Linear effects of glucose levels on voice fundamental frequency in type 2 diabetes and individuals with normoglycemia.
Level 3 - non-randomized controlled study
Prospective observational cohort study with repeated within-subject measurements
PubMed 39198592 · doi:10.1038/s41598-024-69620-z
What was done
Researchers evaluated the relationship between blood glucose and vocal fundamental frequency (F0) in 505 participants. Participants wore a continuous glucose monitor (CGM) and recorded their voices using a custom mobile application up to six times daily over a 2-week period. CGM glucose measurements were paired with the voice recordings to analyze within-individual associations across normoglycemic, prediabetic, and type 2 diabetic subgroups.
What was found
Glucose levels and fundamental frequency (F0) had a significant positive association within an individual. A 1 mg/dL increase in CGM-recorded glucose corresponded to a 0.02 Hz increase in F0 (CI 0.01-0.03 Hz, P < 0.001). This effect was also observed when participants were split into non-diabetic (P = 0.03), prediabetic (P = 0.01), and type 2 diabetic (P = 0.01) classifications.
Why it matters
This study provides empirical within-subject evidence that fluctuations in blood glucose measurably alter vocal fold vibrational frequency. It demonstrates the physiological plausibility of acoustic vocal monitoring while highlighting that predictive models require individual calibration.
Limits
The abstract notes a high intraclass correlation, indicating substantial baseline variation between individuals that prevents generic, unpersonalized predictive models. The study duration was limited to two weeks, and the abstract does not report participant demographics, specific speech tasks, ambient recording conditions, or physiological confounders such as hydration and medication use.
Cited by
- context Diabetes can be detected from acoustic voice analysis due to spectral sound changes associated with dehydration.