Introduction: In the light of the ongoing replication crisis in the field of neuroimaging, it is necessary to assess the possible exogenous and endogenous factors that may affect functional magnetic resonance imaging (fMRI). The current project investigated time-of-day effects in the spontaneous fluctuations (<0.1 Hz) of the blood oxygenation level dependent (BOLD) signal. Method: Using data from the human connectome project release S1200, cross-spectral density dynamic causal modeling (DCM) was used to analyze time-dependent effects on the hemodynamic response and effective connectivity parameters. The DCM analysis covered three networks, namely the default mode network, the central executive network, and the saliency network. Hierarchical group-parametric empirical Bayes (PEB) was used to test varying design-matrices against the time-of-day model. Results: Hierarchical group-PEB found no support for changes in effective connectivity, whereas the hemodynamic parameters exhibited a significant time-of-day dependent effect, indicating a diurnal vascular effect that might affect the measured BOLD signal in the absence of any diurnal variations of the underlying neuronal activations and effective connectivity. Conclusion: We conclude that these findings urge the need to account for the time of data acquisition in future MRI studies and suggest that time-of-day dependent metabolic variations contribute to reduced reliability in resting-state fMRI studies. Impact statement The results from this study suggest that the circadian mechanism influences the blood oxygenation level dependent signal in resting-state functional magnetic resonance imaging (fMRI). The current study urges to record and report the time of fMRI scan acquisition in future research, as it may increase the replicability of findings. Both exploratory and clinical studies would benefit by incorporating this small change in fMRI protocol, which to date has been often overlooked.

Findings from neuroimaging studies show low replication rates, but the origin of the problem is not clear. The current project explored the extent of the time-of-day dependent metabolic variation to the replication crisis. Using data from the human connectome project (HCP) release S1200, cross-spectral density dynamic causal modelling was used to analyse time-dependent effects on the hemodynamic response and effective connectivity parameters in the spontaneous fluctuations (>0.1Hz) of the blood oxygenation level dependent (BOLD) signal. Hierarchical group-parametric empirical Bayes found no support for time-of-day variations in effective connectivity, whereas the hemodynamic parameters exhibited a significant time-of-day dependent effect. It is concluded that these findings urge the need to account for the time of day in data acquisition of future MRI studies. Moreover, data from the human connectome project suggest that the relationship between functional and dynamic connectivity and the BOLD signal might not be direct.

No abstract