On the Relevance of Clinical Assessment Tasks for the Automatic Detection of Parkinson's Disease Medication State from Speech

Alberto Abad; Anna Pompili; Carlos-D. Mart\'inez-Hinarejos; David Gimeno-G\'omez; Isabel Guimar\~aes; Joaquim J. Ferreira; Rita Cardoso; Rub\'en Solera-Ure\~na

REVIEW 3 major objections 4 minor 41 references

On the Relevance of Clinical Assessment Tasks for the Automatic Detection of Parkinson's Disease Medication State from Speech

T0 review · 3 major / 4 minor · reviewed 2026-08-07 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗ Review JSON

Pith's one-line read A single, speaker-independent speech model can tell whether a Parkinson's patient is ON or OFF medication, hitting 88.2% F1.

desk verdict Useful benchmark with a real novelty claim, but the 88.2% F1 is not well supported as evidence of medication-state detection until the label-validity and session-confound problems are addressed. read the letter →

arxiv 2505.15378 v2 pith:KSSAESNO submitted 2025-05-21 eess.AS cs.SD

David Gimeno-Gómez , Rubén Solera-Ureña , Anna Pompili , Carlos-D. Martínez-Hinarejos , Rita Cardoso , Isabel Guimarães , Joaquim J. Ferreira , Alberto Abad This is my paper · ORCID

classification eess.AScs.SD

keywords Parkinson'sdiseasemedicationstatedetectionspeechbiomarkersself-supervisedrepresentationsWav2Vec2prosodyspeaker-independentclassificationFraLusoParkcorpus

verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

The paper presents the first speaker-independent system that identifies whether a Parkinson's disease patient is in the ON or OFF medication state from voice recordings alone. Using the FraLusoPark corpus of 74 patients recorded both off and on medication, the authors compare hand-crafted acoustic descriptors (eGeMAPS) with self-supervised Wav2Vec2 embeddings and two classifiers, a linear SVM and an attention-based deep network. The best configuration — Wav2Vec2 features with a linear SVM on a sentence-reading task with prosodic contrasts — reaches an F1-score of 88.2%. The results indicate that continuous, prosody-rich speech tasks carry the most medication-state information, and that a single generic model can replace per-patient personalization.

What carries the argument

The central objects are two speech representations: eGeMAPS, a compact set of 88 knowledge-based acoustic functionals extracted with openSMILE, and Wav2Vec2.0, a self-supervised model whose 1024-dimensional embeddings are taken from the 7th encoder layer of the multilingual XLS-R variant and aggregated at both frame and utterance levels. These feed either a linear SVM with PCA dimensionality reduction or an attention-based deep network, under three data grouping strategies (task-specific, task-grouping, task-independent) evaluated with nested five-fold cross-validation. The decisive mechanism is the interaction between representation and task: Wav2Vec2 embeddings trained on large multilingual speech corpora capture prosodic variability that eGeMAPS descriptors miss, and that prosodic information is most available in continuous, prosodically demanding speech tasks.

What would settle it

Record the same patients twice in the same medication state under an identical protocol and retrain the best model to distinguish the two sessions; if the F1 stays near 88%, the model is separating sessions rather than medication states, and the central claim would be refuted. A cheaper check is label permutation within each patient (shuffling ON/OFF while keeping speaker partitions), which should destroy performance if the signal is medication-related.

Watch

Extended reading notes

Core claim

The central claim is that the ON/OFF medication state of Parkinson's patients can be identified automatically from speech in a speaker-independent manner, without fine-tuning per patient. The discovery that carries this claim is that self-supervised speech representations (Wav2Vec2.0 embeddings) substantially outperform knowledge-based acoustic descriptors (eGeMAPS) in this setting, and that the choice of speech assessment task is decisive: tasks involving continuous, prosody-rich speech — particularly reading sentences with prosodic properties (PROS-SENT) — yield the best accuracy, with the top result of 88.2% F1 achieved by a linear SVM trained on Wav2Vec2 features. Traditional machine learning classifiers are shown to be competitive with attention-based deep networks, which lowers the computational bar for deployment. The authors also report that gender-specific models do not improve over a gender-independent one, suggesting the underlying vocal cues are shared across genders.

