Centre for Cognitive Science (COGS)

Jonathan Williams: Computational Models of Child Psychiatry

Jonathan Williams, King's College London
18 November 2014

Children love computers, but computations also happen within children. Computational modelling is a small but promising field: programmed simulations of how information is processed. Such models have been made of brain processes leading to both normal and abnormal behaviour – as well as of evolution and social interactions. Modelling of children may advance faster than that of adults, because CAMHS routinely addresses observable behaviour, with the diverse influences on it such as learning, inherited characteristics, family function, and medication.

There are many computational models of children’s mental health, and I briefly summarise several including my own work on ADHD. Currently, modellers constantly work to maximise rigour, novelty, and understandability. Key concepts in current models include adaptation, normative dysfunction, multifactoriality, population variation, emergent properties, and failure of complex systems. Future goals include models that are clinically realistic and collaboratively incrementable; and eventually, models that make useful predictions for individuals and populations.