Proposal for the Design of an Electronic Auscultation System for the Acquisition and Digital Analysis of Heart Sounds

Abstract

Cardiac auscultation is a fundamental procedure in clinical practice for the early detection of cardiovascular diseases. However, in primary care services, it faces limitations due to environmental noise and variability in the interpretation of heart sounds. Given this problem, this research seeks to propose the design of a prototype electronic cardiac auscultation device with digital recording and analysis capabilities using MATLAB, to identify cardiac abnormalities that support diagnosis in primary care settings. A two-phase methodology was applied. In the study phase, the electronic stethoscope prototype was built, and the digital processing system was developed based on a recurrent neural network implemented in MATLAB for the classification of heart sounds. In the validation phase, the device was evaluated under simulated conditions with available recordings and by comparison with the diagnosis made by medical personnel through traditional auscultation. The results obtained show that an electronic device is capable of reliably capturing heart sounds and performing basic classification of normal and abnormal recordings in the identification of murmurs. The study showed that a low-cost prototype can be technically viable to complement traditional auscultation and that, with technical improvements and broader clinical validation, these types of devices have the potential to become a diagnostic support tool in primary care.