Neurophysiological brain mapping of human sleep-wake states

Area of Science:

• Neuroscience
• Computational Neuroscience
• Sleep Science

Background:

• Intracranial electroencephalogram (iEEG) provides detailed brain activity insights.
• Previous work established a spectrum-based model for awake iEEG using normative data.
• The normative database has been expanded to include non-rapid eye movement (NREM) and rapid eye movement (REM) sleep data.

Purpose of the Study:

• To extend the spectrum-based iEEG model to encompass NREM and REM sleep data.
• To analyze spectral characteristics of iEEG across all four arousal states: wake, N2, N3, and REM sleep.
• To develop a reduced, parametric model for iEEG across different arousal states.

Main Methods:

• Normalized amplitude spectra from iEEG data across wake and sleep states were averaged region-wise.
• Spectra were fitted to a multi-component Gaussian distribution.
• A reduced model with five key parameters per brain region was developed and visualized on cortical surface models.

Main Results:

• A lognormal Gaussian mixture model accurately described iEEG in all brain regions and arousal states.
• Model parameters showed smooth transitions between sleep and wake states.
• Primary cortical areas (vision, motor, auditory) and the hippocampus did not show 'awakening' activity during REM sleep.

Conclusions:

• A low-dimensional parametric spectral model effectively describes iEEG across wakefulness and all sleep stages.
• Despite time-domain differences, spectral modeling unifies iEEG characteristics across arousal states.
• Continuous spectral variation suggests a 'continuum' hypothesis for sleep-wake generation in the human brain.