q-bio.NCSep 3, 2026

A Roadmap for MEG Foundation Models

Authors: Philipp Thölke, Hamza Abdelhedi, Yorguin Mantilla-Ramos, Fouad Lbakali, Oumayma Gharbi, Catherine Duclos, Annalisa Pascarella, Vanessa Hadid, 

Organizations: Cognitive and Computational Neuroscience Laboratory (CoCo Lab), Université de Montréal, Montréal, QC, Canada · Department of Psychology, Université de Montréal, Montréal, QC, Canada · Mila – Quebec Artificial Intelligence Institute, Montréal, QC, Canada · IMT Atlantique, Brest, F-29238, France · 

Read PDF
Ask AI
Cite

Copy citation

DOINot available for this paper

Abstract

Foundation models are beginning to reshape brain-signal analysis by moving the field beyond task-specific decoding pipelines toward reusable models pretrained on broad neural datasets. Magnetoencephalography (MEG) is a compelling but still underdeveloped target for this shift: it captures human cortical dynamics at millisecond resolution while offering stronger spatial interpretability than EEG, making it especially valuable for source-resolved studies of perception, language, cognition, and clinical brain function. Yet MEG foundation models remain at an early stage, with only a small number of MEG-specific and MEG-inclusive multi-modal models, modest pretraining corpora, and emerging but still limited benchmarks. This perspective lays down the basic concepts needed to understand MEG foundation models and provides a didactic overview of the field's key design choices, including tokenization, sensor- versus source-space representations, sensor-geometry encoding, backbone architectures, self-supervised objectives, and pretraining data. We then offer a roadmap for future development, organized around native MEG pretraining, adaptation of EEG foundation models, transfer from generic time-series models, and multi-modal integration with EEG, fMRI, MRI, behaviour, and stimulus features. We highlight the need for coordinated infrastructure, including diverse and reusable MEG datasets, rigorous evaluation across subjects, sites, tasks, and clinical settings, and responsible data-sharing practices that address consent, privacy, access, and governance.

Figures & tables

Fig. 1: Pathways to MEG foundation models. Complementary strategies for arriving at a reusable MEG foundation model (MEG FM), differing in where the model starts and what it must be taught about MEG. Each trades investment in data, compute and expressiveness against the others: routes that inherit from EEG or from generic time-series models demand less MEG data and less compute but constrain how much MEG-specific structure the model can express, whereas MEG-native pretraining is the most expressive and the most demanding. All aim at representations that generalize across subjects, tasks, recording sites and clinical settings.
TermDefinition
Core concepts for MEG foundation models
Foundation modelA general-purpose model pretrained on broad and diverse data, then adapted to many downstream tasks rather than trained from scratch for each new task.
PretrainingThe initial large-scale training stage, usually on mostly unlabeled data, in which the model learns general representations before being adapted to specific tasks.
Fine-tuningUpdating some or all of a pretrained model’s parameters on a downstream task so that the model becomes specialized for a new dataset, question, or application.
Linear probingEvaluating a pretrained model by freezing its internal representations and training only a simple linear classifier or regressor on top of them.
Downstream taskAny task performed after pretraining, such as cognitive-state decoding, clinical classification, speech decoding, or behavioral prediction.
TABLE II: Foundation-model terminology
More in this topic

Magnetoencephalography

See all 14 papers

Explore similar work

CardsList
Feb 18, 2026

A Systematic Evaluation of Sample-Level Tokenization Strategies for MEG Foundation Models

Recent success in natural language processing has motivated growing interest in large-scale foundation models for neuroimaging data. Such models often require discretization of continuous neural time series data, a process referred to as 'tokenization'. However, the impact of different tokenization strategies for neural data is currently poorly understood. In this work, we present a systematic evaluation of sample-level tokenization strategies for transformer-based large neuroimaging models (LNMs) applied to magnetoencephalography (MEG) data. We compare learnable and non-learnable tokenizers by examining their signal reconstruction fidelity and their impact on subsequent foundation modeling performance (token prediction, biological plausibility of generated data, preservation of subject-specific information, and performance on downstream tasks). For the learnable tokenizer, we introduce a novel approach based on an autoencoder. Experiments were conducted on three publicly available MEG datasets spanning different acquisition sites, scanners, and experimental paradigms. Our results show that both learnable and non-learnable discretization schemes achieve high reconstruction accuracy and broadly comparable performance across most evaluation criteria, suggesting that simple fixed sample-level tokenization strategies can be used in the development of neural foundation models. The code is available at https://github.com/OHBA-analysis/Cho2026_Tokenizer.

