The DOI matches its Crossref record; 45 of the 45 references it cites were found in a registry or an index; 9 of the 44 sources its claims cite were checked for support.
A deep learning approach for acoustic-based identification of muscle tension dysphonia and spasmodic dysphonia
Open accessAbstract
Problem: Differentiating between spasmodic dysphonia (SD), a neurological disorder, and muscle tension dysphonia (MTD), a behavioral voice disorder, based on auditory perception alone is a common clinical challenge. This diagnostic difficulty can lead to delays in appropriate treatment, highlighting the need for objective and reliable assistive diagnostic tools. Aim: This study aims to develop and validate an artificial intelligence (AI) model based on deep learning to automatically differentiate between healthy voices, SD, and MTD using only voice audio recordings, and to compare its diagnostic performance against human experts. Methods: A retrospective analysis was conducted on 1,597 voice samples (595 healthy, 471 MTD, 531 SD). Voice audio was processed into Log-Mel spectrograms. Pre-trained convolutional neural networks (CNNs), including VGG16, ResNet50, and DenseNet161, were employed for transfer learning to perform both binary (Healthy vs. Disordered) and ternary (Healthy vs. MTD vs. SD) classification. The model's performance was evaluated on a held-out test set and compared to the diagnostic assessments of four otolaryngologists. Results: The AI model achieved an accuracy of 89.5% (AUC = 0.956) in distinguishing healthy voices from disordered ones. For the more complex ternary classification, the model attained an accuracy of 71.6%, with class-specific AUCs of 0.957 (Healthy), 0.731 (MTD), and 0.855 (SD). This performance surpassed that of human experts, who achieved average accuracies of 78.2% in binary classification and 60.6% in ternary classification on the same test set. Conclusion: The deep learning model trained on a Mandarin pathological voice dataset achieves favorable performance in distinguishing SD from MTD using only voice audio signals, with classification results comparable to those of experienced clinical specialists. This technology serves as a promising objective auxiliary tool for the preliminary screening and differential diagnosis of voice disorders.
Overview
CoChat Summary
TL;DR
This retrospective study trained deep learning models on 1,597 voice samples to distinguish healthy voices, muscle tension dysphonia (MTD), and spasmodic dysphonia (SD) from audio alone. The best model outperformed four human experts on both binary and three-class tasks, but the paper notes that its acoustic-only design is an initial screening tool rather than a full clinical diagnosis.
Key findings
- On binary classification (healthy vs. disordered), the model achieved 89.5% accuracy, 92.2% sensitivity, 85% specificity, and AUC = 0.956.
- On three-class classification (healthy vs. MTD vs. SD), the model achieved 71.6% accuracy, 70.2% sensitivity, 85.8% specificity, with class-specific AUCs of 0.957 (Healthy), 0.731 (MTD), and 0.855 (SD).
- Human experts averaged 78.2% accuracy on the binary task and 60.6% on the three-class task; individual three-class accuracies were 59%, 60.5%, 59%, and 57.1%.
- The paper reports the model inference was completed within seconds, whereas four specialists spent several hours evaluating the 324 test recordings.
Methods
- Retrospective study using 1,597 voice samples: 595 healthy, 471 MTD, and 531 SD, collected from two hospitals in China from December 2012 to April 2024.
- Audio was converted to Log-Mel spectrograms; three pre-trained CNNs (VGG16, ResNet50, DenseNet161) were fine-tuned with transfer learning and combined by averaging softmax probabilities.
- Data were split patient-level into 960 training, 313 validation, and 324 test samples; performance was evaluated with AUROC, accuracy, sensitivity/recall, specificity, precision, F1-score, ICC, and McNemar’s test, and compared against four otolaryngology specialists.
Limitations
- The paper states the sample is relatively small and limited to populations in southern China, which may restrict generalizability.
- The model uses audio only; the paper says comprehensive voice-disorder diagnosis usually also requires medical history, symptoms, laryngoscopic findings, auditory perceptual assessment, and other information.
- The paper notes it did not explicitly evaluate performance under real-world background noise or device variability; this is stated as a limitation and suggests robustness in deployment remains untested.
Model-written summary. Check it against the paper before relying on it.
