Compression of functional gradients during early Non-REM sleep

Sleep profoundly alters large-scale brain dynamics, yet its impact on the cortex’s macroscale functional hierarchy remains incompletely understood. We analyzed transitions from wakefulness to NREM1 and NREM2 sleep in a simultaneous EEG–fMRI dataset of 33 healthy adults, deriving cortical gradients and local and global functional connectivity density (lFCD, gFCD) from stage-specific fMRI time series. Cortical gradients remained highly correlated across states (R > 0.92), indicating preservation of the overall hierarchical scaffold, yet both principal and secondary gradients showed progressive compression during NREM2, with prominent posterior/visual involvement. lFCD increased within unimodal regions and closely tracked gradient position, whereas gFCD showed broader, less spatially specific increases, consistent with altered low-frequency synchronization and network integration during early sleep. Complementary analysis of an independent eyes-open/eyes-closed dataset revealed milder, more localized gradient changes during eyes-closed wakefulness, suggesting that early NREM sleep may involve broader sensory disengagement than eye closure alone. Together, these results indicate that early NREM sleep preserves large-scale cortical topology while reducing differentiation along its principal axes and strengthening local connectivity within sensory systems, marking a shift toward more internally dominated cortical dynamics as the brain transitions away from external input.