CoChat
Sign Up
Cited by 326Voice and Speech DisordersParkinson's Disease Mechanisms and TreatmentsPhonetics and Phonology Research

The DOI matches its Crossref record; 37 of the 37 references it cites were found in a registry or an index.

Automated analysis of connected speech reveals early biomarkers of Parkinson’s disease in patients with rapid eye movement sleep behaviour disorder

Open access
Hlavnička, Jan, Čmejla, Roman, Tykalová, Tereza, Šonka, Karel, Růžička, Evžen, Rusz, Jan·2017·Scientific Reports
DOI 10.1038/s41598-017-00047-5PMC PMC5428345
Open this paper in CoChat — free accountOpen PDF

Abstract

For generations, the evaluation of speech abnormalities in neurodegenerative disorders such as Parkinson’s disease (PD) has been limited to perceptual tests or user-controlled laboratory analysis based upon rather small samples of human vocalizations. Our study introduces a fully automated method that yields significant features related to respiratory deficits, dysphonia, imprecise articulation and dysrhythmia from acoustic microphone data of natural connected speech for predicting early and distinctive patterns of neurodegeneration. We compared speech recordings of 50 subjects with rapid eye movement sleep behaviour disorder (RBD), 30 newly diagnosed, untreated PD patients and 50 healthy controls, and showed that subliminal parkinsonian speech deficits can be reliably captured even in RBD patients, which are at high risk of developing PD or other synucleinopathies. Thus, automated vocal analysis should soon be able to contribute to screening and diagnostic procedures for prodromal parkinsonian neurodegeneration in natural environments.

Overview

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

Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.