Load-bearing premise

The system assumes the clinical ON/OFF labels are reliable ground truth, even though only 25 of 74 patients showed a change in MDS-UPDRS-III after medication and two worsened; if the labels do not correspond to a stable acoustic signature, the classifier may be learning recording-session or time-of-day artifacts instead.

Editorial extensions

If this is right

A single deployable model can classify a new patient's medication state without any per-patient training data.
Clinical protocols should favor prosodically rich, continuous speech tasks (such as reading prosodic sentences) when speech is used to monitor medication response.
Self-supervised speech features can serve as effective medication-state biomarkers even without fine-tuning on pathological speech.
Linear SVMs offer near-state-of-the-art performance at much lower computational cost, making remote monitoring more feasible.
The reported 88.2% F1 is specific to this corpus; cross-linguistic and cross-dataset validation is the immediate next test.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

If the learned signal is genuinely about medication state rather than recording conditions, the approach could enable cheap, passive monitoring of treatment response from brief voice captures in home environments.
The observation that only one-third of patients had changed MDS-UPDRS-III motor scores while the classifier still separated ON from OFF suggests speech may reflect a medication effect not well captured by the clinical motor scale — or that the classifier exploits session-level acoustic confounds; a controlled recording protocol with repeated same-state sessions could distinguish these.
The layer-7 embedding choice hints that intermediate self-supervised layers encode prosodic structure; targeted probing against prosodic annotations could reveal which acoustic dimensions carry the ON/OFF signal and make the black-box model more interpretable.
Combining speaker-independent pre-training with light speaker-dependent adaptation (e.g., a few labeled samples from the target patient) could close the remaining gap toward the 95% accuracy reported in the speaker-dependent literature.

Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Desk Editor's Note

read the letter

Short version: this is a competent, clearly-written benchmark that makes a modest but real claim to novelty — first speaker-independent PD medication-state detection from speech — and the systematic comparison across nine clinical tasks is useful. But I would not take the 88.2% F1 at face value as evidence of medication detection; the label reliability and session confounds are serious enough to require re-analysis or strong argument before the headline is accepted.

What's good: the evaluation is carefully set up — speaker-independent nested CV, gender-balanced folds, repeated runs, three sensible data-grouping strategies. The finding that continuous, prosody-rich tasks (PROS-SENT, TEXT, FROG) consistently outperform vowels and didadochokinesia is interesting and aligns with prior work. Showing that SSL embeddings beat eGeMAPS without fine-tuning is a useful data point for the community. The comparison with Pompili's speaker-dependent work is honest.

The soft spots: (1) The ON/OFF labels are clinical protocol labels, not verified speech changes. The paper itself states only 25/74 patients had MDS-UPDRS-III changes after medication, and two worsened. If most patients' motor state doesn't change, what is the model separating? The simplest non-medication explanation is recording-session differences: OFF is recorded after 12h withdrawal, ON after medication, almost certainly in a fixed order. That confounds time of day, fatigue, vocal warm-up, and session acoustics. The paper does not control for this or even discuss it. (2) The 'Dysarthria Level Analysis' section says there was no correlation with severity changes even for subjects whose scores changed. That is an internal tension with the medication-state interpretation, not just a null result. (3) The headline result is selected from hundreds of feature/task/architecture/grouping combinations with no multiple-comparison correction or significance tests. The trend across continuous tasks softens this, but the 88.2% itself is not well supported. (4) No code or data release, which limits reproducibility.

These are not fatal to the paper's value as a benchmark, but they are fatal to the strong claim that the method detects medication state. The authors need to at least address the session confound (e.g., by analyzing a subset with verified motor response, or by comparing feature distributions across sessions), and temper the language from 'detection' to 'discrimination under the recorded protocol.' As is, I'd send it to peer review with the expectation of major revision.

Who it's for: people building speech-based PD monitoring systems, and reviewers who care about benchmark hygiene. I'd cite the task comparison but not the headline number.

Referee Report

3 major / 4 minor

Summary. The paper proposes a speaker-independent system for classifying Parkinson's disease patients' medication state (ON/OFF) from speech. It compares knowledge-based eGeMAPS features against self-supervised Wav2Vec2/XLS-R embeddings, and linear SVMs against attention-based deep networks, across nine clinical speech tasks and three training-data grouping strategies on the FraLusoPark corpus. The reported best result is 88.2% F1 for the PROS-SENT task with Wav2Vec2 features and a linear SVM under the task-specific strategy. The authors conclude that continuous, prosody-rich speech tasks and self-supervised representations are the most effective, and claim this is the first speaker-independent approach to PD medication-state identification from speech.

