Our impact

The Foresight Large Language Model is the world’s first national-scale foundation model of electronic health records, trained on de-identified NHS data from 57 million patients to predict COVID-related and wider health outcomes, supporting earlier, life-saving interventions across physical and mental health.

Read more about Foresight in the Lancet journal article and see how Foresight is being used to train AI on NHS records to predict healthcare needs in the Massive-scale AI on de-identified NHS data project.

Cogstack AI is an information retrieval and AI platform developed at the NIHR Maudsley BRC to unlock the value of electronic health records. Using enterprise search and natural language processing, CogStack enables secure analysis of clinical data to support research, decision-making, and service improvement, and is now widely deployed across UK and international healthcare systems. Cogstack won the Artificial Intelligence in Health and Care Award.

Read more about how Cogstack is unlocking the power of healthcare data.

The RADAR-CNS and RADAR-Base programmes use wearable and smartphone data to track mental and physical health, including use in COVID monitoring research.

RADAR-Base has been globally adopted in over 20 disorder areas, 54 studies and with over 150,000 participants. Used in both mental health (e.g. depression, anxiety, stress, ADHD and psychosis) and physical health (e.g. heart disease, epilepsy and diabetes).

The EPSRC Centre for Doctoral Training in Data-Driven Health (DRIVE-Health) trains researchers in health data science, AI, and machine learning through interdisciplinary PhD training embedded in real-world healthcare. The programme focuses on applying data-driven technologies to address global health challenges, and our close links with DRIVE-Health provide MSc students with clear pathways into doctoral research that combines academic excellence with meaningful impact in healthcare.

The ADHD Remote Technology (ART) programme develops and applies wearable and smartphone-based tools to remotely assess ADHD and related traits in adolescents and adults. Linked to the RADAR-base platform, ART combines digital markers, lifestyle factors, and clinical measures to enable long-term, real-world monitoring, supporting self-management, personalised treatment, and improved outcomes, including during the transition from adolescence to adulthood.

The Ethical Learning of Artificial (eXplainable), Intelligence & Reflection (ELAXIR) tool is an interactive learning tool designed to improve AI literacy and promote ethical reflection on the use of AI in healthcare. Using physical and digital cards, scenarios, and supporting resources, ELAXIR empowers patients, clinicians, researchers, and the public to understand key AI concepts, ethical challenges, and patient rights, supporting responsible, patient-centred AI use in healthcare.

The Mental Health & Neuroscience Clinical Trial Statistics Group runs and supports trials in psychiatry and neuroscience, designing, overseeing and analysing studies that directly inform practice.

Led by Professor Ben Carter, his team has a portfolio of around 50 clinical trials that cover all phases and focus primarily on mental health, ranging from small-scale studies to large, multi-site projects.

For examples of further initiatives from from the group, see the Brief Educational Workshops in Secondary Schools (BESST), and the Pouch study.

The Precision Health Informatics Data Lab (PHIDL), led by Professor Richard Dobson, develops data-driven approaches to support precision, predictive and participatory healthcare. The group integrates ‘omics data with electronic health records, mobile health data and knowledge graphs, working closely with the NIHR Maudsley BRC to translate advanced analytics into improved understanding, diagnosis and treatment of mental disorders.

The NIHR Maudsley BRC Prediction Modelling group, led by Professor Daniel Stahl, works with clinicians at South London and Maudsley NHS Foundation Trust to advance how prediction models are developed and used in precision medicine for mental health. The group develops and applies risk and treatment-response models to improve patient outcomes and support clinical decision-making.

The Voice & Speech Processing group, led by Dr Nicholas Cummins, are developing AI-driven speech technologies to turn speech and conversational signals into digital biomarkers. These tools aim to assess mood, cognition, and neurological health, enabling more reliable, inclusive, and innovative applications in clinical research and practice.

Article

Researchers compared AI-based “ageing clocks” using blood metabolites from over 225,000 UK Biobank participants, showing how metabolite-predicted age can reveal accelerated biological ageing and predict health outcomes, frailty, and lifespan.

Read the study published in Science Advances.

Article

Researchers analysed smartphone-recorded speech from the RADAR-CNS study to track depression symptoms. Slower, quieter speech was linked to greater symptom severity, and personalised machine learning models improved detection.

Research led by Dr Nicholas Cummins – published in the Journal of Affective Disorders.

Article

ARC researchers are visualising mental health patient data to help clinicians deliver more targeted, proactive care, reducing health inequalities and improving outcomes for people with severe mental illness. Read the BMJ article.

Article

Researchers developed an NLP tool to identify adverse drug events in psychiatric health records, enabling analysis across diagnoses, demographics, and drug types with high accuracy and generalisability.

Research led by Professor Richard Dobson and Dr Zina Ibrahim – read the article in PLOS One.

Article

Research finds small rise in exposure to air pollution leads to higher risk of needing treatment.

Research led by Professor Ioannis Bakolis – read the article in the Guardian.

Article

Research shows problematic smartphone use in teens is linked to depression, anxiety, and insomnia, with affected young people twice as likely to experience anxiety and higher rates of disturbed sleep and self-harm.

Research led by Professor Ben Carter – Read the article published in PLOS One.

Article

A novel parenting smartphone app trialled during the COVID-19 pandemic was found to reduce emotional difficulties in children aged four to ten and offered a cost-effective way to support families during lockdown.

Research by Professor Kimberley Goldsmith – Read more here.

Article

Misophonia is a disorder involving strong negative reactions towards sounds such as chewing or snoring. Researchers found that one in five people in the UK suffer from misophonia.

Research by Dr Silia Vitoratou – Read the Guardian article here.