Integrity signals
Registry
Cites
Statistics
Figures






Verification
Claims the paper makes
6 Valid · 1 Not valid · 10 Not checked · 64 Own results and interpretations
The cited sources support only part of this claim — read each source’s verdict below.
In the paper
“In clinical practice, the diagnosis of Spasmodic Dysphonia (SD) primarily relies on clinical professionals' auditory perceptual assessment—an inherently subjective process prone to variability due to differences in expertise and experience (41).”
§ A deep learning approach for acoustic-based identification of muscle tension dysphonia and spasmodic dysphonia > Materials and methods This study adheres to the principles of the Declaration of Helsinki and has been approved by the Ethics and Research Committees of Guangdong Provincial People's Hospital and Shenzhen People's Hospital. From December 2012 to April 2024, a total of 595 healthy voice audio samples and 1,002 pre-treatment voice disorder audio samples were collected from the Otolaryngology Departments of Guangdong Provincial People's Hospital and Shenzhen People's Hospital. The pre-treatment samples included MTD (n = 471) and SD (n = 531). All samples are clinically clearly diagnosed disease specimens, and those with other comorbidities are excluded. All patients and healthy individuals are aged between 18 and 80 years old with no gender restrictions. The samples were recorded at comfortable volume levels using Praat version 4.5, a Shure 8 microphone, in mono sound with a sampling rate of 44,100 Hz and 16-bit resolution. The data was saved in uncompressed.wav format. First, 1,597 samples were randomly divided into three groups in a 3:1:1 ratio, with 960 samples used as the training set, 313 samples as the internal validation set, and 324 samples as the test set. Through random grouping, 358 healthy voice audio samples, 282 MTD samples, and 320 SD samples were selected for the training set (see Table 2). We strictly adopted patient-level data splitting. All data of one patient were entirely assigned to either the training set, validation set or test set, with no patient crossover between different datasets. > Comparison with human experts
What the sources say
- Partly supportedcites [41] Creighton, 2015, Diagnostic delays in spasmodic dysphonia: a call for clinician education, J Voice sourceExperimental — accuracy not yet measured
The Discussion says SD diagnostic criteria are subjective and depend on individual practitioners’ interpretation of examination findings; it also calls for greater clinician education. It does not say that diagnosis primarily relies on auditory perceptual assessment or specifically attribute variability to differences in expertise and experience. (LLM, conf=0.87)
plainstates a cause- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Partly supported
The cited sources support only part of this claim — read each source’s verdict below.
In the paper
“while SD typically requires botulinum toxin injections to alleviate involuntary spasms, MTD responds well to voice therapy alone (12–16).”
§ A deep learning approach for acoustic-based identification of muscle tension dysphonia and spasmodic dysphonia > Introduction
What the sources say
- Supportedcites [12] Ludlow, 2009, Treatment approaches for spasmodic dysphonia: limitations of current approaches, Curr Opin Otolaryngol Head Neck Surg source
The Abstract calls botulinum toxin injection the gold standard for treating spasmodic dysphonia, and the treatment discussion says injections are used primarily in most countries. It also reports that injections reduce voice breaks, supporting the claim that they alleviate symptoms. The paper describes surgery as an alternative, so “requires” is best read here as describing typical treatment, not an absolute necessity. (LLM, conf=0.88)
- Not checked — no full text of the sourcecites [13] Berke, 1999, Selective laryngeal adductor denervation– reinnervation: a new surgical treatment for adductor spasmodic dysphonia, Ann Otol Rhinol Laryngol source
- Not checked — no full text of the sourcecites [14] Chhetri, 2006, Long-term follow-up results of selective laryngeal adductor denervation– reinnervation surgery for adductor spasmodic dysphonia, Laryngoscope source
- Not checked — no full text of the sourcecites [15] Roy, 1997, Manual circumlaryngeal therapy for functional dysphonia: an evaluation of short- and long-term treatment outcomes, J Voice source
- Partly supportedcites [16] Dejonckere, 2012, Tridimensional assessment of adductor spasmodic dysphonia pre- and post-treatment with Botulinum toxin, Eur Arch Otorhinolaryngol sourceExperimental — accuracy not yet measured
The Introduction describes SD as involving spasms and says botulinum toxin injections were the standard treatment for patients in this study. The Conclusion reports improvement after treatment. The paper does not establish that SD generally requires injections, so “typically requires” goes beyond its findings. (LLM, conf=0.82)
plainstates a cause- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Supported
The cited sources support only part of this claim — read each source’s verdict below.
