Global vs. segmental acoustic features for dysarthria assessment in motor neuron diseases
Bibliographic record
Abstract
Dysarthria, a common symptom of motor neuron disease (MND), is primarily assessed through perceptual evaluations that are subjective, time-consuming, and require expert training. Acoustic analysis provides an objective alternative, leveraging either "global" features extracted from the entire speech signal or "segmental" features derived from individual phonemes (e.g., vowels). While segmental features offer finer-grained acoustic insights, their extraction has traditionally required manual segmentation, making the process time-consuming. This study explored the use of the Montreal Forced Aligner (MFA) for automatic vowel segmentation and compared the diagnostic utility of global versus segmental acoustic features in two machine learning tasks: dysarthria detection (i.e., distinguishing healthy controls (HCs) from speakers with dysarthria) and dysarthria severity classification (i.e., pre-, early-, and late-symptomatic stratification). Speech data were collected from 104 speakers with MND and 99 HCs. Global features were computed from voiced segments, while segmental features were derived from automatically aligned vowels. Four tree-based classifiers - Decision Tree, Random Forest, XGBoost, and LightGBM - were trained using 10-fold cross-validation. Feature importance was assessed using SHAP values, and statistical tests identified features with significant group differences. The MFA achieved alignment accuracy of at least 86% for healthy and early-symptomatic speakers, declining to 72% in late-stage dysarthria. For dysarthria detection, global features were significantly more effective than segmental features only in the XGBoost model. In contrast, segmental features significantly outperformed global features in dysarthria severity classification across all ensemble classifiers. These findings support the use of automated segmental analysis as an objective, viable, and clinically meaningful approach for assessing dysarthria in MNDs.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame distilled prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. Learned from the 10,348 direct Codex labels and 10,348 direct Gemma labels. Candidate is the union of thresholded teacher heads; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels or direct frontier model labels.
Codex and Gemma teacher scores by category
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.000 | 0.000 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.000 | 0.000 |
| Bibliometrics | 0.000 | 0.000 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.000 | 0.000 |
| Open science | 0.000 | 0.000 |
| Research integrity | 0.000 | 0.000 |
| Insufficient payload (model declined to judge) | 0.000 | 0.000 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one teacher head, not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".