New research shows: AI detects heart failure based on the sound of your voice

class=”sc-cffd1e67-0 iQNQmc”>

Artificial intelligence (AI) can do a lot these days: whether you use ChatGPT to write texts, generate images or produce music. According to a new study by the German Heart Center of the Charité, AI should now also recognize patients with heart failure, writes the “German Health Portal”.

In collaboration with the Mayo Clinic in the US, the “AI-Based Voice Analysis for Monitoring Patient Hospitalized with Acute Decompensated Heart Failure”, or “VAMP-HF” for short, is intended to provide researchers with initial insights into the effectiveness of voice therapy. analysis with AI. This is especially helpful for people who suffer from “hydropic decompensation”. People who suffer from this disease cannot properly regulate their fluid balance.

Due to the greatly reduced pumping capacity of the heart, fluids are less able to flow from the tissue to the excretory organs. Water builds up, which can be problematic, especially in the lungs. Usually there is a deterioration of the patient’s general condition.

“More fluid in the body leads to a change in the propagation of sound waves”

As senior doctor at the Heart Center Felix Hohendanner explains, the fluid in the body has a direct influence on the sound of our voice. This could therefore be understood as an “early warning system”. “More fluid in the body leads to a change in the propagation of sound waves and therefore also to a change in the voice, which is usually inaudible to the human ear but is nevertheless measurable,” the doctor writes on the portal.

Voting brings many benefits. This measuring method is particularly suitable for carrying out independently at home. “It is painless, fast and uncomplicated, has few sources of error and requires no technical aids other than a recording device such as a smartphone,” writes Felix Hohendanner.

“We are just getting started here”

To make the whole thing work, the Berlin start-up Noah Labs-AI, which has developed the necessary software, examines hundreds of parameters in the voice samples of affected patients. The AI ​​now learns to predict the effectiveness of the treatment and to sound the alarm if the amount of fluid changes.

AI-based voice recognition can also be used for other cardiovascular diseases. Doctor Hohendanner: “We are only at the beginning of a promising development.” (mgf)

Source: Blick