Significance. If the reported result is attributable to medication state rather than to recording-session artifacts, the paper provides the first speaker-independent PD medication-state classifier from speech and a useful practical step toward remote monitoring. The experimental design is generally sound: speaker-independent nested cross-validation with repeated seeds, gender-balanced folds, multiple tasks and feature sets, and an explicit limitations section. The strong performance of SSL features on continuous speech tasks is plausible and partly supported by prior literature. However, two load-bearing issues—the confounding of ON/OFF labels with recording session and the absence of any multiple-comparison control when selecting the headline configuration—mean the central claim is not yet established at the level the paper states.

major comments (3)

[Section 2; Section 4, Dysarthria Level Analysis] The central attribution of the 88.2% F1 score to medication state is not adequately supported because the ON/OFF labels are systematically confounded with the recording protocol. Patients were recorded OFF after at least 12 hours of drug withdrawal and ON at least 1 hour after medication, i.e., in two separate sessions; the text does not describe counterbalancing of session order or a no-medication control condition. The classifier could therefore learn session, time-of-day, fatigue, or vocal-warm-up cues rather than a medication signature. This concern is reinforced by the paper's own observation that only 25 of 74 participants had MDS-UPDRS-III changes after medication (two worsened) and that the dysarthria-level analysis found no correlation with severity changes even for those subjects. The authors should provide a control analysis, for example by evaluating the best model on the subset of patients with documented motor changes, by training on repeated recordings under unchanged medication status, or by demonstrating that the learned decision function is driven by medication-related acoustic changes rather than session artifacts.
[Section 4, Table 2] The headline result is selected as the maximum over a large grid of configurations: 9 tasks × 3 grouping strategies × 2 feature types × 2 architectures, with the hyperparameters of the SVM also tuned in the inner loop. No multiple-comparison correction, hold-out model-selection procedure, or external validation is applied to this best cell, so the 88.2% F1 may partially reflect selection bias. The reported standard deviations quantify fold and seed variability within one cell but not the variability of the maximum over the whole grid. The authors should either report the distribution of results across all cells with an appropriate correction or permutation test, or validate the single best configuration on a set of speakers never used in any model selection.
[Section 4, Gender-based Analysis] The claim of 'robustness across genders' is made without any statistical comparison of the male, female, and gender-independent results in Table 3. For the SVM, the differences among 88.2%, 88.9%, and 87.6% are within the reported standard deviations. The authors should state whether these differences are statistically meaningful and, if not, soften the robustness claim accordingly. This is not the central claim of the paper, but it is part of the evidence for the generality of the method.

minor comments (4)

[Section 3.1] The text refers to 'Wav2Vec2.0' but the cited model is XLS-R (reference [19]); please align the terminology used in the text with the actual pretrained model.
[Section 4, Dysarthria Level Analysis] The dysarthria-level analysis is described only qualitatively; please provide the definition of the severity strata, the number of subjects in each stratum, and the actual performance numbers or statistics supporting the claimed absence of correlation.
[Limitations] The Limitations paragraph should explicitly acknowledge the session-confounding risk and the small number of patients with measured motor change; the current wording mentions data scarcity but not label validity, which is the more serious threat to interpretation.
[Limitations] The phrase 'the no incorporation of demographic information' is grammatically awkward; it should read 'the failure to incorporate demographic information'.

Circularity Check

0 steps flagged · score 1.0 of 10

No significant circularity: the medication-state classifier is an empirical benchmark with held-out speaker-independent evaluation; self-citations are contextual, not load-bearing.

full rationale

The paper's central claim is an empirical result, not a derivation. The target labels are clinical ON/OFF states, the features are pre-trained Wav2Vec2 embeddings or eGeMAPS descriptors, and the classifiers are trained and evaluated under nested cross-validation with speaker-independent splits. The reported 88.2% F1-score comes from held-out test folds, so it is not presupposed by the training procedure. The choice of the 7th Wav2Vec2 encoder layer is adopted from external prior work and is an architectural decision, not a parameter fitted to the target data; even if it were tuned, it would not guarantee the reported performance. The standardization using ON-state training samples is explicitly reported to have no significant effect, so it is not a loaded input. Self-citations to Pompili et al. [17] and Gimeno-Gomez et al. [31] are used for comparison and future-work context, not as load-bearing evidence for the main result. The weakness noted by skeptics—that ON/OFF labels may be confounded with recording session or time of day—is an external-validity concern about what the model learns, not a circularity in the derivation. No step in the paper reduces, by definition or by self-citation, to its own inputs.

