ASL has demonstrated the capability to quantify global fluctuations and resting state network fluctuations in cerebral blood flow (CBF) (1). ASL global signal fluctuations are larger than the resting state network fluctuations and appear uniformly distributed across gray matter (2). Much of the global CBF fluctuations may reflect globally correlated neural activity that would suggest a global resting network. However, physiological noise, such as cardiac and respiratory motion, could potentially contribute to the global signal fluctuations. In this work, we systematically assess the potential sources of ASL global signal fluctuation by monitoring multiple aspects of the cardiac and respiratory signals during repeated ASL measurements and evaluating their contributions to global fluctuations.Ten volunteers were scanned on a GE 3T scanner using the body coil for transmission and an 8 channel head coil for reception. Resting state pCASL(3) was performed using a single shot stack-of-spiral RARE acquisition, with resolution 5.3mmx5.3mmx4mm. Labeling duration was 2s and the post labeling delay was 1.8s and TR was 5s. A baseline ASL scan was first performed to provide a coil phase map and to quantify perfusion for each subject. A background suppression design optimized for just 0.3% background signal was employed. With background suppression, both control and label scans are sensitive to CBF, so control and label scans were performed and processed separately (20min each). Vessel suppression and inferior blood saturation were used to eliminate the impact of cardiac supply. Images were reconstructed with the phase map from the traditional ASL baseline scan, then normalized by individual perfusion level and transformed to standard Montreal Neurological Institute (MNI) space using SPM12. The global resting state signal was extracted as the mean of the same 3D brain mask from each temporal frame. To assess the ASL global signal fluctuation associated with cardiac and respiratory effects, the global signal fluctuation was fit to a second-order Fourier series of cardiac phase (4), cardiac rate (5), respiratory phase and respiratory volume (6). Fitting power was calculated by the percentage of global signal variation explained by each potential contributor. After removing significant cardiac and respiratory contributions, the data were separated into 15 components using independent component analysis. No significant correlations between global ASL signal fluctuations and cardiac and respiratory parameters (fitting power <0.15) (cardiac phase: 0.054±0.028, cardiac rate: 0.041±0.043, respiratory phase: 0.046±0.014, respiratory volume: 0.063±0.074). This excludes these systemic factors as significant contributors to the ASL global signal fluctuation. The amplitudes of fluctuation in the identified networks were reduced after regressing out the global fluctuation signals (control 6.57%, label 7.02%) which can be confirmed with the visual appearance the fluctuation maps (Fig. 1). After regressing out the global signal, the relative fluctuations (to the mean perfusion) of all recognized 8 networks were greatly reduced in both control and label scans (Table 1) and the correlations between networks were significantly reduced (p<0.0001) (Fig. 2). This indicates that the global signal fluctuations are the dominant resting fluctuations of the ASL signal. These data suggest a separate globally correlated resting state network [7] in addition to those region-specific resting state networks (e.g. default mode network). Because background suppressed ASL is robust to many sources of global fluctuation that can appear in BOLD, it may be uniquely suited to the study of this globally correlated functional signal.