Introduction:

In humans but also in rodents, resting state functional magnetic resonance imaging (rs-fMRI) is used to investigate brain function at rest. It is known that the activity of most animal species varies on a 24-h basis controlled by the suprachiasmatic nucleus of the hypothalamus. Recent study has shown that there is a modulation of neuronal activity due to circadian rhythm ¹. However, to our knowledge there is no previous work investigating the impact of circadian rhythm on functional brain connectome in mice. Here, we examined whether functional connectivity changes during the day-night cycle.

Methods:

MRI acquisition: this study was approved by our institutional committee on Ethics (authorization #26779). Nine female C57Bl6/J mice (14 ± 1 months) were studied: 5 in Night Condition (NC) and 4 in Light Condition (LC). Anesthesia was induced with isoflurane (3-4%) followed by s.c. injection of Medetor (0.13mg/kg). The right flank was shaved to facilitate monitoring of the oxygen saturation during rs-fMRI with an MR compatible mouse oximeter. Animals were scanned on a Pharmascan 70/16 US 7T Brucker system using Paravison 6 software. Animals in the nocturnal phase were anesthetized and explored in total darkness. Isoflurane was maintained below 0.6% during BOLD rs-fMRI acquisitions and body temperature between 33-35°C to prevent awakening. Rs-fMRI was acquired using a gradient echo EPI sequence: TE/TR/FA=19.91ms/2500ms/45°; 40 interleaved axial slices, 0.145×0.145×0.4 mm³ spatial resolution, 110 ×110 matrix, 20 dummy scans; 512 repetitions, 22 min total scan duration.

rs-fMRI preprocessing: the data were converted to nifty using MIPAV script and all the preprocessing was performed using FSL’s recommended preprocessing pipeline from FMRIB’s Software Library (FSL version 6). Preprocessing included motion correction, outliers regression using frame displacement (FD) metric ², registration to native space, non-brain removal using Brain Extraction Tool (BET), bias field correction; slice-timing correction using Fourier-space time-series phase-shifting, grand-mean intensity normalization of the entire 4D dataset by a single multiplicative factor and high pass temporal filtering (0.01Hz). The filtered data were regressed according to the changes in intensity using the default FSL DVARS ² metric threshold. Then, global signal regression and low pass filtering (0.1Hz) were applied and data were registered to Allen template using FLIRT followed by nonlinear (FNIRT) warping. Spatial smoothing by a Gaussian kernel and an isotropic resampling resolution of 4 mm were applied on the preprocessed data during the multi-session temporal concatenation Independent Component Analysis (ICA) analysis using Probabilistic ICA ³ as implemented in MELODIC (Multivariate Exploratory Linear Decomposition into Independent Components) Version 3.15, part of FSL. The following data pre-processing was applied to the input data and a total of 40 independent components (IC maps) were extracted based on a previous work ⁴. We used dual regression (DR) (v0.6) for between-subject analysis allowing voxel-wise comparisons of rs-fMRI data ^5,6. Non-parametric permutation based inference analysis ⁷ was performed with subject-specific component spatial maps concatenated across subjects and submitted to voxel-wise between-subject analysis testing for effect of night and day cycle using FSL-randomise ⁸. FSL’s general linear model (GLM) was used to define design contrasts based on unpaired t-test, testing circadian condition effect among the two groups. For each analysis we ran 5000 randomized permutations as recommended, while threshold-free cluster enhancement (TFCE) ⁹ was used for statistical inference to validate the likelihood of extended areas of signal, which also considers information from neighboring voxels. Correction for multiple comparisons across space was applied assuming an overall significance of p < 0.05 using permutation testing and TFCE.

Results:

Of the 40 components derived from group-ICA, 15 were identified as plausible resting state networks according to the networks reported in the literature ^4,10,11 after discarding components at the brain surface and those involving vascular structures and ventricles. Among the interesting ICA components, one appeared statistically different between the two groups. The cluster was significantly increased in the group in Night Condition (figure 1).

Discussion:

Figure 1 shows the only significant difference found in NC for pineal gland which is in accordance with its role in the circadian rhythm by the nighttime secretion of melatonin ^12,13. One caveat of this study is that the animals were disrupted during the diurnal phase (rest phase) because of the MRI exploration. Another limitation is the small number of exclusively female animals. The analysis of the effect of the circadian rhythm on both males and females at different ages is under way.

Conclusion:

Our study showed that functional connectivity increases in the pineal gland during night condition. Our ongoing studies will include more animals and more time points for a better coverage of the day and night cycle, which could lead to the identification of other networks depending on the circadian rhythm.