healthnexus tab references d4d

Grantor	Grant Name	Grant Number
National Institutes of Health	Bridge2AI: Voice as a Biomarker of Health - Building an ethically sourced, bioacoustic database to understand disease like never before.	3OT2OD032720-01S1

📊

Composition

What do the instances represent?

Instances

Name
Instance description
Representation
Voice recordings with derived spectrograms/features and participant-level phenotype/clinical data.
Instance Type
participants, sessions, and recordings (participant_id, session_id, task_name).
Data Type
Derived spectrogram arrays; acoustic, phonetic, and prosodic features; limited transcriptions; tabular phenotype and feature files.
Counts
12,523
Label
Cohort/disease group membership and clinical/phenotype variables; no raw audio included in v1.0.
Sampling Strategies
1. Name
  Targeted clinical cohort sampling
  Is Sample
  sample from patients presenting at specialty clinics at five North American sites
  Is Random
  False
  Source Data
  Adults with conditions affecting voice (voice, neurological/neurodegenerative, mood/psychiatric, respiratory disorders)
  Is Representative
  not stated
  Why Not Representative
  Targeted sampling of specific disorders; pediatric cohort not included in v1.0
  Strategies
  targeted clinical cohort sampling at participating sites

Subpopulations

Name
Cohort categories
Identification
- Disease Cohorts Identified At Enrollment
  Voice disorders; Neurological and Neurodegenerative; Mood and Psychiatric; Respiratory; Pediatric (planned but not included in v1.0).
Distribution
- Adult cohort only in v1.0; 306 participants across five sites in North America.

Distribution Formats

Name
Distribution formats
Description
- Parquet (spectrograms.parquet)
- TSV (phenotype.tsv, static_features.tsv)
- JSON (phenotype.json, static_features.json)

Distribution Dates

Name
Initial release
Description
- 2024-11-27 (v1.0)

🔍

Collection Process

How was the data acquired?

ID *

doi:10.57764/qb6h-em84

Name

Bridge2AI-Voice v1.0

Title

Bridge2AI-Voice: An ethically-sourced, diverse voice dataset linked to health information

Description

The Bridge2AI-Voice project presents a comprehensive, ethically sourced dataset enabling research on voice as a biomarker of health. Bridge2AI-Voice v1.0 provides 12,523 recordings for 306 adult participants collected across five sites in North America, with corresponding demographic, clinical, and validated questionnaire information. The initial release is considered low risk and includes derived data (e.g., spectrograms, acoustic/phonetic/prosodic features) and data dictionaries; original voice audio waveforms are not included. Participants were enrolled in predetermined groups reflecting conditions that manifest in voice (voice disorders, neurological and neurodegenerative disorders, mood and psychiatric disorders, respiratory disorders), with pediatric cohorts planned but not included in v1.0. Data were collected via a standardized protocol and subsequently de-identified using HIPAA Safe Harbor principles.

Language

Issued

2024-11-27

Page

https://doi.org/10.57764/qb6h-em84

Created By

Alistair Johnson
Jean-Christophe Bélisle-Pipon
David Dorr
Satrajit Ghosh
Philip Payne
Maria Powell
Anaïs Rameau
Vardit Ravitsky
Alexandros Sigaras
Olivier Elemento
Yael Bensoussan

Keywords

voice
bridge2ai
audio
biomarker
spectrograms
clinical
health

Addressing Gaps

Name
AddressingGap
Response
Address the lack of large, high-quality, multi-institutional, demographically diverse voice datasets linked to health and clinical data.

Sensitive Elements

Name
Sensitive health data
Description
- Contains health-related clinical and demographic questionnaire information.

Is Deidentified

Name

De-identification status

Description

HIPAA Safe Harbor identifiers removed (e.g., names, contact details, finer-than-year dates, identifiers).
State and province removed; country of data collection retained.
Transcripts of free speech audio removed.
Original audio waveforms omitted in v1.0; only derived spectrograms and features released.

