Collection and Analysis of Repeated Speech Samples: Methodological Framework and Example Protocol
Bibliographic record
Abstract
BACKGROUND: Speech and language biomarkers have the potential to provide regular, objective assessments of symptom severity in several neurological and mental health conditions, both in the clinic and remotely. However, speech and language characteristics within an individual are influenced by multiple variables that can make findings highly dependent on the chosen methodology and study cohort. These characteristics are often not reported adequately in studies investigating speech-based health assessment, which (1) hinders the progress of methodological speech research, (2) prevents replication, and (3) makes the definitive identification of robust biomarkers problematic. OBJECTIVE: This study aims (1) to facilitate replicable speech research by presenting a transparent speech collection and feature extraction protocol and design checklist for other researchers to adapt and design for their own experiments and (2) to demonstrate in a pilot study the feasibility of implementing our example in-laboratory protocol that reduces multiple potential confounding factors in repeated recordings of healthy speech. METHODS: We developed a collection and feature extraction protocol based on a thematic literature review to enable a controlled investigation of within-individual speech variability in healthy individuals. Our protocol comprises the elicitation of read speech, held vowels, and a picture description and extraction of 14 example features relevant to health. We collected speech using a freestanding condenser microphone, 3 smartphones, and a headset to enable a sensitivity analysis across different recording devices. RESULTS: We collected healthy speech data from 28 individuals 3 times in 1 day (the "day" cohort), with the same schedule repeated 8 to 11 weeks later, and from 25 individuals on 3 days within 1 week at fixed times (the "week" cohort). Participant characteristics collected included sex, age, native language, and voice use habits. Before each recording, we collected information on recent voice use, food and drink intake, and emotional state. Recording times were also documented. Analysis relating to exploring within-individual variability within the day and week cohorts, as well as the device-type sensitivity analysis, is ongoing, with findings expected later in 2025. CONCLUSIONS: The wide variability in speech data collection, processing, analysis, and reporting in research on speech's use in clinical trials and practice is the motivation for this paper and the development of the speech curation protocol design checklist. Increased, more consistent reporting and justification of study protocols is urgently required to facilitate speech research replication and translation into clinical practice. INTERNATIONAL REGISTERED REPORT IDENTIFIER (IRRID): DERR1-10.2196/69431.
Fetched live from OpenAlex and de-inverted. Abstracts are not stored in this database: the inverted indexes are 8.6 GB of the frame’s 9.3 GB of text, and the host has 13 GB free.
How this classification was reachedexpand
Full frame distilled prediction
Teacher imitationNot calibrated prevalence, not ground truth. Human validation pending. Learned from the 10,348 direct Codex labels and 10,348 direct Gemma labels. Candidate is the union of thresholded teacher heads; consensus is their intersection. These outputs are machine_predicted_unvalidated and are not human labels or direct frontier model labels.
Codex and Gemma teacher scores by category
| Category | Codex | Gemma |
|---|---|---|
| Metaresearch | 0.003 | 0.003 |
| Meta-epidemiology (narrow) | 0.000 | 0.000 |
| Meta-epidemiology (broad) | 0.001 | 0.000 |
| Bibliometrics | 0.001 | 0.004 |
| Science and technology studies | 0.000 | 0.000 |
| Scholarly communication | 0.000 | 0.000 |
| Open science | 0.000 | 0.000 |
| Research integrity | 0.000 | 0.001 |
| Insufficient payload (model declined to judge) | 0.000 | 0.000 |
Machine scores (provisional)
The two teacher heads of the student model, read on this work. A score orders the frame for review; it never asserts a category, and the validation status ships verbatim with every row.
Baseline scores from an immature model (maturity gate not passed, 7 training rounds). Scores rank; they never assert a category.
score_only:v0-immature-baseline · verbatim from the scoring run: score_only means the number may rank works, and no category label ships from itClassification
machine, unvalidatedMachine predicted; a candidate call from one teacher head, not a consensus.
How this classification was reached, model by model and score by score, is at the end of the page under "How this classification was reached".