AI
What masking geometry works best for EEG foundation models?
Publication: Hugging Face Daily Papers
Author: Not available in collected metadata
Source published: 26/09/2026, 20:00:00 UTC
Briefing created: 29/09/2026, 18:44:20 UTC
Based on a feed excerpt; full article content has not been acquired. Source evidence retrieved 29/09/2026, 08:18:24 UTC.
Briefings
This study formalizes spatio-temporal masking strategies for EEG foundation models by training 58 models across two SSL frameworks (MAE and JEPA) on 12 OpenEEGBench datasets. Both frameworks agree on an optimal masking configuration and shared failure modes, while performance remains robust outside these configurations. A novel JEPA-specific failure mode, tagged bias-inflation collapse, was identified that standard detectors miss. The pipeline achieves REVE-level downstream performance with significantly reduced compute compared to pre-training.
Key points
- The study formalizes spatio-temporal masking strategies for EEG foundation models by training 58 models across two SSL frameworks (MAE and JEPA) on 12 OpenEEGBench datasets.
- Both frameworks agree on an optimal masking configuration and shared failure modes, while performance remains robust outside these configurations.
Context and limitations
Results are limited to two SSL frameworks (MAE and JEPA) and 12 datasets from OpenEEGBench; no causal claims or causal mechanisms are supported.
Generated by Bittrees-hosted qwen35-2b sha256:aaf42c8b7c3c on 29/09/2026, 18:44:20 UTC. Based on publisher-provided feed evidence; may not cover the full original article. Check the source for details.
View archived briefing on IPFS
bafybeief7kzkl4rlpaxzpklnr2vhb3fjdc3lrzhrgcgbhxiqwwt5ijgqca