Circadian rhythm of sleep and wakefulness is a fundamental property of human physiology, which is important for sustaining essential body functions. It has been reported that the physiology of human brain is rhythmic genetically and systematically (Smale et al., 2008). The suprachiasmatic nuclei (SCN), located in a small region of the hypothalamus, is considered to be the primary circadian clock in mammals (Hastings et al., 2003; Saper et al., 2005). SCN activity fluctuates on a daily basis (Hofman and Swaab, 1993). Understanding system properties intrinsic to the SCN, as well as its regulation of peripheral oscillators throughout the brain and body will provide a fundamental contribution to our understanding of the neurophysiological basis of human circadian behavior.

The fluctuating activities of SCN and peripheral oscillators, in concert with the learning and experience over the course of a day may influence neurons activity and the way brain regions communicate with each other. Resting-state functional magnetic resonance imaging (RS-fMRI) technique can investigate brain activity by measuring the variance in the spontaneous fluctuations of the blood oxygen level dependent (BOLD) signal. Based on RS-fMRI, regional homogeneity (ReHo) and amplitude of low frequency fluctuation (ALFF) methods have been used to analyze spontaneous low-frequency (<0.08 Hz) blood oxygenation level-dependent fluctuations of the brain (Zang et al., 2004; Yang et al., 2007). ReHo measures the coherence of low frequency fluctuation within a given area, which can be used to assess the synchrony of neural activity based on the hypothesis that brain activities would more likely occur in clusters than in a single voxel. ALFF is associated with field potential activity of local brain regions (Logothetis et al., 2001) that could be employed to evaluate the intensity of intrinsic or spontaneous neuronal activity of the brain (Yu-Feng et al., 2007). Therefore, ReHo and ALFF can serve as sensitive markers indicating the alteration of brain function. For example, ReHo decreases in the posterior cingulate cortex/precuneus (PCC/PCu) in the AD patients (He et al., 2007), ALFF increases in the anterior portion of the dorsal anterior cingulate cortex in subthreshold depression patients (Li et al., 2014).

A recent research investigated if the brain networks are stable during a 24-h period by acquiring RS-fMRI data from 12 young adults at 3-h intervals over 24 consecutive hours. The authors found resting state sub-network dynamicity, especially the dynamic fluctuations of default mode network (DMN) connectivity and speculated that connectivity may have different spatiotemporal properties on the neural level (Park et al., 2012). Marek et al. (2010) also reported that the orienting and executive attention neuronal networks present time-of-day related variations with a stroop-like task performed five times a day. In our previous work, we have found a diurnal variation pattern of white matter microstructure which may function as the substrates of the phasic neural activities in correspondence to the environment adaptation in a light-dark cycle (Jiang et al., 2014). These analyses of the spontaneous neural activities in response to the diurnal rhythm may provide a new insight toward better understanding of the brain physiology and sleep medicine. Therefore, we aim to investigate the dynamic characteristic of neural activity by evaluating the intra-individual variability of ReHo and ALFF over one light-dark cycle.

The current study demonstrated the diurnal rhythmic dynamic of the neural activity of human brain. Circadian rhythms are widespread in human brain and influenced by the interactions between internal molecular systems and environmental cues (Cirelli, 2009; Mohawk et al., 2012). Core clock genes are known to comprise transcriptional-translational auto-regulatory complexes. These genes make up a group of auto-regulatory loops with diurnal variation of activity and present rhythmic expression of their own and their regulatory target transcrips. Li et al. (2013) reported that there exists a rhythmic rise and fall in the transcriptional activity of hundreds of genes [including three Period homolog (PER 1-2-3) genes; brain and muscle Arnt-like protein-1 (BMSL1) and so on] initiating or responding to the regulation of 24-h behavioral and hormonal cycles. Vicentic et al. (2005) also found that CART (cocaine- and amphetamine-regulated transcript) peptides exhibit a diurnal rhythm in several brain regions including nucleus accumbens, hypothalamus and amygdala. Muto et al. (2016) investigated the local modulation of human brain responses by circadian rhythmicity and sleep debt. They found that the response of some subcortical regions like thalamus, head of caudate nucleus and putamen follows a 24 h circadian rhythmicity, and some brain regions were affected significantly by sleep debt, and some are affected by the interaction of the two factors (Muto et al., 2016). The circadian rhythmicity may substrate the physiological mechanisms of the diurnal variation of ReHo and amplitude fluctuation of neural activity occurred in some of the structures mentioned above in the current study.

The DMN is supposed to be involved in self-referential functions such as internal mentation, recollection and imagination and conceptual processing, containing a set of interacting brain areas that are functionally connected. It can be divided into sub-networks contributing to specific tasks or intrinsic cognitive demand (Mayer et al., 2010; van Buuren et al., 2010). The current study demonstrated that overall the DMN is involved in greater neural activity at resting state in both morning and the evening as compared to the other brain areas, which is consistent with previous findings (Raichle et al., 2001; Fransson, 2005). But some of the structures were found to be more subject to the diurnal alterations at the neural activity level, such as MPFC, medial temporal lobe, PCu and PCC. The divergence of the diurnal variation in these structures may suggest the functional coordination of the anatomically independent gyri. It has been reported that perturbations in DMN activity during wakefulness display co-occurring abnormalities of sleep in a number of disorders (Cherkassky et al., 2006; Garrity et al., 2007; Zhao et al., 2007). The DMN functional connectivity was found to be reduced in sleep deprivation (De Havas et al., 2012). The PCu and PCC, which demonstrated increased activity in the PM group, play a pivotal role in the mediation of intrinsic activity through DMN (Fransson and Marrelec, 2008). In the contrast, the activity of medial temporal lobe decreased in the PM group. Considered in the context of these findings, it’s reasonable to hypothesize that diurnal changes in DMN activity reflect adaptation or compensation response under continued wakeful condition. It may also imply that there is a balance of neural activity within DMN subregions, or a decoupling of regulation between these regions.

