Network topography and tissue-depth-like dependence of cerebrovascular reactivity covariance in the human brain

Resting-state functional networks (RSNs) have demonstrated topographical structures that have been widely used to represent interactions within brain circuits. However, recent work demonstrated that RSNs may coexist with vascular networks, especially during task-induced or spontaneous neuronal activity. In this work, we formally interrogate the topography in the inter-individual covariance of cerebrovascular dynamics using breath-holding fMRI. We clarify the nature and scope of such a topography through cerebrovascular reactivity (CVR) amplitude and lag in a group of healthy older adults. We demonstrate cross-participant spatial synchrony in CVR amplitude and lag co-variations that are confined to RSNs. Such an inter-subject covariance topography is not seen in the BOLD signal response time courses. CVR amplitude covariance is also found to vary in a tissue-depth-like manner in both the grey and white matter, being most variable across brains in deep WM and least variable in superficial cortex. Moreover, WM CVR lag appears to be organized by fibre tracts. This work demonstrates the strong presence of vascular network topography independent of neuronal activity, one that is also modulated by regional differences that resemble tissue depth. This work emphasizes the importance of studying cerebrovascular dynamics as a key constituent of functional segregation.