X-Vectors: New Quantitative Biomarkers for Early Parkinson's Disease Detection From Speech
Many articles have used voice analysis to detect Parkinson's disease (PD), but few have focused on the early stages of the disease and the gender effect. In this article, we have adapted the latest speaker recognition system, called x-vectors, in order to detect PD at an early stage using voice...
Main Authors: | , , , , , , , |
---|---|
Format: | Article |
Language: | English |
Published: |
Frontiers Media S.A.
2021-02-01
|
Series: | Frontiers in Neuroinformatics |
Subjects: | |
Online Access: | https://www.frontiersin.org/articles/10.3389/fninf.2021.578369/full |
id |
doaj-2824f6eb36184efa818b863e2d34132c |
---|---|
record_format |
Article |
spelling |
doaj-2824f6eb36184efa818b863e2d34132c2021-02-19T06:20:09ZengFrontiers Media S.A.Frontiers in Neuroinformatics1662-51962021-02-011510.3389/fninf.2021.578369578369X-Vectors: New Quantitative Biomarkers for Early Parkinson's Disease Detection From SpeechLaetitia Jeancolas0Laetitia Jeancolas1Dijana Petrovska-Delacrétaz2Graziella Mangone3Graziella Mangone4Badr-Eddine Benkelfat5Jean-Christophe Corvol6Jean-Christophe Corvol7Marie Vidailhet8Marie Vidailhet9Stéphane Lehéricy10Stéphane Lehéricy11Stéphane Lehéricy12Habib Benali13Paris Brain Institute—ICM, Centre de NeuroImagerie de Recherche—CENIR, Paris, FranceLaboratoire SAMOVAR, Télécom SudParis, Institut Polytechnique de Paris, Palaiseau, FranceLaboratoire SAMOVAR, Télécom SudParis, Institut Polytechnique de Paris, Palaiseau, FranceSorbonne University, Inserm, CNRS, Paris Brain Institute—ICM, Paris, FranceAssistance Publique Hôpitaux de Paris, Hôpital Pitié-Salpêtrière, Department of Neurology, Clinical Investigation Center for Neurosciences, Paris, FranceLaboratoire SAMOVAR, Télécom SudParis, Institut Polytechnique de Paris, Palaiseau, FranceSorbonne University, Inserm, CNRS, Paris Brain Institute—ICM, Paris, FranceAssistance Publique Hôpitaux de Paris, Hôpital Pitié-Salpêtrière, Department of Neurology, Clinical Investigation Center for Neurosciences, Paris, FranceSorbonne University, Inserm, CNRS, Paris Brain Institute—ICM, Paris, FranceAssistance Publique Hôpitaux de Paris, Hôpital Pitié-Salpêtrière, Department of Neurology, Clinical Investigation Center for Neurosciences, Paris, FranceParis Brain Institute—ICM, Centre de NeuroImagerie de Recherche—CENIR, Paris, FranceSorbonne University, Inserm, CNRS, Paris Brain Institute—ICM, Paris, FranceAssistance Publique Hôpitaux de Paris, Hôpital Pitié-Salpêtrière, Department of Neuroradiology, Paris, FranceDepartment of Electrical & Computer Engineering, PERFORM Center, Concordia University, Montreal, QC, CanadaMany articles have used voice analysis to detect Parkinson's disease (PD), but few have focused on the early stages of the disease and the gender effect. In this article, we have adapted the latest speaker recognition system, called x-vectors, in order to detect PD at an early stage using voice analysis. X-vectors are embeddings extracted from Deep Neural Networks (DNNs), which provide robust speaker representations and improve speaker recognition when large amounts of training data are used. Our goal was to assess whether, in the context of early PD detection, this technique would outperform the more standard classifier MFCC-GMM (Mel-Frequency Cepstral Coefficients—Gaussian Mixture Model) and, if so, under which conditions. We recorded 221 French speakers (recently diagnosed PD subjects and healthy controls) with a high-quality microphone and via the telephone network. Men and women were analyzed separately in order to have more precise models and to assess a possible gender effect. Several experimental and methodological aspects were tested in order to analyze their impacts on classification performance. We assessed the impact of the audio segment durations, data augmentation, type of dataset used for the neural network training, kind of speech tasks, and back-end analyses. X-vectors technique provided better classification performances than MFCC-GMM for the text-independent tasks, and seemed to be particularly suited for the early detection of PD in women (7–15% improvement). This result was observed for both recording types (high-quality microphone and telephone).https://www.frontiersin.org/articles/10.3389/fninf.2021.578369/fullParkinson's diseasex-vectorsvoice analysisearly detectionautomatic detectiontelediagnosis |
collection |
DOAJ |
language |
English |
format |
Article |
sources |
DOAJ |
author |
Laetitia Jeancolas Laetitia Jeancolas Dijana Petrovska-Delacrétaz Graziella Mangone Graziella Mangone Badr-Eddine Benkelfat Jean-Christophe Corvol Jean-Christophe Corvol Marie Vidailhet Marie Vidailhet Stéphane Lehéricy Stéphane Lehéricy Stéphane Lehéricy Habib Benali |
spellingShingle |
