NREM sleep stages specifically alter dynamical integration of large-scale brain networks

Abstract

<jats:title>Abstract</jats:title><jats:p>Functional dissociations in the brain observed during non-rapid eye movement (NREM) sleep have been mainly associated with reduced information integration and impaired consciousness that accompany increasing sleep depth. Most of the studies that evaluated this relation have mostly been focused on <jats:italic>spatial</jats:italic> alterations of brain networks across various vigilance states. Here, we explored the <jats:italic>dynamical</jats:italic> properties of large-scale functional brain networks derived from <jats:italic>transients</jats:italic> or moments of activity changes in fMRI using two complementary studies of simultaneous EEG-fMRI recordings of sleep. Our results revealed distinct alterations in the individual network’s dynamical characteristics across wakefulness and NREM sleep. Positive activations of visual-sensory areas simultaneously deactivate focal subcortical regions known to be involved with arousal regulation. The cerebellum is also found to dissociate into posterior and anterior regions, with the former being predominantly present during wakefulness than in the deep sleep. Most surprisingly, we found that global network activity and pair-wise network interactions increased significantly in NREM stage 2 before an abrupt loss of communication is observed in NREM stage 3. Thus, by providing a spatiotemporal and more accurate network-level representation of brain organization, we were able to capture new features of information integration of consciousness during sleep, and provide concrete evidence for the presence of unstable yet distributed global synchronization in NREM stage 2.</jats:p>

Post-stroke reorganization of transient brain activity characterizes deficits and recovery of cognitive functions

“…The number of clusters was determined using evidence accumulation and, as a result, 16 iCAPs were extracted ( Figure S3a ). The iCAPs corresponded to well-known resting-state networks, in line with our previous results in different datasets ( Figure 1b and Table S2 ) (23–25). Specifically, they included sensory-related networks such as primary and secondary visual areas (iCAP 6), auditory and language network (iCAP 5), with high activations in Heschl gyrus and rolandic operculum, and sensorimotor network (iCAPs 3 and 8), which included pre-, post-, and para-central areas, supplementary motor area, and middle cingulum (only for iCAP 3).…”

“…In particular, networks were bilateral regardless of the side of the lesion, suggesting a preserved organization across brain regions. While this might seem in contradiction with the decrease in interhemispheric connectivity previously reported post-stroke (18, 35–37), it should be noted that fine-grained coordination between brain areas can be occulted when using classical static FC approaches (13, 25). For instance, the correlation between distinct brain areas exhibiting both positive and negative correlations in distinct time-windows would result in a low static functional connectivity, in spite of the dynamic synchronization existing between these regions (25).…”

“…There was an additional task-related iCAP, the fronto-parietal network (iCAP 4), with high activations in the frontal and parietal lobes together with anterior and middle cingulate cortex (ACC and MCC). As previously reported (23–25), the Default Mode Network (DMN) was decomposed into a frontal part (iCAP 12) including frontal lobe and ACC, and a posterior part (iCAP 10) comprising the cuneus, the precuneus, and the superior and inferior parietal lobes. Four iCAPs included regions of the cerebellum (iCAPs 2, 7, 13, and 16).…”

“…Therefore, it follows that a decrease in FC does not necessarily imply a decrease in global network activity. However, decreased FC might reflect an overall reduction in the efficiency of information transfer, which could manifest in the form of aberrant durations of functional networks (12, 25, 31).…”

Dynamic Auditory Remapping Across the Sleep-Wake Cycle

“…5 and 6). The altered auditory map structure presented in N3 might be explained by findings from whole-brain fMRI studies showing a global decrease in effective interactions and breakdown of inter-modular connectivity during deep sleep (Horovitz et al, 2009; Jobst et al, 2017; Tagliazucchi et al, 2013; Tarun et al, 2021). Furthermore, the distinct auditory neural organisation during N3, may help to explain the decline in sensory and cognitive processes that is typically found during deep sleep (Andrillon and Kouider, 2020; Andrillon et al, 2016, 2017; Hennevin et al, 2007; Legendre et al, 2019; Peigneux et al, 2001).…”