MEG-GPT: A transformer-based foundation model for magnetoencephalography data
4 affiliations
- 1Department of Psychiatry, University of Oxford, Oxford, United Kingdom
- 2Oxford Centre for Integrative Neuroimaging (OxCIN), University of Oxford, Oxford, United Kingdom
- 3Nuffield Department of Clinical Neurosciences, University of Oxford, Oxford, United Kingdom
- 4Department of Engineering Science, University of Oxford, Oxford, United Kingdom
Abstract
Modelling the complex spatio-temporal patterns of large-scale brain dynamics is crucial for neuroscience, but traditional methods fail to capture the rich structure in modalities such as magnetoencephalography (MEG). Recent advances in deep learning have enabled significant progress in other domains, such as language and vision, by using foundation models at scale. Here, we introduce MEG-GPT, a transformer-based foundation model that uses time-attention and next time-point prediction. To facilitate this, we also introduce a novel data-driven tokeniser for continuous MEG data, which preserves the high temporal resolution of continuous MEG signals without lossy transformations. We trained MEG-GPT on tokenised brain region time courses extracted from a large-scale MEG dataset (N = 612, eyes-closed rest, Cam-CAN data), and show that the learnt model can generate data with realistic spatio-spectral properties, including transient events and population variability. Critically, it performs well in downstream decoding tasks, improving downstream supervised prediction task, showing improved zero-shot generalisation across sessions (improving accuracy from 0.56 to 0.59) and subjects (improving accuracy from 0.45 to 0.49) compared with a PCA baseline method. Furthermore, we show the model can be efficiently fine-tuned on a smaller labelled dataset to boost performance in cross-subject decoding scenarios. This work establishes a powerful foundation model for electrophysiological data, paving the way for applications in computational neuroscience and neural decoding.
Related
- MEG-Mamba: A Scalable State-Space Foundation Model for Magnetoencephalography
- Transductive Learning From a Transformer Foundation Model Improves the Taxonomic Classification of eDNA Metabarcodes
- Neural oscillations and network specialisation in early childhood captured with next-generation magnetoencephalography
- Variance of human cortical synaptic transmission inferred from non-invasive magnetoencephalography of somatosensory evoked population spikes
- Spatiotemporal dynamics of lexical and morphosyntactic planning during language production revealed by magnetoencephalography
BibTeX
@article{huang2026meggpt,
title = {{MEG-GPT: A transformer-based foundation model for magnetoencephalography data}},
author = {Rukuang Huang and Sungjun Cho and Chetan Gohil and Ōiwi Parker Jones and Mark W. Woolrich},
journal = {Imaging Neuroscience},
year = {2026},
month = jun,
doi = {10.1162/imag.a.1301},
url = {https://doi.org/10.1162/imag.a.1301},
note = {CC BY},
}
RIS
TY - JOUR
TI - MEG-GPT: A transformer-based foundation model for magnetoencephalography data
AU - Rukuang Huang
AU - Sungjun Cho
AU - Chetan Gohil
AU - Ōiwi Parker Jones
AU - Mark W. Woolrich
PY - 2026
DA - 2026/06/26
JO - Imaging Neuroscience
DO - 10.1162/imag.a.1301
UR - https://doi.org/10.1162/imag.a.1301
AB - Modelling the complex spatio-temporal patterns of large-scale brain dynamics is crucial for neuroscience, but traditional methods fail to capture the rich structure in modalities such as magnetoencephalography (MEG). Recent advances in deep learning have enabled significant progress in other domains, such as language and vision, by using foundation models at scale. Here, we introduce MEG-GPT, a transformer-based foundation model that uses time-attention and next time-point prediction. To facilitate this, we also introduce a novel data-driven tokeniser for continuous MEG data, which preserves the high temporal resolution of continuous MEG signals without lossy transformations. We trained MEG-GPT on tokenised brain region time courses extracted from a large-scale MEG dataset (N = 612, eyes-closed rest, Cam-CAN data), and show that the learnt model can generate data with realistic spatio-spectral properties, including transient events and population variability. Critically, it performs well in downstream decoding tasks, improving downstream supervised prediction task, showing improved zero-shot generalisation across sessions (improving accuracy from 0.56 to 0.59) and subjects (improving accuracy from 0.45 to 0.49) compared with a PCA baseline method. Furthermore, we show the model can be efficiently fine-tuned on a smaller labelled dataset to boost performance in cross-subject decoding scenarios. This work establishes a powerful foundation model for electrophysiological data, paving the way for applications in computational neuroscience and neural decoding.
SN - 2837-6056
ER -