Umeå University's logo

Longitudinal studies have become more common in the past years due to their superiority over cross-sectional samples. In light of the ongoing replication crisis, the factors that may introduce variability in resting-state networks have been widely debated. This publication aimed to address the potential sources of variability, namely, time of day, sex, and age, in longitudinal studies within individual resting-state fMRI data. DCM was used to analyze the fMRI time series, extracting EC connectivity measures and parameters that define the BOLD signal. In addition, a two-way ANOVA was used to assess the change in EC and parameters that define the BOLD signal between data collection waves. The results indicate that time of day and gender have significant model evidence for the parameters that define the BOLD signal but not EC. From the ANOVA analysis, findings indicate that there was a significant change in the two nodes of the DMN and their connections with the fronto-parietal network. Overall, these findings suggest that in addition to age and gender, which are commonly accounted for in the fMRI data collection, studies should note the time of day, possibly treating it as a covariate in longitudinal samples.