This is the first study to integrate graph-based representations into conditional positional encoding and introduce adaptive masking within EEG foundation model pretraining and demonstrates that the proposed GCN-based positional encoding and adaptive masking strategies achieve state-of-the-art or competitive performance, highlighting their effectiveness in EEG foundation modeling.
Abstract
Electroencephalography (EEG) foundation models have emerged as promising paradigms for decoding brain activity through large-scale self-supervised pretraining. Notably, most existing approaches employ masked patch reconstruction, in which randomly selected EEG patches are masked and recovered by the network. However, purely random masking may limit representation learning as informative patches from highly correlated channels often remain unmasked, causing the network to underutilize weakly correlated channels or serial dependencies. Furthermore, because these models are typically based on transformer architectures, their performance is influenced by the choice of positional encoding. Although positional encoding has been extensively explored in natural language processing and computer vision, few studies have adapted it to the unique characteristics of EEG signals. To address these limitations, this study introduces a conditional positional encoding method based on graph convolutional networks (GCNs) to explicitly model inter-channel correlations, along with an adaptive masking strategy that enhances representation learning by modulating the difficulty of masked patch reconstruction. To the best of our knowledge, this is the first study to integrate graph-based representations into conditional positional encoding and introduce adaptive masking within EEG foundation model pretraining. Experimental results on ten benchmark datasets demonstrate that the proposed GCN-based positional encoding and adaptive masking strategies, both individually and in combination, achieve state-of-the-art or competitive performance, highlighting their effectiveness in EEG foundation modeling.
Recent electroencephalography (EEG) foundation models increasingly emphasize model scale, potentially limiting their deployment on resource-constrained devices and in everyday environments. We argue that transferable EEG features can instead be learned with a compact architecture and introduce BrainCodes, a lightweight...
Kleanthis Avramidis, Ivan J. Tashev, Dimitra Emmanouilidou· 0 citations
It is argued that, as the field advances toward large-scale foundation-level feature extractors, it is imperative to move beyond simple time-domain reconstruction and to integrate spectral-domain loss terms and gradient-balancing techniques, so that the subtle high-frequency transients critical for diagnosing complex b...
David Aquilué-Llorens, Abel Puigseslloses, A. Blanco et al.· 0 citations
Electroencephalography (EEG) artifact removal remains a significant challenge in the reliable deployment of brain-computer interfaces (BCIs). BCI frameworks, such as motor imagery, SSVEP, and P300, rely on the accurate preservation of cortical rhythms for consistent decoding. Traditional deep learning based EEG denoise...
Vandana Akshath Raj, A. Pai, Puttamadappa Chaluve Gowda et al.· IEEE Access· 0 citations
Electroencephalography (EEG) foundation models are pretrained on thousands of hours of recordings, yet their frozen features transfer poorly to downstream tasks, and this transfer does not consistently improve with larger models or more pretraining data. We show that a vision foundation model that has never seen EEG is...
Foundation Models (FMs) trained on EEG signals promise subject- and dataset-independent pipelines for brain-computer interface (BCI) applications, but have yet to consistently outperform task-specific architectures, and their large representations are costly to compute and store. As a means of advancing the understandi...
L. Arvanitopoulos, A. Dragomir, C. Fidas et al.· bioRxiv· 0 citations
Electroencephalography (EEG) foundation models learn representations that combine information across channels and time. This structure motivates adapters that model the same relationships during downstream learning. However, it is not clear whether such a design outperforms other candidate adapters. Moreover, in order...
Ming-Yang Jiang, Ya-Min Li, Fan Ma et al.· 0 citations
Related blog posts
MIT News · Artificial Intelligence· news.mit.eduOct 13, 2026
Machine-learning networks can learn without detailed instructions from a designer. Those that do may offer insight into how the brain reorganizes itself during learning.
MIT News · Artificial Intelligence· news.mit.eduSep 29, 2026
Professor Sherry Turkle’s new book, “Artificial Intimacy,” offers a withering critique of chatbots and the antisocial dynamics she believes they encourage.