Assumptions & free parameters 4 free parameters · 3 assumptions · 0 invented entities

The central evaluation rests on the clinical labels and the chosen feature pipelines; no new entities are introduced. The tuned hyperparameters are standard machine learning choices and are not used to fabricate a derivation. The main unexamined premises are the reliability of the ON/OFF labels and the transferability of the preprocessing pipeline.

free parameters (4)

PCA components for SVM = 16, 32, or 64 selected by inner cross-validation
Dimensionality of SVM input tuned per fold; a standard model selection hyperparameter.
SVM regularization C = 0.01 to 10 selected by inner cross-validation
Regularization strength tuned by inner loop; no theoretical justification given.
Wav2Vec2 encoder layer = 7th layer out of 24
Chosen based on prior literature [16,27], not justified from first principles for this task.
A-DNN training hyperparameters = learning rate 0.0004, 5 epochs, batch size 8
Set after a preliminary hyperparameter search, not nested within the outer cross-validation loop.

assumptions (3)

domain assumption ON/OFF clinical labels reflect a learnable speech state
The classifier is trained against these labels, yet only 25 of 74 patients showed MDS-UPDRS-III changes after medication, so the acoustic target may be absent for many subjects.
domain assumption The automatic preprocessing from [17] correctly removes silences and SLT instructions
No validation of this preprocessing step is reported on the FraLusoPark corpus; errors would propagate to all downstream features.
domain assumption Wav2Vec2 XLS-R embeddings encode medication-state-relevant prosodic cues
The model is used without fine-tuning, so the relevance of layer 7 features is assumed from prior pathological speech work.

how reviews work

0 comments

Cite this review

Pith. "Pith review of On the Relevance of Clinical Assessment Tasks for the Automatic Detection of Parkinson's Disease Medication State from Speech." pith.science (2026). https://pith.science/paper/KSSAESNO

@misc{pith2026250515378,
  author       = {Pith},
  title        = {Pith review of: On the Relevance of Clinical Assessment Tasks for the Automatic Detection of Parkinson's Disease Medication State from Speech},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/KSSAESNO}},
  note         = {Machine review of arXiv:2505.15378}
}

read the original abstract

The automatic identification of medication states of Parkinson's disease (PD) patients can assist clinicians in monitoring and scheduling personalized treatments, as well as studying the effects of medication in alleviating the motor symptoms that characterize the disease. This paper explores speech as a non-invasive and accessible biomarker for identifying PD medication states, introducing a novel approach that addresses this task from a speaker-independent perspective. While traditional machine learning models achieve competitive results, self-supervised speech representations prove essential for optimal performance, significantly surpassing knowledge-based acoustic descriptors. Experiments across diverse speech assessment tasks highlight the relevance of prosody and continuous speech in distinguishing medication states, reaching an F1-score of 88.2%. These findings may streamline clinicians' work and reduce patient effort in voice recordings.

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

41 extracted references · 40 canonical work pages

[1]

Introduction Parkinson’s disease (PD) is a progressive neurodegenerati ve disorder that primarily affects movement control and speec h production [1]. Patients are faced with articulatory deﬁci ts, leading to dysarthria, and voice disorders, frequently per ceived as dysprosody (resulting in voice modulation and intensity im- pairments) [2]. Within the cou...

work page
[2]

A unique feature of this corpus is the inclusion of a study speciﬁcally designed to analyze the effects of medica - tion in PD patients

Data Material The European Portuguese FraLusoPark corpus [20] is com- posed of 158 speakers, 74 healthy controls and 84 PD pa- tients. A unique feature of this corpus is the inclusion of a study speciﬁcally designed to analyze the effects of medica - tion in PD patients. To this end, participants diagnosed wit h PD were recorded twice: at least 12 hours a...

work page
[3]

Methodology This work is inspired by two well-established frameworks in the ﬁeld of automatic analysis of pathological speech: i) th e traditional approach, which relies on knowledge-based aco us- tic descriptors and conventional machine learning techniq ues, and ii) a more recent paradigm that leverage SSL speech rep- resentations and attention-based dee...

