A novel infrasound and audible machine-learning approach to the diagnosis of COVID-19

Guy Dori; Noa Bachner-Hinenzon; Nour Kasim; Haitem Zaidani; Sivan Haia Perl; Shlomo Maayan; Amin Shneifi; Yousef Kian; Tuvia Tiosano; Doron Adler; Yochai Adir

doi:10.1183/23120541.00152-2022

Figures

Tables
Supplementary Materials

Download figure
Open in new tab
Download powerpoint
FIGURE 1
A map of the 14 sites (A–N) used for acquiring the acoustic data.
Download figure
Open in new tab
Download powerpoint
FIGURE 2
Optional cloud connectivity for the VoqX. BT: Bluetooth; EMR: electronic medical record.
Download figure
Open in new tab
Download powerpoint
FIGURE 3
Flowchart of the pre-processing and machine-learning algorithm. SVM: support vector machine.
Download figure
Open in new tab
Download powerpoint
FIGURE 4
Two features (out of 164) that strongly depend on infrasound: Mel-Frequency Cepstrum Coefficient 1 (MFCC1) as a function of magnitude of breathing frequency for COVID-19, normal and non-COVID-19: a) with infrasound and b) no infrasound.
Download figure
Open in new tab
Download powerpoint
FIGURE 5
Visual sound signatures of the VoqX recorded from: a) a healthy subject, b) a COVID-19 pneumonia patient and c) a patient with non-COVID-19 pneumonia. The colours represent the intensity of breathing sounds (dB full scale).
Download figure
Open in new tab
Download powerpoint
FIGURE 6
Receiver operating characteristic curves after classification with and without infrasound: a) detection of “silent” COVID-19 pneumonia and b) detection of non-COVID-19 pneumonia.

TABLE 1

Selected patient characteristics and physical findings by lung auscultation

	Subjects	Age (years)	Height (cm)	Weight (kg)	Smoker	Gender (F/M)	Lung auscultation
	Subjects	Age (years)	Height (cm)	Weight (kg)	Smoker	Gender (F/M)	Crackles (%)	Wheezes (%)	Crackles and wheezes (%)	Reduced intensity breathing sound (%)	Vesicular breathing sounds (%)
Healthy	141	42±19	168±11	73±21	52	58/83	16	0	0	8	76
Non-COVID-19 pneumonia	61	62±17*	166±9^#,*	81±20	38^#,*	33/28	59^#,*	7^#,*	3^#,*	3	28^#,*
COVID-19 pneumonia	164	58±12*	170±11	78±28	32	77/87	19*	4*	0	2	75

Data are presented as n or mean±sd, unless otherwise stated. F: female; M: male. *: p<0.05 versus healthy; ^#: p<0.05 versus COVID-19 pneumonia.

TABLE 2

Performance measures from 12 runs of random sets

	Specificity (%)	Sensitivity		NPV (%)	PPV		Accuracy
	Specificity (%)	Non-COVID-19 pneumonia (%)	COVID-19 pneumonia (%)	NPV (%)	Non-COVID-19 pneumonia (%)	COVID-19 pneumonia (%)	Accuracy
Test group^#
With IFS	93±4**	70±7	97±2**	90±4**	90±4*	94±3**	92±1**
No IFS	80±7	69±11	93±4	85±5	84±6	85±3	85±2
Cross-validation^¶
With IFS	87±2**	71±7*	93±2**	88±3*	78±6**	89±1**	87±2**
No IFS	75±4	65±8	85±3	79±2	68±5	80±2	78±2
Auscultation by acoustic stethoscope (95% CI)⁺	76 (68–83)	72 (59–82)	25 (17–32)	43 (37–50)	71 (62–79)		52 (44–66)

VoqX data are presented as mean±sd. IFS: infrasound; NPV: negative predictive value; PPV: positive predictive value. ^#: not in the learning process; ^¶: optimisation of the learning process; ⁺: all data. *: p<0.05; **: p<0.01 versus no IFS.

Supplementary Materials

Figures
Tables

Supplementary Material
Please note: supplementary material is not edited by the Editorial Office, and is uploaded as it has been supplied by the author.
Features for machine learning and their p-values for separating the three groups 00152-2022.supplement