In the paper
“while SD typically requires botulinum toxin injections to alleviate involuntary spasms, MTD responds well to voice therapy alone (12–16).”
§ A deep learning approach for acoustic-based identification of muscle tension dysphonia and spasmodic dysphonia > Introduction
What the sources say
- Partly supportedcites [12] Ludlow, 2009, Treatment approaches for spasmodic dysphonia: limitations of current approaches, Curr Opin Otolaryngol Head Neck Surg sourceExperimental — accuracy not yet measured
The Diagnosis of spasmodic dysphonia passage (page not provided) says functional dysphonia can often be effectively treated by voice therapy alone. The paper mentions muscular tension dysphonia elsewhere, but does not specifically say that MTD responds well to voice therapy alone; the claim is more specific and less qualified than the source. (LLM, conf=0.83)
- Not checked — no full text of the sourcecites [13] Berke, 1999, Selective laryngeal adductor denervation– reinnervation: a new surgical treatment for adductor spasmodic dysphonia, Ann Otol Rhinol Laryngol source
- Not checked — no full text of the sourcecites [14] Chhetri, 2006, Long-term follow-up results of selective laryngeal adductor denervation– reinnervation surgery for adductor spasmodic dysphonia, Laryngoscope source
- Not checked — no full text of the sourcecites [15] Roy, 1997, Manual circumlaryngeal therapy for functional dysphonia: an evaluation of short- and long-term treatment outcomes, J Voice source
- Partly supportedcites [16] Dejonckere, 2012, Tridimensional assessment of adductor spasmodic dysphonia pre- and post-treatment with Botulinum toxin, Eur Arch Otorhinolaryngol sourceExperimental — accuracy not yet measured
The Discussion (page unavailable) calls botulinum injection the gold-standard symptomatic treatment for adductor spasmodic dysphonia, supporting the SD portion of the cited sentence. The paper does not address whether muscle tension dysphonia (MTD) responds to voice therapy alone, so it does not support that portion. (LLM, conf=0.84)
boostedstates a cause- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
- Partly supported
A result the paper states — it cites no source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“Moreover, the proposed model demonstrated perfect consistency”
§ A deep learning approach for acoustic-based identification of muscle tension dysphonia and spasmodic dysphonia > Results
boosted- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: unclear — It is unclear whether the paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
A result the paper reports from its own data — it cites no other source for it, so there is none to check. The quote shows it is in the paper.
In the paper
“This performance surpassed that of human experts, who achieved average accuracies of 78.2% in binary classification and 60.6% in ternary classification on the same test set.”
§ A deep learning approach for acoustic-based identification of muscle tension dysphonia and spasmodic dysphonia > Abstract > Results
plain- Asserted: passes — The authors put it forward as true.
- Checkable: passes — It can be checked against evidence.
- Load-bearing: passes — The paper’s conclusions depend on it.
- Based: passes — The paper gives a reason why it holds.
40 more claims
Create a free account to see all 81 claims and every source verdict.
- It may state two separate claims as one. 26
- Probably not a substantial claim. 19
Self-check
Percentages, sums and reported differences re-computed from the paper’s own numbers.
Each contradiction between two passages listed here was confirmed by an independent check.
Rated major: it would touch a result — a value a result rests on, reported and recomputed p-values on opposite sides of the significance level, two passages that conflict with no sign which is right, or a cited work retracted before this paper was published.
In the paper
§ A deep learning approach for acoustic-based identification of muscle tension dysphonia and spasmodic dysphonia > Results
“The accuracy rates were 59% and 60.5% for the 2 senior laryngologists, 59% for the attending laryngologist and 57.1% for the speech therapist (ST).”
§ A deep learning approach for acoustic-based identification of muscle tension dysphonia and spasmodic dysphonia > Abstract > Results
“This performance surpassed that of human experts, who achieved average accuracies of 78.2% in binary classification and 60.6% in ternary classification on the same test set.”
