{
  "abstract": "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 understanding of the Foundation Models' inner mechanisms, we propose a lightweight alternative that exploits the self-attention mechanism of REVE, a pretrained EEG FM, to derive features similar to weighted Functional Connectivity (FC) without any fine-tuning. Inspired by kernel-based FC matrices, we aggregate the raw first-layer attention weights of a single head over the full duration of dataset trials into weighted channel-to-channel adjacency matrices. These matrices replace standard encoder representation vectors as classification features, substantially reducing computational cost. We further introduce a label-free, symmetry-based criterion, motivated by the emergence of near-symmetric attention patterns in trained encoder-decoder transformers, ranking attention heads by how much discriminative information they carry. Classical machine learning classifiers are then trained on the single most informative head's attention-derived FC matrix and achieve balanced accuracy and F1 scores on the SEED-IV emotion-recognition dataset that are competitive with several fully fine-tuned state-of-the-art EEG Foundation Models, requiring at least two orders of magnitude fewer calculations than the standard use of the encoder. Subsequent cortical region analysis revealed a significant over-representation of frontal region connections by ranking the most discriminative FC matrix connectivity edges, validating our methods as consistent with prior affective-computing literature. These results suggest that interpretable, computationally inexpensive linear probing of attention maps is a viable route to deploying EEG Foundation Models in resource-constrained settings.",
  "abstract_shown": true,
  "affiliations": [
    "Neuroengineering and Brain-Computer Interfaces Research Infrastructure, University of Patras, 26504 Patras, Greece"
  ],
  "article_type": "New results",
  "arxiv": null,
  "author_total": 4,
  "authors": [
    {
      "name": "L. N. Arvanitopoulos"
    },
    {
      "name": "A. Dragomir"
    },
    {
      "affs": [
        0
      ],
      "name": "C. A. Fidas"
    },
    {
      "name": "A. Bezerianos"
    }
  ],
  "bibtex": "https://research.draconitic.com/paper/1afy9721ymdh1.bib",
  "cited_by": 0,
  "doi": "10.64898/2026.11.05.764629",
  "fulltext_url": "https://www.biorxiv.org/content/10.64898/2026.11.05.764629v1.full",
  "id": "1afy9721ymdh1",
  "issns": [],
  "landing_url": "https://www.biorxiv.org/content/10.64898/2026.11.05.764629v1",
  "language": null,
  "licence": "CC-BY",
  "licence_label": "CC BY",
  "licence_source": "article",
  "licence_url": "https://creativecommons.org/licenses/by/4.0/",
  "metadata_licence": null,
  "native_subjects": [
    [
      "biorxiv",
      "neuroscience"
    ]
  ],
  "openalex": null,
  "pdf_url": "https://www.biorxiv.org/content/10.64898/2026.11.05.764629v1.full.pdf",
  "pmcid": null,
  "pmid": null,
  "published": "2026-11-05",
  "publisher": null,
  "ris": "https://research.draconitic.com/paper/1afy9721ymdh1.ris",
  "sources": [
    {
      "first_seen": "2026-11-07T00:21:29Z",
      "native_id": "10.64898/2026.11.05.764629",
      "source": "biorxiv",
      "url": "https://www.biorxiv.org/content/10.64898/2026.11.05.764629v1"
    }
  ],
  "subjects": [
    {
      "field": "biology",
      "origin": "biorxiv:neuroscience",
      "subfield": "neuroscience"
    }
  ],
  "title": "Probing EEG Foundation Model Attention Maps for Lightweight Interpretable Emotion Recognition",
  "title_text": "Probing EEG Foundation Model Attention Maps for Lightweight Interpretable Emotion Recognition",
  "type": "preprint",
  "updated_at": "2026-11-08T00:21:11Z",
  "url": "https://research.draconitic.com/paper/1afy9721ymdh1",
  "venue": "bioRxiv",
  "versions": []
}