Interpreting glottal flow dynamics for detecting COVID-19 from voice

TL;DR

This paper introduces a CNN-based method analyzing differences in inferred and modeled glottal flow waveforms to detect COVID-19 from voice, revealing respiratory impairments through voice production anomalies.

Contribution

It proposes a novel approach combining physical modeling and deep learning to identify COVID-19 related voice anomalies, which has not been explored before.

Findings

Effective detection of COVID-19 from voice using the proposed method

Identification of significant voice features linked to respiratory impairment

Demonstrated viability on a clinical dataset

Abstract

In the pathogenesis of COVID-19, impairment of respiratory functions is often one of the key symptoms. Studies show that in these cases, voice production is also adversely affected -- vocal fold oscillations are asynchronous, asymmetrical and more restricted during phonation. This paper proposes a method that analyzes the differential dynamics of the glottal flow waveform (GFW) during voice production to identify features in them that are most significant for the detection of COVID-19 from voice. Since it is hard to measure this directly in COVID-19 patients, we infer it from recorded speech signals and compare it to the GFW computed from physical model of phonation. For normal voices, the difference between the two should be minimal, since physical models are constructed to explain phonation under assumptions of normalcy. Greater differences implicate anomalies in the bio-physical factors that contribute to the correctness of the physical model, revealing their significance indirectly. Our proposed method uses a CNN-based 2-step attention model that locates anomalies in time-feature space in the difference of the two GFWs, allowing us to infer their potential as discriminative features for classification. The viability of this method is demonstrated using a clinically curated dataset of COVID-19 positive and negative subjects.