work page
[4]

Results & Discussion A comprehensive overview of our experimental results for au - tomatic, speaker-independent PD medication state identiﬁ cation from speech is presented in Table 2, showing the outcomes for the three data grouping strategies and different combinati ons of model architectures and speech features. Results are prese nted at the utterance l...

work page
[5]

Conclusions & Future Work In this paper, we presented our work on the automatic identiﬁ - cation of PD medication states from speech, showing its pote n- tial for more reliable remote monitoring and personalized t reat- ment scheduling. Compared to prior work [17], our speaker- independent approach offers two key advantages: i) the ability to leverage lar...

work page
[6]

Acknowledgements The work of Gimeno-G´ omez and Mart´ ınez-Hinarejos was par- tially supported by GV A through Grants CIACIF/2021/295 and CIBEFP/2023/167, by Grant PID2021-124361OB-C31 funded by MCIN/AEI/10.13039/501100011033 and ERDF/EU. The work of Solera-Ure˜ na, Pompili and Abad was supported by Portuguese national funds through Fundac ¸ ˜ ao para a C...

doi:10.13039/501100011033 2021
[7]

Parkinson’s Disease: Clinical Features a nd Diagno- sis,

J. Jankovic, “Parkinson’s Disease: Clinical Features a nd Diagno- sis,” Journal of Neurology, Neurosurgery & Psychiatry , vol. 79, no. 4, pp. 368–376, 2008

work page 2008
[8]

Intonation an d Speech Rate in Parkinson’s Disease: General and Dynamic Aspects and Responsiveness to Levodopa Admission,

S. Skodda, W. Gr¨ onheit, and U. Schlegel, “Intonation an d Speech Rate in Parkinson’s Disease: General and Dynamic Aspects and Responsiveness to Levodopa Admission,” Journal of V oice, vol. 25, no. 4, pp. e199–e205, 2011

work page 2011

Show all 41 references

[9]

Nonlinear progression of Parkinson disease as determined by serial positron emission tomographic imagin g of striatal ﬂuorodopa F 18 activity,

R. Hilker et al. , “Nonlinear progression of Parkinson disease as determined by serial positron emission tomographic imagin g of striatal ﬂuorodopa F 18 activity,” Archives of neurology, vol. 62, no. 3, pp. 378–382, 2005

2005
[10]

Speech Defects o f Parkinsonian Patients: Effects of Levodopa Therapy on Spee ch Intelligibility,

K. K. Nakano, H. Zubick, and H. R. Tyler, “Speech Defects o f Parkinsonian Patients: Effects of Levodopa Therapy on Spee ch Intelligibility,” Neurology, vol. 23, no. 8, pp. 865–865, 1973

1973
[11]

Speech Changes in Parkinson’s Disease During Treatment with L-dopa,

V . I. Wolfe et al., “Speech Changes in Parkinson’s Disease During Treatment with L-dopa,” Journal of Communication Disorders , vol. 8, no. 3, pp. 271–279, 1975

1975
[12]

Perceptual Characteristics of Parkinso- nian Speech: A Comparison of the Pharmacological Effects of Levodopa across Speech and Non-Speech Motor Systems,

E. Plowman-Prine et al., “Perceptual Characteristics of Parkinso- nian Speech: A Comparison of the Pharmacological Effects of Levodopa across Speech and Non-Speech Motor Systems,” Neu- roRehabilitation, vol. 24, no. 2, pp. 131–144, 2009

2009
[13]

Automatic detection of Parkinson’s disease in running speech spoken in three different languag es,

J. R. Orozco-Arroyave et al., “Automatic detection of Parkinson’s disease in running speech spoken in three different languag es,” The Journal of the ASA , vol. 139, no. 1, pp. 481–500, 2016

2016
[14]

Parkinson’s Disease and Aging: Analysis of Thei r Ef- fect in Phonation and Articulation of Speech,

T. Arias-V ergara, J. C. V´ asquez-Correa, and J. R. Orozc o- Arroyave, “Parkinson’s Disease and Aging: Analysis of Thei r Ef- fect in Phonation and Articulation of Speech,” Cognitive Compu- tation, vol. 9, no. 6, pp. 731–748, 2017

