Neuronal Oscillations Group

We investigate oscillatory dynamics as fundamental mechanisms shaping attention, working memory, and the integration of information across distributed systems, and we aim to translate these mechanistic insights into the study of neurodiversity and psychiatric conditions including dyslexia, attention-deficit/hyperactivity disorder (ADHD), and autism spectrum conditions. A central translational aim of our research is to identify mechanistically grounded insight that can improve diagnostic precision and inform targeted intervention strategies.

Our research focusses on core cognitive operations that are essential for naturalistic behaviours, including reading and speech comprehension. Both processes critically rely on the dynamic coordination of attention and working memory to select, maintain, and integrate information over time. During fluent reading, rapid eye movements sample visual input, and working memory supports the ongoing integration of visual, phonological, lexical, and semantic information across fixations. Similarly, successful speech comprehension depends on the continuous updating of information in working memory as linguistic input unfolds over time. We investigate how neuronal oscillations—particularly in the alpha band—orchestrate these processes, enabling temporal coordination across brain networks.

In the future, we will examine how atypical oscillatory dynamics relate to behavioural phenotypes in ADHD and autism, conditions in which attentional control, working memory, and the temporal coordination of information processing are often altered. By linking neural dynamics to cognitive and behavioural profiles in these populations, we aim to advance oscillation-based insight that can stratify neurodiverse profiles and contribute to mechanistically informed clinical approaches.

We adopt a network-based and multimodal methodological framework, integrating eye-tracking, magnetoencephalography (MEG), intracranial recordings, computational modelling, and multimodal combinations of EEG, fMRI, and TMS. Our development of Rapid Invisible Frequency Tagging (RIFT) enables precise tracking of information processing streams in the brain, and we are advancing optically pumped magnetometer MEG (OPM-MEG) technology for paediatric and clinical applications.

Attention to Reading

In our previous research, we found that parafoveal previewing allows adults to process semantic and orthographic word properties. Now, using OPM-MEG, we plan to investigate how and when parafoveal processing develops in childhood. We will develop a paediatric OPM-MEG and eye-tracking setup, aiming to explore the oscillatory neural mechanisms behind naturalistic reading, therefore contributing to developmental cognitive neuroscience. In future, we hope to apply our findings to study neurodivergence and clinically relevant reading abnormalities in children.

Working memory in natural language comprehension

By using MEG and OPM-MEG, we are investigating how working memory supports real-time interpretation of speech and reading. Specifically, we are exploring working memory-based reactivation of encoded representations during natural language processing and the brain areas involved in supporting this.

Attention-deficit/hyperactivity disorder and brain oscillations

In collaboration with researchers at Beijing Normal University, we investigate the role of neuronal oscillations in the allocation of spatial attention in children diagnosed with attention-deficit/hyperactivity disorder (ADHD). This work aims to identify oscillatory mechanisms—particularly those governing attentional selection and control—that may be altered in neurodivergent populations. By linking behavioural measures of attention to neural dynamics, our goal is to provide mechanistic insight into the neural substrates of attentional difficulties in ADHD. In the longer term, this research seeks to refine diagnostic characterisation and to inform the development of interventions that support neurodivergent individuals.

The role of subcortical regions in spatial attention

We investigate how subcortical hubs—particularly the thalamus and basal ganglia—contribute to higher cognitive functions, with a specific focus on the allocation of spatial attention. Using structural MRI and MEG, we relate hemispheric asymmetries in these nuclei to lateralised attentional performance and to attention-dependent modulation of posterior alpha-band oscillations. Our findings indicate that caudate asymmetry is linked to pseudoneglect, and that the globus pallidus is a predictor of individual differences in alpha modulation during attention tasks. In collaboration with researchers at South China Normal University, we are now extending this programme with intracranial recordings from deep brain stimulation (DBS) electrodes in patients with Parkinson’s disease, enabling direct tests of how output regions from the basal ganglia interact with cortical rhythms during cognitive control and attention.

Computational modelling

We develop computational models at the interface of neuroscience and artificial intelligence (NeuroAI) to investigate how biologically grounded temporal dynamics can enhance computation in artificial systems. Our aim is to incorporate key neural principles—such as brain oscillations and repetition suppression—into deep neural network architectures, and to test whether these mechanisms confer computational advantages analogous to those observed in the human brain. These models are constrained by MEG data from visual processing paradigms, including signatures of representational sharpening and multiplexing controlled by brain oscillations. By integrating empirical neural constraints with modern deep learning frameworks, this work seeks both to advance mechanistic models of human cognition and to inform the development of more efficient and flexible artificial systems.

OPM-MEG development and gaze decoding

In collaboration with our colleagues at the University of Birmingham, we aim to develop novel OPM-MEG systems for paediatric functional brain recordings, including infant recordings. Using the system at Oxford, we are planning to investigate gaze decoding, retinotopic mapping, and mechanisms of spatial attention. These projects will contribute to the ongoing development of OPM-MEG research and support the projects in our research group.

The FLUX Pipeline for MEG data analysis

We are actively developing a toolbox for processing SQUID- and OPM-MEG data, termed the FLUX pipeline. Originating from the Cogitate Consortium, a major collaborative effort to investigate neural mechanisms of consciousness, FLUX is implemented in MNE-Python and uses a dataset on visuospatial attention to demonstrate the full analysis workflow. The pipeline aims to standardise MEG data processing while also providing an educational resource for those new to MEG research. We frequently run interactive FLUX workshops in Oxford and worldwide.