Skip to main content
SearchLoginLogin or Signup

Review 3: "The Acoustic Dissection of Cough: Diving into Machine Listening-based COVID-19 Analysis and Detection"

This preprint reports on a machine learning model for detecting COVID-19 by analyzing patients’ cough sounds. Reviewers deemed the findings potentially informative and promising, with a few limitations that could be addressed.

Published onApr 07, 2022
Review 3: "The Acoustic Dissection of Cough: Diving into Machine Listening-based COVID-19 Analysis and Detection"
1 of 2
key-enterThis Pub is a Review of
The Acoustic Dissection of Cough: Diving into Machine Listening-based COVID-19 Analysis and Detection

AbstractPurposeThe coronavirus disease 2019 (COVID-19) has caused a crisis worldwide. Amounts of efforts have been made to prevent and control COVID-19’s transmission, from early screenings to vaccinations and treatments. Recently, due to the spring up of many automatic disease recognition applications based on machine listening techniques, it would be fast and cheap to detect COVID-19 from recordings of cough, a key symptom of COVID-19. To date, knowledge on the acoustic characteristics of COVID-19 cough sounds is limited, but would be essential for structuring effective and robust machine learning models. The present study aims to explore acoustic features for distinguishing COVID-19 positive individuals from COVID-19 negative ones based on their cough sounds.MethodsWith the theory of computational paralinguistics, we analyse the acoustic correlates of COVID-19 cough sounds based on the COMPARE feature set, i. e., a standardised set of 6,373 acoustic higher-level features. Furthermore, we train automatic COVID-19 detection models with machine learning methods and explore the latent features by evaluating the contribution of all features to the COVID-19 status predictions.ResultsThe experimental results demonstrate that a set of acoustic parameters of cough sounds, e. g., statistical functionals of the root mean square energy and Mel-frequency cepstral coefficients, are relevant for the differentiation between COVID-19 positive and COVID-19 negative cough samples. Our automatic COVID-19 detection model performs significantly above chance level, i. e., at an unweighted average recall (UAR) of 0.632, on a data set consisting of 1,411 cough samples (COVID-19 positive/negative: 210/1,201).ConclusionsBased on the acoustic correlates analysis on the COMPARE feature set and the feature analysis in the effective COVID-19 detection model, we find that the machine learning method to a certain extent relies on acoustic features showing higher effects in conventional group difference testing.

RR:C19 Evidence Scale rating by reviewer:

  • Reliable. The main study claims are generally justified by its methods and data. The results and conclusions are likely to be similar to the hypothetical ideal study. There are some minor caveats or limitations, but they would/do not change the major claims of the study. The study provides sufficient strength of evidence on its own that its main claims should be considered actionable, with some room for future revision.



Summary: The authors analyzed acoustic features extracted from cough sounds for COVID-19 detection, and further developed machine learning models for COVID-19 detection. Three subtasks were considered according to the patients’ symptoms. Specifically, the COMPARE feature set of 6373 dimensions was adopted, and feature importance was ranked. Different machine learning techniques were investigated and compared for COVID-19  detection. All experiments were conducted using the COUGHVID database. 

The methods used to analyze the feature importance and the machine learning methods for COVID-19 detection are reasonable, and the analysis and performance support the claim. Only some minor caveats could be improved: 

  • In table 5, it is hard to justify whether the higher performance is due to the gender/age or the number of samples. As it is imbalanced dataset, the authors should take this into account when comparing the performance in gender/age subgroups. Therefore, the claim of performance in male over female and the  age group might need further justification. 

  • Some machine learning techniques may suffer from the ‘curse of dimensionality’, such as SVM. Therefore,  the lower performance of SVM or nonlinear models when compared with linear models may not be owing to the model itself, but rather to the inappropriate features. The authors might need to justify this in system comparisons in table 4. Further, compared to other approaches employing deep learning using the same dataset, the reported performance is low. This raises the concern that if deep learning features significantly outperform this feature set, what are the main advantages of the feature analysis? Are there any other potential ways to use these most important features to achieve better performance?  

  • In table 2, it is observed that different groups of features show effectiveness in three subtasks, and task 2 requires most of the features. If the authors could discuss the potential reason, it would provide some insights and significantly enhance the understanding. 

The presented feature analysis could provide some insights but needs further justification in comparison to state-of-the-art deep learning features and system performance.  

The literature review is relatively comprehensive and covers the broad picture of disease detection using audio sounds, as well as the specific task of COVID-19 detection. The careful consideration in categorizing audio samples with and without symptoms is good. 

The manuscript is well-structured and clearly presented. The writing quality is also good.

1 of 3

No comments here

Why not start the discussion?