SungJun Cho, Chetan Gohil, Rukuang Huang +2

Oxford Centre for Human Brain Activity, University of Oxford, Oxford OX3 7JX, U.K. · Nuffield Department of Clinical Neurosciences, University of Oxford, Oxford OX3 9DU, U.K. · Department of Psychiatry, University of Oxford, Oxford OX3 7JX, U.K. +1

Time Series TokenizationTime-Series Foundation Models
AskPDF
Mar 4, 2026

Standing on the Shoulders of Giants: Rethinking EEG Foundation Model Pretraining via Multi-Teacher Distillation

Pretraining for electroencephalogram (EEG) foundation models has predominantly relied on self-supervised masked reconstruction, a paradigm largely adapted from and inspired by the success of vision and language foundation models. However, unlike images and text, EEG datasets are notoriously expensive to collect and characterized by low signal-to-noise ratio. These challenges introduce difficulties in scaling the EEG foundation models and capturing the underlying neural semantics through reconstruction. In this work, we ask the question: can we stand on the shoulders of well-established foundation models from well-represented modalities to bootstrap the pretraining of EEG foundation models? We first demonstrate that mainstream foundation models, such as those from vision and time series, transfer surprisingly well to EEG domain. To this end, we propose the Multi-Teacher Distillation Pretraining (MTDP) framework for pretraining EEG foundation models via a two-stage multi-teacher distillation. In the first stage, we introduce a learnable gating network to fuse representations from diverse teachers (e.g., DINOv3 and Chronos) via a masked latent denoising objective. In the second stage, we distill the fused representation into an EEG foundation model. Extensive evaluations across 9 downstream tasks and 12 datasets demonstrate that our MTDP-based EEG foundation model outperforms its self-supervised counterparts while requiring only 25% of the pretraining data.

Chenqi Li, Yu Liu, Shuo Zhang +2

Department of Engineering Science, University of Oxford.

EEG Foundation ModelsMulti-Teacher Knowledge DistillationCross-Modal Knowledge Distillation
AskPDF
May 12, 2026

What Do EEG Foundation Models Capture from Human Brain Signals?

Clinical electroencephalogram (EEG) analysis rests on a hand-crafted feature catalog refined over decades, e.g., band power, connectivity, complexity, and more. Modern EEG foundation models bypass this catalog, learn directly from raw signals via self-supervised pretraining, and match or outperform feature-engineered baselines on most clinical benchmarks. Whether the two representations align is an open question, which we decompose into three sub-questions: what does the model learn, what does the model use, and how much can be explained. We answer them with layer-wise ridge probing, LEACE-style cross-covariance subspace erasure, and a transparent classifier benchmarked against a random-feature baseline. The audit covers three foundation models (CSBrain, CBraMod, LaBraM), five clinical tasks (MDD, Stress, ISRUC-Sleep, TUSL, Siena), and a 6-family 63-feature lexicon. Of the 945945 (model, task, feature) units, 648648 (68.6%68.6\%) are representation-causal and 199199 (21.1%21.1\%) are encoded-only. Across tasks, 5050 features qualify as universal candidates with strong support (all three architectures RC) in two or more tasks. Frequency-domain features dominate, but the other five families each contribute substantial causal mass. Confirmed features recover, on average, 79.3%79.3\% of the foundation model's advantage over the random baseline, with a clean task gradient (MDD ≈0.99\approx 0.99 down to Stress ≈0.56\approx 0.56): tasks near ceiling are almost fully recovered by the lexicon, while harder tasks leave a non-trivial residual that pinpoints a concrete target for future concept discovery.

Ling Tang, Qian Chen, Jilin Mei +6

Shanghai Artificial Intelligence Laboratory · Shanghai Jiao Tong University · Tongji University +2

Frequency-Domain Feature LearningRepresentation ProbingEEG Foundation Models
AskPDF