In the present study, we utilized resting-state alpha-band EEG data to examine the effects of sleep deprivation and recovery sleep by comparing the differences among NS, SD and RS sessions. The alpha-band activation of SD sessions decreased over a wide range of cortical regions compared with that of NS sessions, especially in the precuneus, posterior cingulate cortex, cingulate gyrus, and paracentral lobule. Compared with NS sessions, the alpha-band functional connectivity of SD sessions decreased, with the precuneus and posterior cingulate cortex as the most critical nodes. In addition, there was a trend toward increased alpha-band activation and functional connectivity in RS sessions compared with SD sessions.
This study showed decreased alpha-band power in SD sessions compared with NS sessions, which was consistent with previous research [13, 39, 40]. Evidence has shown that there is a negative correlation between alpha power and subjective sleepiness , which was supported by our results. The association between alpha power and sleepiness seems to be global, indicating that the attention and working memory involved in alpha-band oscillations may be related globally to sleepiness [12, 15].
In this study, the brain regions involved in decreased activation included the cingulate gyrus, precuneus, paracentral lobule, and posterior cingulate cortex (BAs 31/7/5/23/30), which are among the most often reported active regions after sleep deprivation in many fMRI studies [41, 42]. Therefore, these cortices may play an important role in maintaining wakefulness. In particular, it was documented that the precuneus and posterior cingulate cortex play a pivotal role in regulating the internal activities of the DMN [43, 44]. Perturbations of DMN activity during wakefulness have been identified in many diseases accompanied by abnormal sleep, such as schizophrenia  and anxiety disorders , which may demonstrate that sleep modulates the DMN and maintains its function.
Furthermore, SD sessions showed reduced widespread functional connectivity compared with that of NS sessions. This result is in line with those of fMRI findings of sleep deprivation [19, 47-49]. In addition, the results were also supported by previous studies that investigated functional connectivity in diseases with sleep abnormalities. Fingelkurts et al. reported that compared to control subjects, depression patients showed a desynchronization of the alpha band, mainly in the right anterior and left posterior brain areas .
Moreover, it is noteworthy that the functional connectivity network changed after sleep deprivation and was mainly distributed in the limbic and parietal cortex. These regions were found to be related to cognitive functions such as semantic processing  and attention  as well as working memory . The reduced functional connectivity of these areas in the current results may indicate that these cognitive abilities are affected by sleep deprivation. The present analysis revealed that the precuneus and posterior cingulate cortex make the greatest contribution to the network, which are considered to be the pivotal area of the DMN and play an important role in mediating intrinsic activities . Considering structural and functional connectivity [54, 55], our results suggested that the precuneus and posterior cingulate cortex are neural hubs damaged by sleep deprivation.
After a night of recovery sleep following sleep deprivation, alpha-band activation and functional connectivity did not return to normal levels, indicating that one night of sleep recovery cannot eliminate the damage caused by 36 of sleep deprivation. In general, sleep has a recovery and organizing effect on the cortical activity of wakefulness [56, 57]. Although the sleep recovery effect was not significant in our results, the difference between RS and SD sessions was smaller than the difference between NS and SD sessions, thus confirming the homeostatic regulation of sleep to a certain extent .
The current alpha-band power spectrum results are consistent with the source localization results, which show that alpha-band power is decreased at both the scalp level and the source level after sleep deprivation. Similarly, the results of source location and functional connectivity are consistent, indicating that sleep deprivation greatly influences the DMN to which the precuneus and posterior cingulate cortex belong. Altogether, our complementary results showed that after sleep deprivation, the simultaneous decrease in cortical activation and connectivity weakened local processing and brain region cooperative processing. Based on the high coincidence of the alpha-band activation source and alpha-band functional connections of key node positions and the positive coupling of activation power and connectivity strength after sleep deprivation and recovery sleep, we cannot exclude the possibility that the alpha-band connectivities between the DMN and other brain regions may be modulated by oscillation power. This seems to be consistent with the idea that nerve synchronization influences functional integration [27, 59], indicating that power fluctuations in DMN alpha-band oscillations lead to cortical interaction changes.
Notably, there are several limitations to this study. First, no control group was set up to eliminate the possible influence of circadian rhythm changes on EEG recordings. The three EEG acquisition sessions in this study did not occur at the same time of day, and EEG data may be potentially affected by participants’ circadian rhythms. Fortunately, studies have confirmed that EEG changes caused by are hardly affected by circadian rhythms [60, 61]. Second, the spatial resolution of the source localization and connectivity analysis was not very high. The spatial resolution of EEG sources increases with the number of electrodes, so high electrode density recording is more reliable in EEG rhythm source analysis. The use of a standard MRI template instead of individual MRIs for source localization further decreases the possible spatial resolution. Third, all of our subjects were men, so the results should be extrapolated to women with some caution. Fourth , the current study examined only EEG changes in the alpha band caused by sleep deprivation and recovery sleep, while possible changes in other frequency bands were not taken into account.