Digital biomarkers to assist in the diagnosis of Autism Spectrum Disorder

Autism spectrum disorder (ASD) is a complex neurodevelopmental disorder that affects individuals from an early age. The purpose of this research is to study the screening performance and effectiveness of an artificial intelligence (AI) based solution using vocal, visual and emotion recognition properties to assist in diagnosing ASD. The AI tool M.appy, developed by Vitafluence.ai, is designed to automatically process the voice samples, gaze patterns, facial expressions, and emotional recognition data provided by the participants via a digital platform. The aim of the newly developed tool is to provide immediate assistive diagnostics which will enable faster access to required therapies. In the planned project, we aim to assess objective biomarkers that can be captured using widely accessible digital devices such as voice biomarkers to screen and monitor the expression of ASD traits in a cohort of individuals with neurodevelopmental disorders comparing the data with a control group. We will collect voice samples, track eye movements and facial expressions, and obtain behavioral data from emotion recognition tasks of the participants, process the obtained data, and enter it in our deep learning models to generate probabilities of belonging on the autism spectrum. The biomarkers were created to represent a basis for the generation of a simplistic and brief app-based assessment of ASD likelihood in the future.