Chronic Obstructive Pulmonary Disease
Conditions
Keywords
Acoustic features, Statistic analysis, Mel frequency cepstral coefficients, Vowel quadrilateral
Brief summary
The Study aims to identify the changes in voice caused by Chronic Obstructive Pulmonary Disease (COPD) statistically for identifying the potential digital voice biomarkers.
Detailed description
1058 voice recordings collected via mobile phones over time in form of the Swedish utterance of vowel a will be analysed using vowel quadrilateral space and the Mann-Whitney U test. The significance threshold p-value will be set to 0.05, and vocal features consisting of 101 base line acoustic features and MFCC features will be analysed between the COPD and HC groups. The most significant features that remain under the significance threshold will be reported.
Interventions
OTHERChronic Obstructive Pulmonary Disease (COPD)
The base line acoustic and mel frequency ceprtum coefficients will be analysed statistically between COPD voices and healthy control groups to highlight the correlation between COPD and its effects on voice production.
Sponsors
Excellence Center at Linköping - Lund in Information Technology (ELLIIT)
Blekinge Institute of Technology
Study design
Observational model
CASE_CONTROL
Time perspective
CROSS_SECTIONAL
Eligibility
Sex/Gender
ALL
Age
18 Years to No maximum
Healthy volunteers
Yes
Inclusion criteria
* being 18 years old and older.
Exclusion criteria
* being under 18 years old and older.
Design outcomes
Primary
| Measure | Time frame | Description |
|---|---|---|
| Difference in Acoustic Voice Features Between COPD and Healthy Control Participants Difference in Acoustic Voice Features Between COPD and Healthy Control Participants | 30 weeks | This study assesses differences in baseline acoustic features (e.g., F0, HNR, jitter, shimmer, and formant frequencies), MFCC features (1-13 and derivatives Δ, ΔΔ), and vowel quadrilateral metrics (VSA from F1 and F2 of the sustained vowel 'a') between COPD and healthy controls. Acoustic features are extracted using Parselmouth and Praat, while MFCC features are computed with Librosa. Descriptive statistics (median, mean, std) are calculated, and group differences are tested for significance using the Mann-Whitney U test. Units include median, mean, std for each feature and VSA (Hertz²). Results will include p-values for group comparisons. |
Countries
Sweden
Outcome results
None listed