Rated minor: it would not touch a result — a value no result rests on, a p-value that stays on the same side of the significance level, or a conflict that is clerical, resolved elsewhere in the paper, or outside its own findings.
In the paper
§ A deep learning approach for acoustic-based identification of muscle tension dysphonia and spasmodic dysphonia > Materials and methods > Statistical analysis
“Statistical analyses were conducted using Python 3.91 (40).”
§ A deep learning approach for acoustic-based identification of muscle tension dysphonia and spasmodic dysphonia > Materials and methods > Data preprocessing and model building
“The machine learning process was conducted using Python 3.9.1 and PyTorch 2.0.1 on Ubuntu 22.04. The technical roadmap is illustrated in the following Figure 1.”
Cited works that were retracted
Works this paper cites that were later retracted, from the retraction notices publishers deposit. Major when the notice came out before this paper was published.
Rated major: it would touch a result — a value a result rests on, reported and recomputed p-values on opposite sides of the significance level, two passages that conflict with no sign which is right, or a cited work retracted before this paper was published.
In the paper
“Zakariah M, 2022, An analytical study of speech pathology detection based on MFCC and deep neural networks, Comput Math Methods Med, 2022, 7814952, 10.1155/2022/7814952”
Cited work 10.1155/2022/7814952 · retracted 2023-12-13 · notice 10.1155/2023/9829813 · this paper published 2026-07-09
Citations
Cites
1 retracted
An Analytical Study of Speech Pathology Detection Based on MFCC and Deep Neural Networks
Zakariah, M., Reshma, B., Alotaibi, Y. A., Guo, Y., Tran-Trung, K., & Elahi, M. M. (2022) · Computational and Mathematical Methods in Medicine · doi:10.1155/2022/7814952 · Cited by 34
RetractedGrad-CAM: Visual Explanations from Deep Networks via Gradient-Based Localization
Selvaraju, R. R., Cogswell, M., Das, A., Vedantam, R., Parikh, D., & Batra, D. (2017) · 2017 IEEE International Conference on Computer Vision (ICCV) · doi:10.1109/iccv.2017.74 · Cited by 24,513
Sarker, I. H. (2021) · SN Computer Science · doi:10.1007/s42979-021-00815-1 · Cited by 2,678 · cited 1×
“As a subset of AI, deep learning utilizes multi-layered neural networks to automatically learn from data and can perform complex tasks such as image recognition, speech recognition, and natural language processing (27).”
Ardila, D., Kiraly, A. P., Bharadwaj, S., Choi, B., Reicher, J. J., Peng, L., Tse, D., Etemadi, M., Ye, W., Corrado, G., et al. (2019) · Nature Medicine · doi:10.1038/s41591-019-0447-x · Cited by 2,167
Introduction to Machine Learning, Neural Networks, and Deep Learning.
Choi, R. Y., Coyner, A. S., Kalpathy–Cramer, J., Chiang, M. F., & Campbell, J. P. (2020) · PubMed · doi:10.1167/tvst.9.2.14 · Cited by 1,012 · cited 1×
“Compared with other machine learning algorithms, multi-layered neural networks exhibit superior performance in analyzing and learning from unstructured data, reduce the need for human supervision, and possess the flexibility to adapt to a variety of use cases (28).”
The impact of artificial intelligence in medicine on the future role of the physician
Ahuja, A. S. (2019) · PeerJ · doi:10.7717/peerj.7702 · Cited by 789 · cited 1×
“Artificial intelligence (AI) is revolutionizing the field of medicine by providing powerful tools for patient data analysis (26).”
Yamashita, R., Long, J., Longacre, T., Peng, L., Berry, G., Martin, B., Higgins, J., Rubin, D. L., & Shen, J. (2021) · The Lancet Oncology · doi:10.1016/s1470-2045(20)30535-0 · Cited by 391 · cited 1×
“Current artificial intelligence (AI) technologies have demonstrated remarkable capabilities in healthcare diagnostics, particularly in image-based recognition systems (17 – 20).”
