Biomarkers for AI-assisted diagnosis of autism in adults: A feasibility study

Autism prevalence has increased from 1 in 500 in 1998 to 1 in 35 in 2023 (25 years later). This has led to massive NHS waiting times for diagnosis particularly for adults. Diagnosis is currently labour-intensive and requires the expertise from a very limited number of highly qualified individuals. Quality of diagnosis may also be an issue, due to the push to diagnose. We want to develop cheap, portable, reliable technology to aid, speed up and provide high quality diagnosis. As an initial step towards this goal, we will do a feasibility study in adults of three technologies simultaneously – functional Near Infrared Spectroscopy (using low-power near-infrared light on the surface of the head to measure the haemodynamic brain response), speech analysis, and sleep/movement data. We will also talk to autistic adults and their clinicians about their experience of the technologies and co-design the next steps of the project with them.