Detection of Atrial Fibrillation Through Voice Signal Processing

This is a prospective, single-center cohort study of 40 patients undergoing cardioversion for atrial fibrillation at the Minneapolis Heart Institute Foundation. The study is designed to evaluate whether Vital Audio's investigational vocal biomarker software can reliably distinguish between regular (sinus) and irregular (atrial fibrillation) heart rhythms. Participants will serve as their own controls. They will complete three structured voice recordings: pre-cardioversion, immediately after cardioversion, and at a follow-up visit (1 week to 1 month later). At each time point, ECG data will be collected as the reference standard. Voice samples will be captured through automated phone calls initiated via the Vital Audio platform, with each call lasting about one minute. The software analyzes short vowel sounds in real time to extract heart rate, heart rate variability, and rhythm classification (regular vs. irregular). Data are stored securely, and results are available only to study staff. Patients and their providers will not receive study results directly. The study is minimal risk, with no cost to participants. Compensation is provided for participation. While individuals may not personally benefit, the study aims to generate feasibility data on voice-based heart rhythm monitoring, potentially reducing reliance on costly consumer devices or frequent clinic visits. Findings will support future large-scale trials and may contribute to improved remote monitoring strategies for atrial fibrillation.