2017
[15]

Long-Term Average Spectral Feature-Based Parkinson’s Disease Detection from Speech,

A. Lahoti et al. , “Long-Term Average Spectral Feature-Based Parkinson’s Disease Detection from Speech,” in Adv. Machine In- telligence and Signal Processing. Springer, 2022, pp. 603–612

2022
[16]

Characterization Methods for the Detection of Multiple V oice Disorders: Neurological, Functional, and Laryngeal Diseases,

J. R. Orozco-Arroyave et al. , “Characterization Methods for the Detection of Multiple V oice Disorders: Neurological, Functional, and Laryngeal Diseases,” IEEE Journal of Biomedical and Health Informatics, vol. 19, no. 6, pp. 1820–1828, 2015

2015
[17]

Automatic Detection of Parkinson’s Disease: An Experimental Analysis of Common Speech Production Tasks Used for Diagnosis,

A. Pompili et al. , “Automatic Detection of Parkinson’s Disease: An Experimental Analysis of Common Speech Production Tasks Used for Diagnosis,” in International Conference on Text, Speech, and Dialogue. Springer, 2017, pp. 411–419

2017
[18]

Multi-Class Detection of Pathological Speech with Latent Features: How does it Perform on Unseen Data?

D. Wagner et al., “Multi-Class Detection of Pathological Speech with Latent Features: How does it Perform on Unseen Data?” in Proc. of Interspeech, 2023, pp. 2318–2322

2023
[19]

Exploiting Foundation Models and Speech Enhancement for Parkinson’s Disease Detection from Speech in Real-World Operative Conditions,

M. La Quatra et al., “Exploiting Foundation Models and Speech Enhancement for Parkinson’s Disease Detection from Speech in Real-World Operative Conditions,” in Proc. of Interspeech, 2024, pp. 1405–1409

2024
[20]

On the Use of a Foundation Acoustic Model to Identify Highly Relevant Phonetic Information of Parkinson’s Speech,

D. Escobar-Grisales, C. R´ ıos-Urrego, and J. Orozco-A rroyave, “On the Use of a Foundation Acoustic Model to Identify Highly Relevant Phonetic Information of Parkinson’s Speech,” in W ork- shop on Engineering Applications . Springer, 2024, pp. 71–81

2024
[21]

Automatic Classiﬁcation of Parkin- son’s Disease Using Wav2V ec Embeddings at Phoneme, Syllable, and Word Levels,

J. D. Gallo-Aristiz´ abal et al., “Automatic Classiﬁcation of Parkin- son’s Disease Using Wav2V ec Embeddings at Phoneme, Syllable, and Word Levels,” in Text, Speech, and Dialogue. Springer Na- ture Switzerland, 2024, pp. 313–323

2024
[22]

Interpretable speech features vs. DNN embed- dings: What to use in the automatic assessment of Parkinson’ s disease in multi-lingual scenarios,

A. Favaro et al., “Interpretable speech features vs. DNN embed- dings: What to use in the automatic assessment of Parkinson’ s disease in multi-lingual scenarios,” Computers in Biology and Medicine, vol. 166, p. 107559, 2023

2023
[23]

Assessment of Parkinson’s Disease Medica- tion State Through Automatic Speech Analysis,

A. Pompili et al. , “Assessment of Parkinson’s Disease Medica- tion State Through Automatic Speech Analysis,” in Proc. of In- terspeech, 2020, pp. 4591–4595

2020
[24]

The Geneva Minimalistic Acoustic Parameter Set (GeMAPS) for V oice Research and Affective Computing,

F. Eyben et al., “The Geneva Minimalistic Acoustic Parameter Set (GeMAPS) for V oice Research and Affective Computing,” IEEE Trans. on Affective Computing, vol. 7, no. 2, pp. 190–202, 2016

2016
[25]

XLS-R: Self-supervised Cross-lingual Speech Representation Learning at Scale,

A. Babu et al. , “XLS-R: Self-supervised Cross-lingual Speech Representation Learning at Scale,” in Proc. of Interspeech, 2022, pp. 2278–2282

2022
[26]

Dysarthria in Individuals with Parkinson’s Disease: A Protocol for a Binational, Cross-Sectional, Cas e- Controlled Study in French and European Portuguese (FraLu- soPark),