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

Integrity signals

Registry

Crossref · 10.1038/s41598-017-00047-5

Cites

37 of 37 references verified0 citation errors · 0 not checkable

Statistics

Figures

Figure 1 Process diagram illustrating the principle of the automated segmentation algorithm on a speech signal sample. (A) Original speech sample depicted using oscillographic sound pressure plot. (B) Red ‘x’ marks illustrate recognized voiced speech depicted in dimensions of PWR and ACR in two consecutive recognition windows. (C) Speech sample representing voiceless intervals shorter than 300 ms after removal of voiced intervals from original speech sample. (D) Green ‘o’ marks illustrate recognized unvoiced speech depicted in dimensions of the first linear-frequency cepstral coefficient (LFCC1) and the second linear-frequency cepstral coefficient (LFCC2) in two consecutive recognition windows. (E) Speech sample representing respiration intervals longer than 200 ms after removal of voiced and unvoiced intervals from original speech sample. (F) Blue ‘o’ marks illustrate recognized respirations depicted in dimensions of the first linear-frequency cepstral coefficient (LFCC1) and the third linear-frequency cepstral coefficient (LFCC3) in two consecutive recognition windows. (G) Resulting intervals of the segmentation plotted in time. PWR = power, ACR = variance of autocorrelation function.
Figure 1 Process diagram illustrating the principle of the automated segmentation algorithm on a speech signal sample. (A) Original speech sample depicted using oscillographic sound pressure plot. (B) Red ‘x’ marks illustrate recognized voiced speech depicted in dimensions of PWR and ACR in two consecutive recognition windows. (C) Speech sample representing voiceless intervals shorter than 300 ms after removal of voiced intervals from original speech sample. (D) Green ‘o’ marks illustrate recognized unvoiced speech depicted in dimensions of the first linear-frequency cepstral coefficient (LFCC1) and the second linear-frequency cepstral coefficient (LFCC2) in two consecutive recognition windows. (E) Speech sample representing respiration intervals longer than 200 ms after removal of voiced and unvoiced intervals from original speech sample. (F) Blue ‘o’ marks illustrate recognized respirations depicted in dimensions of the first linear-frequency cepstral coefficient (LFCC1) and the third linear-frequency cepstral coefficient (LFCC3) in two consecutive recognition windows. (G) Resulting intervals of the segmentation plotted in time. PWR = power, ACR = variance of autocorrelation function.CC BY · PubMed Central
Figure 2 Mind map illustrating basic principles of individual acoustic speech features. MFCC = Mel-frequency cepstral coefficients.
Figure 2 Mind map illustrating basic principles of individual acoustic speech features. MFCC = Mel-frequency cepstral coefficients.CC BY · PubMed Central
Figure 3 Results of acoustic speech analyses. Bars represent mean values and error bars represent SD values. Repeated measures analysis of variance (RM-ANOVA) was used to test for group differences: GROUP (PD vs. RBD vs. controls): corrected *p < 0.05, **p < 0.01, ***p < 0.001 after Bonferroni adjustment; TASK (reading passage vs. monologue): corrected # p < 0.05, ## p < 0.01, ### p < 0.001 after Bonferroni adjustment. None of the features showed significant GROUP × TASK interaction. RST = rate of speech timing, AST = acceleration of speech timing, DPI = duration of pause intervals, EST = entropy of speech timing, DUS = duration of unvoiced stops, DUF = decay of unvoiced fricatives, DVI = duration of voiced intervals, GVI = gaping in-between voiced intervals, RSR = rate of speech respiration, PIR = pause intervals per respiration, RLR = relative loudness of respiration, LRE = latency of respiratory exchange, PD = Parkinson’s disease, RBD = rapid eye movement sleep behaviour disorder.
Figure 3 Results of acoustic speech analyses. Bars represent mean values and error bars represent SD values. Repeated measures analysis of variance (RM-ANOVA) was used to test for group differences: GROUP (PD vs. RBD vs. controls): corrected *p < 0.05, **p < 0.01, ***p < 0.001 after Bonferroni adjustment; TASK (reading passage vs. monologue): corrected # p < 0.05, ## p < 0.01, ### p < 0.001 after Bonferroni adjustment. None of the features showed significant GROUP × TASK interaction. RST = rate of speech timing, AST = acceleration of speech timing, DPI = duration of pause intervals, EST = entropy of speech timing, DUS = duration of unvoiced stops, DUF = decay of unvoiced fricatives, DVI = duration of voiced intervals, GVI = gaping in-between voiced intervals, RSR = rate of speech respiration, PIR = pause intervals per respiration, RLR = relative loudness of respiration, LRE = latency of respiratory exchange, PD = Parkinson’s disease, RBD = rapid eye movement sleep behaviour disorder.CC BY · PubMed Central
Figure 4 Flowchart describing the procedure of speech test experiment. (A) The motor score of UPDRS was examined in each RBD subject by a well-trained neurologist. (B) RBD subjects were separated into a subgroup of motor negatives (asymptomatic, UPDRS III* ≤ 3) and a subgroup of motor positives (symptomatic, UPDRS III* > 3). (C,D) Speech performances of all newly diagnosed, untreated PD patients, RBD subjects, and controls were analyzed using the set of designed speech features. (E) The most distinctive parkinsonian speech patterns were determined as the best resulting combination of speech features for differentiating between newly diagnosed, untreated PD patients and controls. (F) All RBD subjects were separated into a subgroup of speech positives (subjects with unchanged speech performance) and a subgroup of speech negatives (subjects with speech performance closer to PD speakers) based on the most distinctive parkinsonian speech pattern obtained through comparison between newly diagnosed, untreated PD and controls. PD = Parkinson’s disease, RBD = rapid eye movement sleep behaviour disorder, UPDRS = Unified Parkinson’s Disease Rating Scale; UPDRS III* = motor part of the UPDRS III score after removal of action tremor.
Figure 4 Flowchart describing the procedure of speech test experiment. (A) The motor score of UPDRS was examined in each RBD subject by a well-trained neurologist. (B) RBD subjects were separated into a subgroup of motor negatives (asymptomatic, UPDRS III* ≤ 3) and a subgroup of motor positives (symptomatic, UPDRS III* > 3). (C,D) Speech performances of all newly diagnosed, untreated PD patients, RBD subjects, and controls were analyzed using the set of designed speech features. (E) The most distinctive parkinsonian speech patterns were determined as the best resulting combination of speech features for differentiating between newly diagnosed, untreated PD patients and controls. (F) All RBD subjects were separated into a subgroup of speech positives (subjects with unchanged speech performance) and a subgroup of speech negatives (subjects with speech performance closer to PD speakers) based on the most distinctive parkinsonian speech pattern obtained through comparison between newly diagnosed, untreated PD and controls. PD = Parkinson’s disease, RBD = rapid eye movement sleep behaviour disorder, UPDRS = Unified Parkinson’s Disease Rating Scale; UPDRS III* = motor part of the UPDRS III score after removal of action tremor.CC BY · PubMed Central
Figure 5 Selected pairs of representative acoustic features depicting the most distinctive speech pattern allowing differentiation between PD and controls. Individual speakers plotted in speech dimensions represented by (A) DPI in the monologue and RST in reading task, and (B) DPI in the reading task and DUS in monologue. The solid gray line represents the border of discrimination between speech negatives (subjects with unchanged speech performance) and speech positives (subjects with speech performance closer to PD speakers). The ‘.’ marks represent PD, ‘o’ for RBD and ‘x’ for controls. PD = Parkinson’s disease, RBD = rapid eye movement sleep behaviour disorder, DPI = duration of pause intervals, RST = rate of speech intervals, DUS = duration of unvoiced stops.
Figure 5 Selected pairs of representative acoustic features depicting the most distinctive speech pattern allowing differentiation between PD and controls. Individual speakers plotted in speech dimensions represented by (A) DPI in the monologue and RST in reading task, and (B) DPI in the reading task and DUS in monologue. The solid gray line represents the border of discrimination between speech negatives (subjects with unchanged speech performance) and speech positives (subjects with speech performance closer to PD speakers). The ‘.’ marks represent PD, ‘o’ for RBD and ‘x’ for controls. PD = Parkinson’s disease, RBD = rapid eye movement sleep behaviour disorder, DPI = duration of pause intervals, RST = rate of speech intervals, DUS = duration of unvoiced stops.CC BY · PubMed Central

