ADHD in children is a leading neurodevelopmental disorder, often recognised as one of the most common causes for referral to child psychology and psychiatric clinics. Affecting about one in 20 children [1], it manifests through persistent, impairing levels of inattention, hyperactivity, and impulsivity. ADHD treatment for children typically involves a behavioural therapy and, if necessary, medication.
These therapies are not always effective or convenient: behavioural interventions require frequent attendance at multiple sessions and are often subject to long waiting lists, while pharmacological treatments can lead to undesirable side effects. Thus, there is an urgent need to develop personalised therapeutic approaches that can be delivered in home or community settings.
Neurofeedback intervention has been suggested as an alternative treatment for children with ADHD [2, 3]. However, the proposed protocols are often based on spatially imprecise measurements of neural activity, especially using electroencephalography (EEG), and the use of overly simplistic data analysis methods. Furthermore, these interventions show substantial variability in their outcomes, largely due to the limited understanding of the neural mechanisms underlying ADHD.
This research aims to develop a neurofeedback system that utilises comprehensive mapping of neural activity using the more spatially precise optically pumped magnetoencephalography (OPM), combined with advanced machine learning methods and a novel gamified paradigm. The system has the potential to alleviate ADHD symptoms in children by directly modulating the neural mechanisms underlying this condition.
Research challenges
Currently, the neural mechanisms underlying ADHD in children poorly understood. The existing MEG/EEG datasets are often limited to basic experimental conditions which impede the exploration and understanding of the complex neural mechanisms.
Current neurofeedback approaches utilise limited spatial and spectral characteristics of neural activity and largely ignore changes in neural activity caused by the stimulation itself (i.e., not adaptive). Moreover, the methodologies fail to incorporate recent advances in brain imaging technologies, computational modelling, and machine learning.
Research plan
WP1: Acquire OPM data in children with and without ADHD while playing a video game involving attentional switching between visual and auditory stimuli.
WP2: Develop machine learning models to detect specific neural patterns associated with difficulties of attentional switching (e.g., delayed or missed response) between visual and auditory modalities.
WP3: Develop a neurofeedback system that tracks neural activity in real-time and, using machine learning, sends a feedback signal to the user when the error-associated neural pattern is detected. Attention behaviour in the game will be measured to assess improvement.
Novelty
The proposed neurofeedback system leverages state‑of‑the‑art OPM, supported by a multimillion‑pound investment from Aston University and the Institute for Health and Neurodevelopment. This technology will be integrated with advanced signal processing and machine learning techniques, enabling a robust, evidence based approach grounded in a deep understanding of the neural mechanisms underlying ADHD. The project’s close partnership with Birmingham Community Healthcare NHS Foundation Trust ensures a strong pathway to real world impact and meaningful societal benefits.
References
[1] Faraone, S. et al., Nature Reviews 2024; 10:11
[2] Westwood, S. et al., JAMA Psychiatry 2025; 82(2):118-129.
[3] Schönenberg, M. et al., Lancet Psychiatry 2017; 4(9): 673-684.