Acquisition Methods

Name
Data acquisition
Description
- Standardized protocol at specialty clinics; demographic, clinical, and validated questionnaires; voice tasks (e.g., sustained vowel).
- Custom tablet application used; headset microphone when possible.
- Some participants completed multiple sessions.
Was Directly Observed
yes (voice recordings)
Was Reported By Subjects
yes (validated questionnaires)
Was Inferred Derived
yes (spectrograms, acoustic/phonetic/prosodic features, transcriptions)
Was Validated Verified
Standardized data collection protocol; IRB/REB oversight.

Collection Mechanisms

Name
Collection mechanisms
Description
- Custom tablet application for standardized data capture; headset microphone when feasible.
- Export and conversion from REDCap using an open-source b2aiprep library.
Used Software
Name URL Version
REDCap 3.20.0
b2aiprep https://github.com/sensein/b2aiprep

Name	URL	Version
REDCap		3.20.0
b2aiprep	https://github.com/sensein/b2aiprep

Data Collectors

Name
Data collection team
Description
- Project investigators at five North American clinical sites.

Ethical Reviews

Name
Ethics and review
Description
- Data collection and sharing approved by the University of South Florida Institutional Review Board.
- Submitted for review to the University of Toronto Research Ethics Board.

Preprocessing Strategies

Name
Audio preprocessing and feature extraction
Description
- Raw audio converted to mono and resampled to 16 kHz with a Butterworth anti-aliasing filter.
- Spectrograms computed via STFT with 25 ms window, 10 ms hop, 512-point FFT.
- Acoustic features via OpenSMILE; phonetic/prosodic features via Parselmouth and Praat.
- Transcriptions generated with OpenAI's Whisper Large model (free-speech transcripts later removed from release).
Used Software
Name Version
OpenSMILE
Parselmouth
Praat
Torchaudio 2.1
OpenAI Whisper Large

Name	Version
OpenSMILE
Parselmouth
Praat
Torchaudio	2.1
OpenAI Whisper Large

Cleaning Strategies

Name
De-identification and release filtering
Description
- HIPAA Safe Harbor removal; state/province removed; country retained.
- Free-speech transcripts removed; only derived data released (no raw audio).

Labeling Strategies

Name
Transcription
Description
- Automatic transcriptions generated using OpenAI's Whisper Large model; free-speech transcripts removed prior to release.
Used Software
- Name
  OpenAI Whisper Large

Raw Sources

Name
Raw audio availability
Description
- Raw audio waveforms are not included in v1.0; only spectrograms and derived features are provided. Future releases aim to include voice waveforms with additional security precautions.

External Resources

Name
External resources
External Resources
- Documentation Website
  https://docs.b2ai-voice.org
- Zenodo Record (bridge2ai Voice Redcap V3.20.0)
  https://doi.org/10.5281/zenodo.14148755
- B2aiprep Preprocessing Library
  https://github.com/sensein/b2aiprep
Archival
- Version-specific and latest DOIs provided.
Restrictions
- Registered, credentialed access; signed DUA and training required.

Retention Limit

Name

Retention limits

Description

Not specified in source.

Is Tabular

mixed (tabular TSV/JSON dictionaries and array-based Parquet spectrograms)

Subsets

Description	ID	Media Type	Name	Path	Title
Parquet dataset containing 513xN spectrograms per recording with participant_id, session_id, and tas...	spectrograms.parquet	application/x-parquet	spectrograms.parquet	spectrograms.parquet	Spectrograms derived from voice waveforms
Tab-delimited participant-level demographics, acoustic confounders, and validated questionnaire resp...	phenotype.tsv	text/tab-separated-values	phenotype.tsv	phenotype.tsv	Participant-level phenotype data
JSON data dictionary describing columns in phenotype.tsv.	phenotype.json	application/json	phenotype.json	phenotype.json	Phenotype data dictionary
Tab-delimited features derived from raw audio (one row per recording).	static_features.tsv	text/tab-separated-values	static_features.tsv	static_features.tsv	Recording-level static features
JSON data dictionary describing columns in static_features.tsv.	static_features.json	application/json	static_features.json	static_features.json	Static features data dictionary

Motivation

Composition

Collection Process

Uses

Distribution

Maintenance