Canada Spends Logo

Explorez notre analyse du budget 2026

Grant: Master's Award: Canada Graduate Scholarships

University of Alberta

$17,500.00

Hey Siri, Does My Cough Sound Contagious? Classifying Respiratory Disease Using Acoustic Attributes of Voluntary & Involuntary Coughs

EF 2022-23

Résumé
Assessment of cough symptoms offers clinicians valuable information to diagnose patients with respiratory conditions. Primary care physicians may not have as many tools or as much experience as respiratory experts to identify diseases based on coughs, but are often the first point of contact for these patients. Artificial intelligence (AI) cough recognition tools show promise in utilizing acoustic features to consistently distinguish between coughs of different diseases. This study will aid the development of an accessible AI cough classification tool to pinpoint respiratory pathologies in primary care. To determine the extent to which acoustic cough patterns can facilitate diagnosis, we will collect voluntary and involuntary cough data from 60 healthy adults with aerosolized capsaicin. Using the speech analysis software Praat, the data will be converted into spectrograms. Acoustic correlates of the cough signal (e.g., total power, rise-time, peak frequency, duration) will be characterized and compared. The AI will be pre-trained with 2,800 recordings from the cough sound database COUGHVID, which all have been expertly diagnosed for medical abnormalities from sound pattern. The collected coughs will then be used to train the AI. External validation will be performed using clinical cough samples collected from patients with respiratory diseases attending an otolaryngology clinic. Our AI tool will provide rapid and accurate diagnoses, ensuring that physicians can provide more efficient care, as well as alleviate patient burden by reducing exposure to unnecessary investigations, such as radiation. Other than diagnosis, AI tools harnessing acoustic cough signals have the potential to assist clinicians in judging the transmissibility and intensity of respiratory conditions.
Programme
Master's Award: Canada Graduate Scholarships
Bénéficiaire
Emplacement
Alberta, Canada
Type de bénéficiaire
Academia
Type
Grant
Source
CIHR Funding Decisions Database

Tel que publié

Id
486454
Name
Hiew Carole
Region
Alberta
Country
Canada
Orgtype
Institution Paid; Research Institution
Prcname2
Special Cases - Awards Programs
Projecttitle
Hey Siri, Does My Cough Sound Contagious? Classifying Respiratory Disease Using Acoustic Attributes of Voluntary & Involuntary Coughs
Programtype2
Studentship Programs
Orgnameinp2
University of Alberta
Cihramount2
$17,500
Cihrequipment2
$0
Programname2
Master's Award: Canada Graduate Scholarships
Keyworddelim
Acoustic Analysis; Ai Diagnostics; Artificial Intelligence; Cough Classification; Cough Signature; Cough-Based Diagnosis; Machine Learning; Primary Care; Respiratory Diseases; Respiratory Illness Diagnosis
Abstract
Assessment of cough symptoms offers clinicians valuable information to diagnose patients with respiratory conditions. Primary care physicians may not have as many tools or as much experience as respiratory experts to identify diseases based on coughs, but are often the first point of contact for these patients. Artificial intelligence (AI) cough recognition tools show promise in utilizing acoustic features to consistently distinguish between coughs of different diseases. This study will aid the development of an accessible AI cough classification tool to pinpoint respiratory pathologies in primary care. To determine the extent to which acoustic cough patterns can facilitate diagnosis, we will collect voluntary and involuntary cough data from 60 healthy adults with aerosolized capsaicin. Using the speech analysis software Praat, the data will be converted into spectrograms. Acoustic correlates of the cough signal (e.g., total power, rise-time, peak frequency, duration) will be characterized and compared. The AI will be pre-trained with 2,800 recordings from the cough sound database COUGHVID, which all have been expertly diagnosed for medical abnormalities from sound pattern. The collected coughs will then be used to train the AI. External validation will be performed using clinical cough samples collected from patients with respiratory diseases attending an otolaryngology clinic. Our AI tool will provide rapid and accurate diagnoses, ensuring that physicians can provide more efficient care, as well as alleviate patient burden by reducing exposure to unnecessary investigations, such as radiation. Other than diagnosis, AI tools harnessing acoustic cough signals have the potential to assist clinicians in judging the transmissibility and intensity of respiratory conditions.
Competitiondate
202212
Approvedterm2
1 yr 0 mth
Pinamesdelim
Hiew, Carole
Approvedterm
1 an 0 mois
Cihrcontribution2
~ $17,500~ $0~