The Flow has not summarised this story yet — read it at arXiv Machine Learning.
arXiv:2606. 08594v1 Announce Type: new Abstract: Deep learning EEG denoising architectures have scaled from tens of thousands to tens of millions of parameters, yet no prior study has isolated model capacity as the experimental variable or tested whether reconstruction metrics predict downstream neural-signal utility.
BRIDGE-EEG is an efficient multi‑task EEG classification pipeline that leverages self‑supervised pretraining while dramatically reducing model size. It maps heterogeneous EEG recordings to a unified 62‑channel time‑frequency representation, pretrains an SE‑ResNet18 teacher with SimCLR, and distills it into smaller SE‑ResNet8 and SE‑ResNet4 students. The compact models achieve accuracy comparable to or better than larger foundation models on abnormality detection and emotion recognition, and they consume up to three times less energy on edge devices, enabling deployment on wearable hardware.
arXiv:2607. 03094v1 Announce Type: new Abstract: Electroencephalography (EEG) decoding for brain-computer interfaces (BCIs) faces a major challenge: substantial inter-subject variability limits effective cross-subject generalization.
The study investigates whether Low‑Rank Adaptation (LoRA) can adapt three pretrained EEG foundation models—LaBraM‑base, REVE‑base, and REVE‑large—for binary left‑ vs. right‑hand motor imagery decoding in stroke patients. Using subject‑wise five‑fold cross‑validation on the PhysioNet EEG Motor Movement/Imagery Dataset and a binary subset of the UET175 stroke dataset, LoRA significantly improved accuracy for LaBraM‑base (0.822) and REVE‑base (0.957) on the healthy cohort, but only REVE‑base LoRA achieved high performance (0.847±0.194) on stroke data, with a best mean accuracy of 0.952 in leave‑one‑subject‑out evaluation. The results demonstrate that healthy‑benchmark performance does not guarantee transfer to stroke EEG, highlighting the need for target‑domain adaptation and subject‑level assessment in rehabilitation BCIs.
arXiv:2606. 24394v1 Announce Type: cross Abstract: Electroencephalography (EEG) is the dominant non-invasive modality for brain-computer interfaces (BCIs), yet reliable decoding of motor imagery is hampered by inter- and intra-individual variability.
arXiv:2607. 22733v1 Announce Type: cross Abstract: We investigate whether a generative model can supply useful synthetic motor-imagery (MI) electroencephalography (EEG) trials that improve the accuracy of independent downstream classifiers.