An international team, led by researchers at the King’s Maudsley Partnership for Children & Young People, have received £2.2million in Wellcome Discovery funding to use AI to predict children at risk of OCD.

Obsessive-compulsive symptoms can be present in one in five children. Obsessions can be intrusive, unwelcome thoughts and compulsions may appear as repetitive, lengthy rituals.

Whilst most children and young people grow out of these symptoms, a small percentage continue to experience obsessions and compulsions that then can impair a person’s life, which can lead to a formal diagnosis of Obsessive Compulsive Disorder (OCD). By this stage, OCD can be more challenging to treat, with more chronic symptoms predicting poorer outcomes.

“Talking to young people living with OCD and their parents, it’s clear that intervening early, before symptoms spiral out of control, could prevent a lot of distress and anguish. This project aims to do just that for children not just in the UK, but in Brazil, Sweden and beyond.” – Professor Philip Shaw, Director – King’s Maudsley Partnership for Children & Young People

Currently, identifying the children most at risk of developing full OCD is difficult, and this inability to predict the onset of OCD is halting advances in treatment and understanding. Prediction is a vital step to help implement early interventions, and tailor interventions to the needs of each young person.

The aim of this Wellcome funded project is to develop a tool to pinpoint who is likely to develop OCD and when this may be. To do this, easily obtainable medical information, often found in medical records, will be combined with more complex information on genetic make-up and brain features; to find out which type of information helps most in prediction.

“OCD is often thought of as an adult condition, but symptoms frequently emerge in childhood, and that is precisely where the greatest opportunity lies. This project builds a prediction tool that combines routinely collected medical information, the kind already sitting in clinical records, with genetic and brain imaging data to identify which children are most likely to go on to develop full OCD. AI gives us the ability to integrate those complex, multi-layered signals in a way that simply wasn’t feasible before, and to do so across diverse populations in the UK, Brazil, and Sweden.” – Professor Gustavo Sudre, Professor of Genomic Neuroimaging and Artificial Intelligence

Once this prediction tool has been developed, the team will create a framework that incorporates a diverse range of views, including those of people with lived experience of OCD. The framework and tool will be piloted with parents of children deemed at risk of

OCD, with a view of creating a scalable early intervention tool for OCD. This will involve culturally adapting, enriching and evaluating a digital early intervention too, including conducting open trials in the UK and Brazil to evaluate the intervention with parents of children identified as being at-risk, through the prediction tool developed earlier in the project.

Professor Philip Shaw and Professor Gustavo Sudre from the King’s Maudsley Partnership for Children & Young People will lead an international team including, Professor Elizabeth Shephard from the University of Sao Paulo, Professor Georgina Krebs from University College London, Professor David Mataix-Cols from Karolinska and Dr Nick Sireau and Dr Margherita Zenoni from Orchard OCD.

“Families affected by OCD often ask whether it is possible to prevent the condition in the next generation,” said Professor Mataix-Cols of Karolinska Institutet, one of the international centres involved in the project. “With the support of the Wellcome Trust, we will enhance and culturally adapt our prototype intervention for children at increased risk and evaluate it in the UK and Brazil. Our ambition is to intervene before symptoms become disabling by equipping parents with practical, evidence-based tools that can reduce risk.”

“Obsessive-Compulsive Disorder often emerges silently, with early warning signs that are difficult to interpret or act upon. By developing tools that can identify who is most at risk and when, this project has the potential to shift OCD from a condition we react to, to one we can anticipate and prevent. This represents a fundamental change in how we approach OCD: moving toward earlier intervention, personalised support, and ultimately better outcomes for children and families. Crucially, this research is grounded in lived experience and spans multiple countries, ensuring the solutions developed are both scientifically robust and truly meaningful to those affected.” – Dr Nick Sireau, Co-founder and Trustee, Orchard OCD

The researchers have worked extensively with over 150 people with lived experience on the early stages of this study and will continue to work closely with those with lived experience of OCD as the research progresses to ensure the findings and meaningful and shared widely.

The Centre of Neuroimaging Sciences at Denmark Hill Campus, King’s College London, is now home to a new investigational GE HealthCare MAGNUS 3T MRI scanner, the first of its calibre in Europe. GE HealthCare is a 30 year-long industry-partner of researchers at King’s College London.

The MAGNUS system is one of the first high-performance head-only scanners, specialised for brain imaging. The new technology will allow scientists to push the boundaries of what is possible in neuroscience and psychiatric research, particularly regarding brain development.

Previously unobtainable resolution

The GE HealthCare MAGNUS 3T achieves a level of detail up to 10 times finer than previous state-of-the-art commercial systems, measuring the brain in far tinier 3D volumes (voxels). Crucially, previous scanners relied on mathematical models that could only estimate broad tissue properties. The MAGNUS system allows researchers to unlock more advanced models and reveals specific features such as individual axon (the fibres that transmit electrical signals between neurons) diameters and cell sizes.

