Anesthesia affects brain functional connectivity (Stuart et. al. 2006, Desai et. al. 2009). Yet, the link between neurophysiological effects of anesthetics such as alteration of local/global neural activity and functional networks and its clinical effects, i.e. loss of consciousness is not understood. EEG as readout of neural activity has been widely used for studying underlying mechanisms (Patrick et. al. 2013). However, EEG provides poor spatial resolution and thus does not yield information on anesthesia-induced changes in cerebral functional connectivity. In contrast, resting-state fMRI enables studying functional networks and their modulation by pharmacological interventions, though at inferior temporal resolution. In this study, we evaluated the effects of a commonly used anesthetic, isoflurane, on the brain functional networks in mice using resting-state fMRI. Both static and dynamic functional connectivity analysis have been applied to capture the effect of the varying the anesthetic dose on stationary and transient network patterns.Data acquisition: Resting-state fMRI data sets of anesthetized mice were collected on a Bruker BioSpec 94/30 system operating at 9.4T using a gradient-echo echo-planar imaging (GE-EPI) sequence with 1s temporal resolution. Anesthesia regimen included isoflurane with doses of 1.1%, 1.2% and 1.5% in an air/oxygen mixture (4:1). Animals were mechanically ventilated throughout the experiment. Data analysis: After preprocessing and realignment, concat-ICA was applied using the MELODIC toolbox of FSL followed by dual regression analysis and ‘randomize’. These results were used for Network Modeling (Smith et. al. 2011) of the brain networks using FSLNets. We also calculated the dynamic functional connectivity networks using the dictionary learning approach (Leonardi et. al. 2013, Leonardi et. al. 2014). Time courses were extracted from the ICA maps and fed into the dFC algorithm. Simple building blocks (referred as ‘atoms’) of whole brain connectivity were estimated using dictionary learning algorithm (Leonardi et. al. 2014) with 30 folds each made of 200 iterations. The patterns thus extracted explained more than 50% of the variance in the sliding window correlation matrix. These building blocks were estimated for each anesthetic dose separately.Dual regression analysis revealed static functional connectivity between homotopic regions in the two hemispheres. Network analysis was performed for selected ICs using FSL Nets that revealed significant loss of functional connectivity between the homotopic regions in the two hemispheres with the increased dose of isoflurane (Fig. 1). Increasing the isoflurane dose led to a general decrease in static correlation values all over the brain, though the susceptibility varied across brain regions. This is illustrated for two brain networks, the lateral cortical network (LCN) and the default mode network (DMN), for which an average correlation values is given (averaged over the various ROIs contained in the network, Fig.2). The dynamic functional connectivity analysis revealed significant interactions among functional networks that were not apparent from the conventional analysis (Fig. 3). Louvain algorithm was used for estimating the modular functional connectivity matrix. dFC analysis using dictionary-learning method (Leonardi et. al. 2013) displayed significant interaction between the various modules identified (LCN, DMN, thalamic and limbic networks). Increasing the isoflurane dose led to decreased interactions among these modules (comparison 1.1 and 1.2% isoflurane, Fig. 3). At 1.5% isoflurane spatial segregation is largely lost, i.e. the individual atoms (dynamic functional connectivity states) displayed no structure, i.e. they were rather ‘all off’ or ‘all on’. We might speculate that this loss of spatial segregation might be associated to loss of consciousness as suggested in earlier studies, which observed ‘global’ synchronization across the brain during deep anesthesia (Grandjean et al. 2014). Brain functional networks are dependent on the anesthetic doses. While static FC analysis with increasing dose of anesthetic revealed a steady loss of functional correlation among homotopic brain regions, dynamic FC analysis revealed a loss of spatial segregation across brain functional networks, indicative of a deep anesthetic state. Dynamic functional network analysis allowed the analysis of network changes that were not apparent in the conventional static analysis.