Godino-Llorente, J. I., Gomez-Vilda, P., & Blanco-Velasco, M. (2006) · IEEE Transactions on Biomedical Engineering · doi:10.1109/tbme.2006.871883 · Cited by 322
Muscle misuse voice disorders: description and classification
Morrison, M. D., & Rammage, L. A. (1993) · Acta Oto-Laryngologica · doi:10.3109/00016489309135839 · Cited by 319 · cited 1×
“Despite their distinct pathophysiologies, these conditions frequently exhibit overlapping perceptual features and nearly indistinguishable laryngoscopic findings, particularly in cases with atypical presentations (6 – 11).”
Mikhael, P. G., Wohlwend, J., Yala, A., Karstens, L., Xiang, J., Takigami, A. K., Bourgouin, P. P., Chan, P., Mrah, S., Amayri, W., et al. (2023) · Journal of Clinical Oncology · doi:10.1200/jco.22.01345 · Cited by 288
Research works printed but not linked: 3
- Ludlow, 1995, Management of the spasmodic dysphonias, 436
- Roy, 2000, Toward a theory of the dispositional bases of functional dysphonia and vocal nodules: exploring the role of personality and emotional adjustment, 461
- Salhi, 2010, Voice disorders identification using multilayer neural network, Int Arab J Inf Technol, 2, 177
Metadata
- Authors
- Zhou, Z., Cheng, Y., Ren, Q., Li, Y., Yuanyue, X., Kang, J., Lu, C., & Ge, P.
- Affiliations
- Jinan University (CN)
- Key Laboratory of Guangdong Province (CN)
- Guangdong Academy of Medical Sciences (CN)
- Vanderbilt University Medical Center (US)
- Department of Otolaryngology Head Neck Surgery, Shenzhen People's Hospital (The Second Clinical Medical College, Jinan University, The First Affiliated Hospital of Southern University of Science and Technology)
- Corresponding author
- Kang, Jing
- Lu, Cheng
- Ge, Pingjiang
- Venue
- Frontiers in Digital Health
- Publisher
- Frontiers Media SA
- Published
- 2026-07-09
- Volume
- 8
- Pages
- 1690534-1690534
- ISSN
- 2673-253X (ISSN-L)
- PMC
- PMC13391920
- OpenAlex
- W7167808085
- Semantic Scholar
- 4b62d99372ae3ed1e64c94b573486459a5a37186
- Cited by
- 1
- References
- 44
- Citations by year
- 2026: 1
- FWCI
- 4.51
- Type
- journal article
- Language
- en
- Open access
- gold
- doi.org
- Licence
- Locations
- Frontiers in Digital Health (publishedVersion, cc-by)
- PubMed (publishedVersion)
- PubMed Central (publishedVersion, cc-by)
- DOAJ (DOAJ: Directory of Open Access Journals) (submittedVersion)
- Funders
- Basic and Applied Basic Research Foundation of Guangdong Province
- Topics
- Health Sciences › Health Professions › Speech and Hearing › Voice and Speech Disorders
- Health Sciences › Health Professions › Health Information Management › Artificial Intelligence in Healthcare
- Physical Sciences › Computer Science › Artificial Intelligence › Speech Recognition and Synthesis
- Keywords
- spasmodic dysphonia, muscle tension dysphonia, voice disorders, differential diagnosis, acoustic voice analysis, transfer learning, convolutional neural networks, log-mel spectrogram, pathological voice classification
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.
The claim verdicts this report counts are experimental: their accuracy is not yet measured.
A language model drew the claims from the paper on 2026-10-01, under version 3 of the claim definition (model deepseek-v4.1-flash, prompt extract-v3).
A second model labelled each claim on 2026-10-01 (model jev-1.13-20260917, definition revision gold-guideline-v0.2).
Cited sources are checked with a language model’s help (gpt-6-sol).
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: A deep learning approach for acoustic-based identification of muscle tension dysphonia and spasmodic dysphonia. https://app.cochat.ai/papers/10.3389/fdgth.2026.1690534
- The paper
- Zhou, Z., Cheng, Y., Ren, Q., Li, Y., Yuanyue, X., Kang, J., Lu, C., & Ge, P. (2026). A deep learning approach for acoustic-based identification of muscle tension dysphonia and spasmodic dysphonia. Frontiers in Digital Health. https://doi.org/10.3389/fdgth.2026.1690534
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.