The Flow has not summarised this story yet — read it at Hugging Face Trending Papers.
arXiv:2609.24324v1 Announce Type: new Abstract: Electroencephalography (EEG) provides a non-invasive window into dynamic brain activity, yet modeling long-horizon EEG sequences remains challenging du...
arXiv:2607. 21402v1 Announce Type: new Abstract: Self-supervised foundation models have recently shown strong potential for electroencephalogram (EEG)-based analysis.
arXiv:2608.29304v1 Announce Type: new Abstract: Translating electroencephalography (EEG) into functional magnetic resonance imaging (fMRI) is important for medical neuroimaging, clinical brain-state...
NeuroDETR is a novel multiscale discrete spatiotemporal tokenization framework designed for asynchronous EEG event detection. It combines structured tokenization, event-aware feature refinement, and a sparse attention module called FOCAL to capture both brief and sustained neural responses while preserving spatial hierarchy across EEG channels. Experiments on eight diverse EEG datasets show that NeuroDETR achieves state‑of‑the‑art performance in jointly classifying events and estimating their onsets and offsets.
The paper investigates how EEG embeddings from signal processing and foundation models encode trait and state information. Using a hierarchy metric and kNN purity on a BCI dataset, it finds that most embeddings form a macro‑trait hierarchy with micro‑states inside, which hampers state discrimination. Normalizing embeddings reduces trait leakage and improves state classification accuracy by about 10%, and fine‑tuning shifts the hierarchy toward macro‑state organization, enhancing decoding performance.
BrainCodes is a lightweight EEG foundation model that uses a quantization-based neural tokenizer and a spatiotemporal encoder to learn transferable EEG features from over 800 hours of data across more than 1,000 participants. The model is pretrained with a self‑supervised objective that encourages spatial connectivity and sensitivity to narrowband neural oscillations, enabling it to match or surpass state‑of‑the‑art performance on tasks such as sleep staging, attention assessment, and cognitive load estimation. It achieves these results while using up to 97% fewer trainable parameters, making it suitable for low‑resource wearable devices, and its evaluation framework analyzes the geometry and invariances of the learned representations.