Spectral features from frequency and time-frequency domains enable traditional ML to match or exceed attention-based DL performance in EEG disease diagnosis on small datasets, with attention unable to capture stable spectral signatures.
The two decades brainclinics research archive for insights in neurophysiology (tdbrain) database
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.LG 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
No citing papers match the current filters.