EEG patterns reveal internal dynamics of sleep stage N3
Achim Schilling1, Patrick Krauss1, Konstantin Tziridis1, Maximilian Traxdorf2, Holger Schulze1; 1Experimental Otolaryngology University of Erlangen-Nuremberg, 2Department of Otolaryngology University of Erlangen-Nuremberg
Classic visual sleep stage scoring is based on EEG frequency band analysis of 30s epochs and is commonly performed by highly trained medical sleep specialists using additional information from submental EMG and eye movements (EOG). We have recently provided the proof-of-principle in 40 subjects that sleep stages can be consistently differentiated solely on the basis of spatial 3-channel EEG patterns based on root-mean-square (RMS) amplitudes and without using classic EEG frequency band analysis or graphic elements such as K-complexes, spindles, vertex waves and posterior occipital sharp transients. We visualize the spatial cortical activity patterns based on RMS amplitude vectors using multidimensional scaling. It turns out that these patterns naturally divide into different clusters that correspond to visually scored sleep stages. As could be demonstrated, the clusters are highly reproducible between different subjects. Especially the cluster associated with the REM sleep stage seems to be extremely robust and very different from the one associated with the wake state. Furthermore, we found that the sleep stage N3 shows a characteristic dynamic behavior that is consistent across different occurrences during sleep within the same subject and also across different subjects: root-mean-square amplitudes first increase monotonically, then stay constant and finally decrease monotonically again, with each period lasting for several minutes. Interestingly, the described dynamics first start in occipital areas, then spreads to pre-frontal areas and finally reach the central sulcus, with a temporal delay in the order of magnitude of minutes.
Topic Area: METHODS: Electrophysiology