PHINN-EEG uses topological time-series analysis to improve dream-state EEG classification

The PHINN-EEG framework introduces persistent homology-inspired neural network methods for dream mentation analysis, moving beyond traditional power spectral density features. By extracting Dynamic Betti Curves from multichannel pre-awakening EEG epochs via sliding-window Takens delay embeddings and Vietoris-Rips filtrations, the model characterizes the geometric architecture of neural activity.

• PHINN-EEG targets an AUC of 0.82-0.90 on the 1,462-awakening open-access subset of the DREAM database, outperforming existing PSD and catch22 benchmarks.
• The study includes a topology-conditioned rectified flow model for dream-state EEG synthesis, using a spectral-conditioned flow model as an ablation baseline to isolate topological conditioning effects.
• The research proposes candidate Betti transition archetypes linking topology to phenomenological dream report categories as an exploratory hypothesis space.

This work represents a paradigm shift from spectral energy to phase-space geometry in neural rare-event detection, with potential future implications for wearable BCI dream monitoring.