Frequency-Dependent Connectivity in Large-Scale Resting-State Brain
Networks during Sleep
Abstract
Functional connectivity (FC) during sleep has been shown to break down
as non-rapid eye movement (NREM) sleep deepens before returning to a
state closer to wakefulness during REM sleep. However, the specific
spatial and temporal signatures of these fluctuations in connectivity
patterns remain poorly understood. The goal of this study was to
investigate how frequency-dependent network-level FC fluctuates during
nocturnal sleep in healthy young adults using high-density
electroencephalography (hdEEG). Specifically, we examined
source-localized FC in resting-state networks during NREM2, NREM3, and
REM sleep in the first three sleep cycles of 29 participants. Our
results showed that FC within and between all resting-state networks
decreased from NREM2 to NREM3 sleep in multiple frequency bands and in
all sleep cycles. The data also highlighted a complex modulation of
connectivity patterns during the transition to REM sleep whereby delta
and sigma bands hosted a persistence of the connectivity breakdown in
all networks, whereas a reconnection was observed in the default mode
(DMN) and the attentional networks in frequency bands characterizing
their organization during wake (i.e., alpha and beta bands,
respectively). Finally, all network pairs (except the visual network)
showed higher gamma-band FC during REM sleep in cycle three compared to
earlier cycles during the night. Altogether, our results unravel the
spatial and temporal characteristics of the well-known breakdown in
connectivity observed as NREM sleep deepens. They also shed light on a
complex pattern of connectivity during REM sleep that is consistent with
both breakdown and reconnection processes that are network- and
frequency-specific.