Laetitia Jeancolas Laetitia Jeancolas Dijana Petrovska-Delacrétaz Graziella Mangone Graziella Mangone Badr-Eddine Benkelfat Jean-Christophe Corvol Jean-Christophe Corvol Marie Vidailhet Marie Vidailhet Stéphane Lehéricy Stéphane Lehéricy Stéphane Lehéricy Habib Benali X-Vectors: New Quantitative Biomarkers for Early Parkinson's Disease Detection From Speech Frontiers in Neuroinformatics Parkinson's disease x-vectors voice analysis early detection automatic detection telediagnosis |
author_facet |
Laetitia Jeancolas Laetitia Jeancolas Dijana Petrovska-Delacrétaz Graziella Mangone Graziella Mangone Badr-Eddine Benkelfat Jean-Christophe Corvol Jean-Christophe Corvol Marie Vidailhet Marie Vidailhet Stéphane Lehéricy Stéphane Lehéricy Stéphane Lehéricy Habib Benali |
author_sort |
Laetitia Jeancolas |
title |
X-Vectors: New Quantitative Biomarkers for Early Parkinson's Disease Detection From Speech |
title_short |
X-Vectors: New Quantitative Biomarkers for Early Parkinson's Disease Detection From Speech |
title_full |
X-Vectors: New Quantitative Biomarkers for Early Parkinson's Disease Detection From Speech |
title_fullStr |
X-Vectors: New Quantitative Biomarkers for Early Parkinson's Disease Detection From Speech |
title_full_unstemmed |
X-Vectors: New Quantitative Biomarkers for Early Parkinson's Disease Detection From Speech |
title_sort |
x-vectors: new quantitative biomarkers for early parkinson's disease detection from speech |
publisher |
Frontiers Media S.A. |
series |
Frontiers in Neuroinformatics |
issn |
1662-5196 |
publishDate |
2021-02-01 |
description |
Many articles have used voice analysis to detect Parkinson's disease (PD), but few have focused on the early stages of the disease and the gender effect. In this article, we have adapted the latest speaker recognition system, called x-vectors, in order to detect PD at an early stage using voice analysis. X-vectors are embeddings extracted from Deep Neural Networks (DNNs), which provide robust speaker representations and improve speaker recognition when large amounts of training data are used. Our goal was to assess whether, in the context of early PD detection, this technique would outperform the more standard classifier MFCC-GMM (Mel-Frequency Cepstral Coefficients—Gaussian Mixture Model) and, if so, under which conditions. We recorded 221 French speakers (recently diagnosed PD subjects and healthy controls) with a high-quality microphone and via the telephone network. Men and women were analyzed separately in order to have more precise models and to assess a possible gender effect. Several experimental and methodological aspects were tested in order to analyze their impacts on classification performance. We assessed the impact of the audio segment durations, data augmentation, type of dataset used for the neural network training, kind of speech tasks, and back-end analyses. X-vectors technique provided better classification performances than MFCC-GMM for the text-independent tasks, and seemed to be particularly suited for the early detection of PD in women (7–15% improvement). This result was observed for both recording types (high-quality microphone and telephone). |
topic |
Parkinson's disease x-vectors voice analysis early detection automatic detection telediagnosis |
url |
https://www.frontiersin.org/articles/10.3389/fninf.2021.578369/full |
work_keys_str_mv |
AT laetitiajeancolas xvectorsnewquantitativebiomarkersforearlyparkinsonsdiseasedetectionfromspeech AT laetitiajeancolas xvectorsnewquantitativebiomarkersforearlyparkinsonsdiseasedetectionfromspeech AT dijanapetrovskadelacretaz xvectorsnewquantitativebiomarkersforearlyparkinsonsdiseasedetectionfromspeech AT graziellamangone xvectorsnewquantitativebiomarkersforearlyparkinsonsdiseasedetectionfromspeech AT graziellamangone xvectorsnewquantitativebiomarkersforearlyparkinsonsdiseasedetectionfromspeech AT badreddinebenkelfat xvectorsnewquantitativebiomarkersforearlyparkinsonsdiseasedetectionfromspeech AT jeanchristophecorvol xvectorsnewquantitativebiomarkersforearlyparkinsonsdiseasedetectionfromspeech AT jeanchristophecorvol xvectorsnewquantitativebiomarkersforearlyparkinsonsdiseasedetectionfromspeech AT marievidailhet xvectorsnewquantitativebiomarkersforearlyparkinsonsdiseasedetectionfromspeech AT marievidailhet xvectorsnewquantitativebiomarkersforearlyparkinsonsdiseasedetectionfromspeech AT stephanelehericy xvectorsnewquantitativebiomarkersforearlyparkinsonsdiseasedetectionfromspeech AT stephanelehericy xvectorsnewquantitativebiomarkersforearlyparkinsonsdiseasedetectionfromspeech AT stephanelehericy xvectorsnewquantitativebiomarkersforearlyparkinsonsdiseasedetectionfromspeech AT habibbenali xvectorsnewquantitativebiomarkersforearlyparkinsonsdiseasedetectionfromspeech |
_version_ |
1724261597025665024 |