New AI-enhanced diagnostic voice test for early detection of dementia

image credit- shutterstock

Vivid Mind, a company based in Warsaw, Poland, has developed a simple short voice test that leverages artificial intelligence and leading-edge sound processing for early detection of dementia. The screening test has high sensitivity, and clinical tests have reported a high degree of accuracy.
 
The company's patented approach to audio decomposition, separation and analysis uses sound vectoring methods and algorithms called sound objects. It is based on a sinusoidal model that offers a radically new approach to computational analysis of the auditory scene.

The Vivid Mind screening method using voice as a biomarker allows one to first represent the sound spectrum in an accurate way, and then build interpretable features containing relevant information about a subject’s control over their voice. The approach focuses on a sustained utterance of a single vowel, and thus has global utility. The test itself can even be administered remotely without the help of a specialist. Vivid Mind’s testing and patient experience suggest their platform can detect dementia as far as eight years out, a rough estimate that they admit needs more validation.