S. Pinto et al. , “Dysarthria in Individuals with Parkinson’s Disease: A Protocol for a Binational, Cross-Sectional, Cas e- Controlled Study in French and European Portuguese (FraLu- soPark),” British Medical Journal, vol. 6, no. 11, 2016

2016
[27]

(Dys)Prosody in Parkinson’s Disease: Effects of Medication and Disease Duration on Intonation and Prosod ic Phrasing,

S. Frota et al. , “(Dys)Prosody in Parkinson’s Disease: Effects of Medication and Disease Duration on Intonation and Prosod ic Phrasing,” Brain Sciences, vol. 11, no. 8, 2021

2021
[28]

Movement Disorder Society-Sponsored Re- vision of the Uniﬁed Parkinson’s Disease Rating Scale (MDS- UPDRS): Scale Presentation and Clinimetric Testing Result s,

C. G. Goetz et al. , “Movement Disorder Society-Sponsored Re- vision of the Uniﬁed Parkinson’s Disease Rating Scale (MDS- UPDRS): Scale Presentation and Clinimetric Testing Result s,” Movement Disorders, vol. 23, no. 15, pp. 2129–2170, 2008

2008
[29]

Frenchay Dysarthria Assessment (FDA-2) in Parkinson’s Disease: Cross-Cultural Adaptation and Psych ome- tric Properties of the European Portuguese V ersion,

R. Cardoso et al. , “Frenchay Dysarthria Assessment (FDA-2) in Parkinson’s Disease: Cross-Cultural Adaptation and Psych ome- tric Properties of the European Portuguese V ersion,” Journal of neurology, vol. 264, pp. 21–31, 2017

2017
[30]

Exploring Deep Spectrum Representations via Attention-Based Recurrent and Convolutional Neural Netwo rks for Speech Emotion Recognition,

Z. Zhao et al. , “Exploring Deep Spectrum Representations via Attention-Based Recurrent and Convolutional Neural Netwo rks for Speech Emotion Recognition,” IEEE Access , vol. 7, pp. 97 515–97 525, 2019

2019
[31]

Multi -Triplet Loss-Based Models for Categorical Depression Recognition from Speech,

A. Mallol-Ragolta, M. Milling, and B. Schuller, “Multi -Triplet Loss-Based Models for Categorical Depression Recognition from Speech,” in Proc. IberSPEECH, 2024, pp. 31–35

2024
[32]

Opensmile: the munich versatile and fast open-source audio feature extractor,

F. Eyben, M. W¨ ollmer, and B. Schuller, “Opensmile: the munich versatile and fast open-source audio feature extractor,” i n ACM International Conference on Multimedia, 2010, p. 1459–1462

2010
[33]

Automatic Parkinson’s disease detection from speech: Layer selection vs adaptation of foundation models ,

T. Purohit et al. , “Automatic Parkinson’s disease detection from speech: Layer selection vs adaptation of foundation models ,” in ICASSP, 2025

2025
[34]

Layer-Wise Analy sis of a Self-Supervised Speech Representation Model,

A. Pasad, J.-C. Chou, and K. Livescu, “Layer-Wise Analy sis of a Self-Supervised Speech Representation Model,” in Proc. of IEEE ASRU, 2021, pp. 914–921

2021
[35]

Exploring Digital Speech Biomarkers of Hy- pokinetic Dysarthria in a Multilingual Cohort,

D. Kovac et al. , “Exploring Digital Speech Biomarkers of Hy- pokinetic Dysarthria in a Multilingual Cohort,” Biomedical Signal Processing and Control, vol. 88, p. 105667, 2024

2024
[36]

Recent Advances in Drug Therapy for Parkinson’s Disease,

H. Murakami et al. , “Recent Advances in Drug Therapy for Parkinson’s Disease,” Intern Med., vol. 62, no. 1, pp. 33–42, 2023

2023
[37]

Unveiling Interpretability in Self- Supervised Speech Representations for Parkinson’s Diagno sis,

D. Gimeno-G´ omez et al. , “Unveiling Interpretability in Self- Supervised Speech Representations for Parkinson’s Diagno sis,” IEEE Journal of Selected Topics in Signal Processing , pp. 1–14, 2025

2025
[38]

Towards an Automatic Evaluation of the Dysarthria Level of Patients with Parkinson’s Disease,