The occipital lobe shows significant decrease in ReHo and ALFF map in the PM group than that in the AM group. These findings are contrary to our initial assumption. The occipital lobe is mainly related to the visual information processing. The photic stimulus duration has an effect of “temporal integration” on neurons (Dkhissi-Benyahya et al., 2000), which means the neural activity should increase in the PM group. Also, decreased ADC (apparent diffusion coefficient) mainly occurred in the occipital lobe in the PM group in our previous study on diurnal variation of white matter using diffusion tensor imaging (DTI), which is believed to associate with the increased neuron activation (Jiang et al., 2014). But Buysse Daniel et al. (2004) measured relative glucose metabolism in the morning and in the evening with PET, and found significantly lower glucose metabolism in the evening than in the morning in occipital lobe, including cuneus, medial occipital gyrus, lingual gyrus, and occipitotemporal gyrus, which was interpreted to reflect increasing homeostatic sleep drive and/or the restorative effect of sleep. Muto et al. (2016) reported that the neural response of occipital gyrus are modulated by the interaction of circadian rhythmicity and sleep debt. These investigations suggest that sleep debt may have dominant control in the occipital lobe in the evening under current experiment condition.

Precentral and postcentral gyri have higher homogeneous neural activity and low frequency fluctuation in the morning than in the evening. This may suggest that during the evening hours, the somatosensory and motor functions represent a weakened response to the afferent and/or efferent signals. The exact biological mechanisms behind ReHo and ALFF remain unclear up to date. ReHo was defined as the temporal homogeneity of the regional BOLD signal (Zang et al., 2004), while ALFF reflects the amplitude of low-frequency fluctuation in the range of 0.01–0.08 Hz (Yu-Feng et al., 2007). Logothetis and Wandell (2004) reviewed that the hemodynamic response primarily reflects the neuronal input to the relevant area of the brain and synaptic potentials are the strongest cause of the BOLD response. The lower ReHo and ALFF in precentral and postcentral gyri found in this study may indicate that the synchronization of neural activity in these areas is disrupted and the amplitude of low-frequency fluctuation of BOLD signal decreases, possibly corresponding to the deteriorated somatosensory and motor response to the external stimulus and weakened local field potential as a result of increased level of exhaustiveness, sleepiness, and absentmindedness after a day experience.

The prefrontal cortex (including medial- and inferior-orbital frontal gyrus, and anterior cingulate) shows greater cerebral BOLD response in the PM group than in the AM group. These regions are reported to be involved in alertness, attention and higher-order cognition processes. However, Thomas et al. (2000) found that brain activity and function decrease in prefrontal cortex and thalamus after 24 h of sleep deprivation using positron emission tomography (PET), which implies that the need for recuperation in these areas is greater than other brain regions. On the contrary, Drummond and Brown (2001) reported that prefrontal cortex and parietal lobes showed increased activation in verbal learning and divided attention tasks after total sleep deprivation using fMRI, which may suggest a specific compensatory mechanism designed to combat increased sleepiness. Increased evening activity in the prefrontal cortex was also identified in this study. We hypothesize that the increased sleepiness or fatigue after a day of activity may induce neurochemical consequences for the brain to adapt to fulfillment the demand for the alertness as well as the higher cognitive performance as dark cycle approaches.

It should be noted that there are several limitations in this study. For example, the resting state fMRI signal is partially affected by cardiac pulsation, respiration, vasoconstriction, instrumental and thermal sources of noise. We didn’t record the physiological parameters mentioned above during the fMRI data acquisition. During the data preprocessing procedure, a temporal band-pass filter was used to reduce low frequency drift and high frequency physiological noise. This may alleviate, although not completely eliminate the effect on accuracy and reliability for the current study.

Sex related difference in cerebral structure and function was not taken into account in this study, which may confound the interpretation of the results. Evidences have been accumulating that sex disparity matters in the circadian modulation of mammal brain (Mong et al., 2011). Women are found to have shorter intrinsic periods and earlier phase of melatonin rhythms as well as greater night-time impairment in cognitive performance as compared to men (Duffy et al., 2011; Santhi et al., 2016), but the intrinsic circadian periods of both gender are close to 24 h. In addition, healthy people have different preferences in sleep pattern such as morning “lark” or evening “owl” chronotypes. This may also embed complicated mechanisms of rhythmic dynamics of human brain that this study did not reach.

Lastly, the fMRI data were acquired from only 16 young subjects at two time points. The sample size was relatively small. This can’t and limit us in providing more significant findings. A larger sample size covering multiple age groups and time points are expected to provide stronger statistical power with reduced individual effects and enable a more detailed tracking of neural activities during the time course of a day–night cycle.

Further study is needed to optimize the experiment design, like recruitment of single gender subjects, melatonin assays to assess circadian phase, performing different cognitive tasks that are differentially affected by sleep drive and circadian rhythmicity.

Neural activity of healthy young human brain exhibits a diurnal pattern which may be related to circadian plasticity or accumulated waking, which may substrate the neurobiological and behavioral mechanisms of sleep hygiene and sleep disorders. Factors of time of day should be taken into account in the functional neuroimaging based neural studies even under resting conditions.

Conceived and designed the experiments: LZ and CJ. Performed the experiments: CJ, LY, and LZ. Analyzed the data: CJ, LY, SS, CS, XL, GX, and LZ. Wrote the paper: CJ and LZ.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.