Studying developing brains

Precisely measuring axons, cellular size and their densities is crucial to researching brain development in children and young people. In a developing brain, new connections are forming. The new scanner allows highly accurate quantification of these changes in the living brain tissue.

For several developmentally linked conditions such as ADHD, autism and psychosis, the MAGNUS scanner provides a valuable tool to measure how white matter and connections develop differently in children and young people.

“What is exciting is that we can use this novel clinical scanner to generate extremely high-quality images of brain structure and function for all ages” commented Professor Steve Williams.

Importance for children and young people

Professor Steve Williams, Head of Neuroimaging at the School of Neuroscience, and his team have spent years working to make scanning more accessible to children and young people.

Conventional scanning is extremely loud and can be time consuming. This can be challenging for young children who struggle to stay still for long periods and those with ADHD or autism who may be sensitive to noise.

The MAGNUS system is compatible with new silent scanning methods previously developed by Professor Williams’ team in collaboration with GE HealthCare and with support from the NIHR Maudsley BRC.

“Scanners can be noisy. For people feeling anxious, who are neurodiverse, those with tinnitus or auditory hallucinations, the noise can make scanning intolerable. It was important to us that we made the scanners as comfortable and quiet as possible for everyone.” – Professor Steve Williams, Head of Neuroimaging at the School of Neuroscience.

The researchers have also developed imaging methods which tolerate movement during the scan, and the new MRI technologies allow for two to three times faster scanning.

“With conventional whole-body scanners, in order to image every organ, you have to compromise,” commented Professor Williams. “The GE HealthCare MAGNUS 3T scanner is much more optimised for the head and neck, which means we can also choose to prioritise data quality, resolution or scan time. This is particularly good for children, who may not want to be in a scanner for a long time.”

These developments, specifically aimed at children and young people, will enable research into developmental conditions such as autism, ADHD and psychosis to be more inclusive

In recognition of its vital role for children and young people, the new scanner was partly funded by UK Research and Innovation (UKRI) as part of the King’s Maudsley Partnership for Children and Young People.

“The new scanner is better for kids. It’s in a lovely room, makes less noise, and during many of the scans, the child can watch a movie or listen to their favourite music, to help them feel at ease.  We want scanning to be great for science and fun for kids,” commented Professor Philip Shaw, Director of the King’s Maudsley Partnership.

An opportunity for collaboration

The scanner will be available for use by researchers across the IoPPN, NIHR Maudsley BRC, the King’s Maudsley Partnership, King’s School of Dentistry and clinicians in King’s Health Partners.

The scanner also offers opportunities for collaboration further afield across the UK and internationally. Plans are developing for an international consortium between groups using these scanners across the globe.

Professor Adam Hampshire from the School of Neuroscience, alongside Professor Philip Shaw from the School of Academic Psychiatry, and Director of the King’s Maudsley Partnership, the study will explore whether social media and AI can alter children’s development and impact real world outcomes with both positive and negative implications.

“There is growing concern about how online AI platforms and social media may be shaping the cognitive and social development of children and young people. These concerns are already influencing major policy decisions, yet there remains limited evidence about what the actual effects are, who is most vulnerable, and whether some impacts may even be beneficial.” – Professor of Cognitive and Computational Neuroscience

Using Cognitron, an online assessment platform, the team will deliver large-scale (up to 40,000 people) assessments in children from REACT, a unique randomly sampled UK cohort. Then, pseudo-trial modelling will estimate the potential causal pathways where social media and AI use affect cognitive development, mental health, lifestyle and education.

Subsequently, the team will recruit around 800 children and randomise them into two groups. One group will receive a smartphone immediately, and the other will be in the delayed-access arm, where smartphone ownership will be delayed by around six months. This will enable the researchers to develop experimental confirmation of causal effects on cognitive development, mental health and real-world function over 24 months.

“The project tackles a key aspect of the critically important issue of how social media impacts on child development. Many parents ask us, as mental health care professionals, how old their child should be before they get their first smartphone. This study will give the sort of evidence we need to give helpful, informed answers to this important question.” – Professor Philip Shaw, Director of King’s Maudsley Partnership for Children & Young People

The aim of ONSET-Mobile is to help inform parental guidance and national digital health policy around smartphone ownership and use. This will help answer a frequently disputed question, when is the appropriate time for children to be given a smartphone? Alongside this, the study will develop a unique dataset to advance developmental cognitive neuroscience.

The device – which uses an approach called trigeminal nerve stimulation (TNS) – was cleared for use by the US Food and Drugs Administration (FDA) to treat ADHD in 2019 based on a small study. These new findings from a larger multicentre trial, published in the journal Nature Medicine, suggest authorities should revisit the original evidence that supported the FDA clearance. Notably, TNS is currently not recommended for use in the UK by NICE guidelines.