J. C. V´ asquez-Correa et al., “Towards an Automatic Evaluation of the Dysarthria Level of Patients with Parkinson’s Disease, ” Jour- nal of Communication Disorders, vol. 76, pp. 21–36, 2018

2018
[39]

Interrater Reliability of the Uniﬁed Parkin- son’s Disease Rating Scale Motor Examination,

M. Richards et al. , “Interrater Reliability of the Uniﬁed Parkin- son’s Disease Rating Scale Motor Examination,” Movement Dis- orders, vol. 9, no. 1, pp. 89–91, 1994

1994
[40]

Evaluation of Parkinson’s Disease: Relia- bility of Three Rating Scales,

A. Ginanneschi et al., “Evaluation of Parkinson’s Disease: Relia- bility of Three Rating Scales,” Neuroepidemiology, vol. 7, no. 1, pp. 38–41, 1988

1988
[41]

Effect of Thalamotom y and Levodopa Therapy on the Speech of Parkinson Patients,

C. E. Quaglieri and G. G. Celesia, “Effect of Thalamotom y and Levodopa Therapy on the Speech of Parkinson Patients,” Eur Neu- rol., vol. 15, no. 1, pp. 33–9, 1977

1977

Pith tools

Reviewed August 7, 2026 · model on record in the stance chip above.

[1] [1]

Introduction Parkinson’s disease (PD) is a progressive neurodegenerati ve disorder that primarily affects movement control and speec h production [1]. Patients are faced with articulatory deﬁci ts, leading to dysarthria, and voice disorders, frequently per ceived as dysprosody (resulting in voice modulation and intensity im- pairments) [2]. Within the cou...

work page

[2] [2]

A unique feature of this corpus is the inclusion of a study speciﬁcally designed to analyze the effects of medica - tion in PD patients

Data Material The European Portuguese FraLusoPark corpus [20] is com- posed of 158 speakers, 74 healthy controls and 84 PD pa- tients. A unique feature of this corpus is the inclusion of a study speciﬁcally designed to analyze the effects of medica - tion in PD patients. To this end, participants diagnosed wit h PD were recorded twice: at least 12 hours a...

work page

[3] [3]

Methodology This work is inspired by two well-established frameworks in the ﬁeld of automatic analysis of pathological speech: i) th e traditional approach, which relies on knowledge-based aco us- tic descriptors and conventional machine learning techniq ues, and ii) a more recent paradigm that leverage SSL speech rep- resentations and attention-based dee...

work page

[4] [4]

Results & Discussion A comprehensive overview of our experimental results for au - tomatic, speaker-independent PD medication state identiﬁ cation from speech is presented in Table 2, showing the outcomes for the three data grouping strategies and different combinati ons of model architectures and speech features. Results are prese nted at the utterance l...

work page

[5] [5]

Conclusions & Future Work In this paper, we presented our work on the automatic identiﬁ - cation of PD medication states from speech, showing its pote n- tial for more reliable remote monitoring and personalized t reat- ment scheduling. Compared to prior work [17], our speaker- independent approach offers two key advantages: i) the ability to leverage lar...

work page

[6] [6]

Acknowledgements The work of Gimeno-G´ omez and Mart´ ınez-Hinarejos was par- tially supported by GV A through Grants CIACIF/2021/295 and CIBEFP/2023/167, by Grant PID2021-124361OB-C31 funded by MCIN/AEI/10.13039/501100011033 and ERDF/EU. The work of Solera-Ure˜ na, Pompili and Abad was supported by Portuguese national funds through Fundac ¸ ˜ ao para a C...

doi:10.13039/501100011033 2021

[7] [7]

Parkinson’s Disease: Clinical Features a nd Diagno- sis,

J. Jankovic, “Parkinson’s Disease: Clinical Features a nd Diagno- sis,” Journal of Neurology, Neurosurgery & Psychiatry , vol. 79, no. 4, pp. 368–376, 2008

work page 2008

[8] [8]

Intonation an d Speech Rate in Parkinson’s Disease: General and Dynamic Aspects and Responsiveness to Levodopa Admission,

S. Skodda, W. Gr¨ onheit, and U. Schlegel, “Intonation an d Speech Rate in Parkinson’s Disease: General and Dynamic Aspects and Responsiveness to Levodopa Admission,” Journal of V oice, vol. 25, no. 4, pp. e199–e205, 2011

work page 2011