Characterizing systemic physiological effects on the blood oxygen level dependent signal of resting-state fMRI in...

Systemic physiological dynamics, such as heart rate variability (HRV) and respiration volume per time (RVT), are known to account for significant variance in the blood oxygen level dependent (BOLD) signal of resting-state functional magnetic resonance imaging (rsfMRI). However, synchrony between these cardiorespiratory changes and the BOLD signal could be due to neuronal (i.e., autonomic activity inducing changes in heart rate and respiration) or vascular (i.e., cardiorespiratory activity facilitating hemodynamic changes and thus the BOLD signal) effects and the contributions of these effects may differ spatially, temporally, and spectrally. In this study, we characterize these brain-body dynamics using a wavelet analysis in rapidly sampled rsfMRI data with simultaneous pulse oximetry and respiratory monitoring of the Human Connectome Project. Our time-frequency analysis across resting-state networks (RSNs) revealed differences in the coherence of the BOLD signal and heartbeat interval (HBI)/RVT dynamics across frequencies, with unique profiles per network. Somatomotor (SMN), visual (VN), and salience (VAN) networks demonstrated the greatest synchrony with both systemic physiological signals when compared to other networks; however, significant coherence was observed in all RSNs regardless of direct autonomic involvement. Our phase analysis revealed distinct frequency profiles of percentage of time with significant coherence between BOLD and systemic physiological signals for different phase offsets across RSNs, suggesting that the phase offset and temporal order of signals varies by frequency. Lastly, our analysis of temporal variability of coherence provides insight on potential influence of autonomic state on brain-body communication. Overall, the novel wavelet analysis enables an efficient characterization of the dynamic relationship between cardiorespiratory activity and the BOLD signal in spatial, temporal, and spectral dimensions to inform our understanding of autonomic states and improve our interpretation of the BOLD signal.