To create an ethically sourced flagship dataset to enable future research in artificial intelligence and support critical insights into the use of voice as a biomarker of health
Grantor
Name
National Institute of Biomedical Imaging and Bioengineering (NIBIB)
Grant
Name
Bridge2AI: Voice as a Biomarker of Health
Grant Number
3OT2OD032720-01S1
📊
Composition
What do the instances represent?
Representation
Voice recordings and derived data
Data Type
Derived features including spectrograms, MFCCs, acoustic features, and phonetic/prosodic features
Counts
12,523
Identification
Five main cohort categories:
- Voice Disorders
- Neurological and Neurodegenerative Disorders
- Mood and Psychiatric Disorders
- Respiratory disorders
- Pediatric Voice and Speech Disorders
Bridge2AI-Voice: An ethically-sourced, diverse voice dataset linked to health information
A comprehensive collection of data derived from voice recordings with corresponding clinical information. Contains 12,523 recordings from 306 participants collected across five sites in North America. Participants were selected based on known conditions which manifest within the voice waveform including voice disorders, neurological disorders, mood disorders, and respiratory disorders.
Role
Name
ORCID
Affiliation
Principal Investigator
Alistair Johnson
-
-
Principal Investigator
Jean-Christophe Bélisle-Pipon
-
-
Principal Investigator
David Dorr
-
-
Principal Investigator
Satrajit Ghosh
-
-
Principal Investigator
Philip Payne
-
-
Principal Investigator
Maria Powell
-
-
Principal Investigator
Anais Rameau
-
-
Principal Investigator
Vardit Ravitsky
-
-
Principal Investigator
Alexandros Sigaras
-
-
Principal Investigator
Olivier Elemento
-
-
Principal Investigator
Yael Bensoussan
-
-
2025-01-17
Description
Data collected using custom tablet application with headset in specialty clinics
Description
- Audio converted to monaural and resampled to 16 kHz
- Butterworth anti-aliasing filter applied
- Spectrograms computed using short-time FFT
- 60 MFCCs extracted
- Acoustic features extracted using OpenSMILE
- Phonetic/prosodic features computed using Parselmouth and Praat
- Transcriptions generated using OpenAI Whisper
Description
Data collection and sharing approved by University of South Florida Institutional Review Board
Description
MIT Laboratory for Computational Physiology
Description
De-identification steps:
- HIPAA Safe Harbor identifiers removed
- State/province removed
- Transcripts of free speech removed
- Raw audio waveforms omitted from initial release
🚀
Uses
What (other) tasks could the dataset be used for?
Description
Bridge2AI Voice Registered Access License
📤
Distribution
How will the dataset be distributed?
10.13026/249v-w155
🔄
Maintenance
How will the dataset be maintained?
1.1
Generated on 2025-11-09 10:17:35 using Bridge2AI Data Sheets Schema