Verification

Claims the paper makes

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

Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.

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

Cites

Show all 37 references

Research works printed but not linked: 3

  1. Duffy, J. R. Motor Speech Disorders: Substrates, Differential Diagnosis and Management, 3rd ed., Mosby, St. Louis (2013).
  2. Weismer, G. & McNeil, M. Articulatory characteristics of Parkinsonian dysarthria: Segmental and phrase-level timing, spirantization, and glottal-supraglottal coordination. The dysarthrias: Physiology, acoustics, perception, management, College-Hill Press, San Diego 101–130 (1984).
  3. American Academy of Sleep Medicine. International Classification of Sleep Disorders, Third Edition: Diagnostic and Coding Manual, Westchester, Illinois: American Academy of Sleep Medicine (2014).

Cited by

Sorted by citations

See all 326

Related

Metadata

Authors
Hlavnička, J., Čmejla, R., Tykalová, T., Šonka, K., Růžička, E., & Rusz, J.
Affiliations
  • Czech Technical University in Prague (CZ)
  • Charles University (CZ)
Corresponding author
Rusz, Jan
ORCID
Venue
Scientific Reports
Publisher
Springer Science and Business Media LLC
Published
2017-02-02
Volume
7
Issue
1
Pages
12-12
ISSN
2045-2322 (ISSN-L)
DOI
10.1038/s41598-017-00047-5
PMC
PMC5428345
OpenAlex
W2580210409
Cited by
326
References
37
Citations by year
2026: 11 · 2025: 25 · 2024: 25 · 2023: 30 · 2022: 28 · 2021: 71 · 2020: 44 · 2019: 46 · 2018: 33 · 2017: 11
FWCI
27.12
Type
journal article
Language
en
Open access
Licence
Locations
  • Scientific Reports (publishedVersion, cc-by)
  • PubMed Central (publishedVersion, cc-by)
  • PubMed (publishedVersion)
  • DOAJ (DOAJ: Directory of Open Access Journals) (submittedVersion, cc-by-sa)
  • Cvut DSpace (Czech Technical University) (submittedVersion, cc-by)
Funders
  • Grantová Agentura České Republiky (GA16-07879S)
  • Ministerstvo Zdravotnictví Ceské Republiky (16-28914A, 15 – 28038A)
Topics
  • Health Sciences › Health Professions › Speech and Hearing › Voice and Speech Disorders
  • Health Sciences › Medicine › Clinical Neurology › Parkinson's Disease Mechanisms and Treatments
  • Social Sciences › Psychology › Experimental and Cognitive Psychology › Phonetics and Phonology Research
Keywords
Parkinson's disease, rapid eye movement sleep behavior disorder, automated speech analysis, acoustic biomarkers, early diagnosis, prodromal neurodegeneration, dysphonia, cardiac arrhythmias, synucleinopathies, speech screening
MeSH terms
Adult, Aged, Aged, 80 and over, Articulation Disorders, Female, Humans, Male, Middle Aged, Parkinson Disease, Pattern Recognition, Automated, Respiration, Biomarkers, REM Sleep Behavior Disorder, Dysphonia

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