The Biology of Autism

We aim to identify markers to predict how a person may develop and respond to interventions

These features are called biomarkers, short for biological markers. They can be any objective measure, such as the presence of a certain gene, a pattern of brain activity or a score on a questionnaire. Biomarkers could be used in the process of autism diagnosis, to predict who may develop co-occurring symptoms, and to select which interventions, supports or treatments will work best for an individual. 

To identify biomarkers, we will follow participants over time in what are called longitudinal studies. Together, the studies will cover the entire life-span, from the womb to adulthood. Participants will include autistic and non-autistic people, as well as siblings of children with autism. The research will focus on six factors thought to play a role in the core and frequently associated characteristics of autism: executive function; responses to emotion; processing of rewards; social cues; sensory information and unpredictable situations. 

Types of biomarker
  • Diagnostic biomarkers help to diagnose autism
  • Likelihood biomarkers help to predict who may develop autism before they display autistic behaviours
  • Stratification biomarkers group together people who share certain characteristics
  • Prognostic biomarkers predict a person’s future development, for example how their symptoms may progress through childhood or the likelihood of developing co-occurring conditions, such as anxiety
  • Predictive biomarkers help clinicians to decide which therapy may be most desired and most helpful to different individuals
Biomarkers and genetics

We will explore whether any combination of genes could be biomarkers for particular patterns of development or for certain co-occurring conditions. However, some people have concerns about genetic research in autism. For example, some are worried that it could lead to a test during pregnancy to identify whether a baby will develop autism and that such a test could be used in eugenics. AIMS-2-TRIALS is not developing a genetic test and is ethically opposed to eugenics. It is also unlikely that a genetic test will ever be specific or accurate enough to predict whether a person will develop autism. This is because, for most autistic people, many thousands of common genetic variants contribute to causing their autism and, even then, this genetic predisposition interacts with important environmental factors. The exception to this – in about 5% of autistic people – is that there are a small number of rare genetic variants or mutations that are linked to autism.

Biomarkers beyond medicine

We recognise that while some autistic people and their families do want medicines to treat aspects of their autism or co-occurring conditions, others may prefer non-medical interventions or choose not to receive any support. There are many ways in which our work on biomarkers could inform and support the development of non-medical treatment options.

Any biomarkers that we identify could next be tested by researchers across the world to confirm that the biomarkers work as expected in people from different places and backgrounds. To make this easier, we will share our data and study designs.  When a biomarker is successfully validated, the next step will be to gain approval to use it in tests of new treatments and in clinical practice by liaising with international regulatory authorities.

We hope to produce a video with the Autism Representatives to explain biomarkers in more detail.

KCL Eva Loth

Lead: Eva Loth,
King’s College London

Jan Buitelaar,
Radboud University

Jörg Hipp,

Daniel Umbricht,