
Language
en
Open access
Licence
Locations
  • Nature Communications (publishedVersion, cc-by)
  • PubMed Central (publishedVersion, cc-by)
  • PubMed (publishedVersion)
  • eScholarship (California Digital Library) (submittedVersion, cc-by)
  • RePEc: Research Papers in Economics (submittedVersion)
  • DOAJ (DOAJ: Directory of Open Access Journals) (submittedVersion, cc-by-sa)
Funders
  • U.S. Department of Health & Human Services | NIH | National Institute of Neurological Disorders and Stroke (U01 NS117765)
  • Chinese Ministry of Science and Technology | Department of S and T for Social Development (2022ZD0212300)
  • Shanghai Science and Technology Development Foundation (21PJD007)
  • Science and Technology Commission of Shanghai Municipality (22PJ1410500, 2018SHZDZX01)
  • National Natural Science Foundation of China (32371154)
  • Shanghai Municipal Education Commission (2023ZKZD13)
Topics
  • Social Sciences › Psychology › Experimental and Cognitive Psychology › Phonetics and Phonology Research
  • Life Sciences › Neuroscience › Cognitive Neuroscience › Neurobiology of Language and Bilingualism
  • Life Sciences › Agricultural and Biological Sciences › Ecology, Evolution, Behavior and Systematics › Animal Vocal Communication and Behavior
Keywords
tonal languages, Mandarin Chinese, vocal fold tension, pitch control, articulatory kinematics, population coding, direct cortical stimulation
MeSH terms
Humans, Language, Larynx, Motor Cortex, Pitch Perception, Speech Perception

About these checks

Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.

Statistics and arithmetic are recomputed, not judged.

Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.

What counts as a claim ›How we measure confidence ›How we check this ›

Are you an author, or do you see a problem with this page? Request a review

How to cite

Cite CoChat for the checks on this page, and the paper for its own findings.

This page
CoChat. (2026). Citation check: Neural control of lexical tone production in human laryngeal motor cortex. https://app.cochat.ai/papers/10.1038/s41467-023-42175-9
The paper
Lu, J., Li, Y., Zhao, Z., Liu, Y., Zhu, Y., Mao, Y., Wu, J., & Chang, E. F. (2023). Neural control of lexical tone production in human laryngeal motor cortex. Nature Communications. https://doi.org/10.1038/s41467-023-42175-9

Data sources

Bibliographic data: Semantic Scholar, OpenAlex, Crossref.

Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).

Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.