03 | Engineering Health: Causality in Healthcare AI Hub (CHAI)

First Floor | Kitchen

The CHAI - Causality in Healthcare AI - Hub is here to conduct ground breaking research, drive innovation, and build a community.  

CHAI aims to develop a fully explainable causal AI platform specifically addressing unique challenges from healthcare across prevention, diagnosis, and treatment. CHAI is supported by the UKRI AI programme and the Engineering and Physical Sciences Research Council.

The CHAI Hub sits within the School of Engineering at The University of Edinburgh, and is located at the Usher Building, working in partnership with the Usher Institute and wider College of Medicine and Veterinary Medicine.

This poster provides an overview of the EPSRC £12m AI Hub, detail our aims and potential impact, encourage collaboration within our ecosystem, and lay out ways that people can engage with us.


This poster provides an overview of the AI Hub's research aims. It will present early research findings and lay out research plans for the next four years.


Our multidisciplinary team addresses a major healthcare priority of epilepsy care for children at the community level by developing a person-specific solution to detect and monitor childhood epilepsy.

We identified this gap thanks to our ongoing engagement with families of children with epilepsy and charitable organisations such as the UK Infantile Spasms Trust.

Epilepsy affects over 50 million globally and around 600,000 people in the UK. It is particularly common in young children, impacting their development and quality of life, and having potentially far-reaching consequences into adulthood. Major barriers to access healthcare still exist, particularly the brainwave test needed to confirm the diagnosis and monitor epilepsy, making the diagnosis and monitoring process unfair to families from remote locations or deprived socioeconomic backgrounds.

Taking Infantile Spasms – a severe form of childhood-epilepsy – as proof of concept, we have demonstrated that appropriate machine learning algorithms can reliably detect this condition from a reduced number of electroencephalogram (EEG) channels. Our PPIE with families and practitioners has identifiers barriers and enablers of early diagnosis of Infantile Spasms as well.

We will continue our work towards co-delivering a fair and equitable solution for person-specific early diagnosis, monitoring and prediction of response to treatment of childhood epilepsies. Our vision is to develop a solution applicable at home and community healthcare settings (GP, pharmacy, etc.) while addressing issues of fairness and equity due to sociodemographic characteristics and yet unaddressed data biases head-on, ultimately improving quality of life for children and their families.

Exhibitors: Javier Escudero, Bartlomiej Chybowski, Iva Peh, Alfredo Gonzalez-Sulser and Jay Shetty


This article was published on