The trial was run in collaboration with University of Southampton and funded by a partnership between the National Institute for Health and Care Research (NIHR) and Medical Research Council (MRC) with further support from the NIHR Maudsley Biomedical Research Centre.

Attention-Deficit/Hyperactivity Disorder (ADHD) affects 5-8 per cent of school-age children worldwide and is associated with age-inappropriate problems with attention and/or hyperactivity and impulsivity that can impair everyday functioning. Stimulant medications improve symptoms in 70 % of those who take them in the short term but there is less evidence of their long-term effects.

To provide an alternative to medication researchers have developed and trialled approaches that use non-invasive stimulation of the brain, working on the regions that have been identified as influential in ADHD.

One of these approaches involves stimulating the trigeminal nerve (TNS), targeting a branch of this facial nerve which is thought to activate the brainstem and from there other brain regions that may be relevant to ADHD, in particular the locus coeruleus, which plays a role in arousal which is typically diminished in people with ADHD. TNS is thought to stimulate other brain regions associated with attention such as frontal and thalamic areas via the brainstem in a bottom-up manner.

A previous small trial in the US with 62 children diagnosed with ADHD has shown that when TNS is applied every night for 8 hours for one month it is effective in reducing symptoms – this research led to its clearance by the FDA for use in the US. However, the control condition involved no stimulation and blinding was not tested after one month, raising questions about a potential placebo effect.

This new UK clinical trial across two sites in London and Southampton tested TNS in a wider range of 150 children and adolescents diagnosed with ADHD aged between 8 and 18 years old and applied a more rigorous placebo condition. Half of the sample received real TNS for about 9 hours every night for 4 weeks through battery-powered electrodes applied to the forehead. The other half of the sample received the ‘sham’ condition where electrodes were still applied to the forehead every night for 4 weeks but participants only received 30 seconds of stimulation every hour at a lower frequency and pulse width, which are thought to be non-effective and hence act as a “control” condition.

Professor Katya Rubia, Professor of Cognitive Neuroscience at the Institute of Psychiatry, Psychology & Neuroscience (IoPPN), King’s College London and senior author of the study said: “Our study shows how important it is to design an appropriate placebo condition in clinical trials of brain therapies. There is a large placebo effect with high-tech brain therapies, in particular for patients and families that have an expectation that they can adjust brain differences associated with ADHD. It is hence paramount to control for placebo effects in modern brain therapies to avoid false hopes.”

Dr Aldo Conti, postdoctoral researcher at IoPPN and at Florence Nightingale Faculty of Nursing, Midwifery & Palliative Care, King’s College London and first author on the study said: “This multicentre trial was designed to address key limitations of the previous pilot study that informed FDA clearance of TNS for ADHD, particularly by using a rigorously controlled sham condition that supported successful blinding across the treatment period. Unlike the earlier study, which was limited to younger children, we also included adolescents, a clinically important group given well-documented challenges with long-term medication adherence. These design choices enabled a more robust and clinically relevant evaluation of TNS.”

By comparing the groups researchers evaluated effectiveness of TNS by assessing the symptoms of ADHD as reported by parents, alongside other outcomes such as mind-wandering and attention, depression and anxiety, and sleep.

The trial showed that TNS was safe with no serious adverse events and most participants considered it a mild or no burden to use. However, the results showed no significant change in ADHD symptoms, objective measures of hyperactivity, attention and associated behaviours around mood and sleep.

Professor Samuele Cortese, NIHR Research Professor at University of Southampton and study lead for the Southampton site, stated: “Rigorous evidence, such as that generated by this study, is essential for supporting shared decision-making regarding interventions for ADHD. It empowers individuals with ADHD and their families to make informed choices about the treatment of ADHD. Clinicians, individuals with ADHD, and their families need to know which treatments work, and which do not based on the best evidence.”

This project was funded by the Efficacy and Mechanism Evaluation (EME) Programme which is a partnership of the Medical Research Council (MRC) and National Institute for Health and Care Research (NIHR). The trial was run by the King’s Clinical Trials Unit and recruitment involved the Child and Adolescent Mental Health Services (CAMHS) clinics within the following NHS trusts: South London and Maudsley NHS Foundation Trust, Hampshire and Isle of Wight Healthcare (previously known as SOLENT NHS Trust), Central and North-West London NHS Foundation Trust, Oxleas NHS Foundation Trust and South-West London and St. George’s Mental Health NHS Trust.

External trigeminal nerve stimulation in youth with ADHD: a randomized, sham-controlled, phase 2b trial by Conti, A.A. et al was published in Nature Medicine. https://doi.org/10.1038/s41591-025-04075-x