Shao-Jun Xia, Jing Bao, Anlan Sun +4 more
Sleep spindles are short bursts of oscillation that appear in EEG during sleep and carry real physiological meaning. Detecting them is awkward because they are low-amplitude events occupying a tiny fraction of a long recording, so most of what a detector sees is not a spindle.
Previous methods struggle in two specific ways the authors call out: pinning down exactly where a spindle starts and ends, and coping when several occur close together. Both follow from treating detection as classification of fixed windows, which forces a continuous event into a grid it does not respect. A spindle that straddles two windows, or two spindles inside one, are the awkward cases.
SpindleFlexNet reframes the task so boundaries and counts are what the model predicts, rather than by-products of a windowing choice made before the data was seen.
Sleep spindle is a physiologically significant biomedical signal in electroencephalographic (EEG) waveforms, which is typically a low-amplitude event in sleep. Due to the small signal ratio in the overall EEG, previous detection methods have limited capability to capture its start and end points and lack flexibility in handling multi-spindle scenarios. To address the gap, we address the problem from a new perspective and introduce SpindleFlexNet, the first framework in this…
DiffEEG: A Self-Supervised Denoising Diffusion Model for Learning EEG Generic Representations
arXiv (BCI) · July 13, 2026Learning Residual Kinematic Corrections for Continuous Neural Decoding via Reinforcement Learning
arXiv (neural decoding) · July 13, 2026Event-based Neural Decoding for Neuroprosthetic Motor Control
arXiv (EEG) · July 12, 2026Graph Bispectrum for Nonlinear Mode Interactions