Send to

Choose Destination
Brain Connect. 2017 Dec;7(10):627-634. doi: 10.1089/brain.2017.0545.

Detecting Perfusion Pattern Based on the Background Low-Frequency Fluctuation in Resting-State Functional Magnetic Resonance Imaging Data and Its Influence on Resting-State Networks: An Iterative Postprocessing Approach.

Author information

1 MR Collaboration, Siemens Healthcare China , Beijing, China .
2 Department of Psychiatry (UPK), University of Basel , Basel, Switzerland .
3 Department of Mental Health and Psychiatry, University Hospitals of Geneva , Geneva, Switzerland .
4 Faculty of Medicine, University of Geneva , Geneva, Switzerland .
5 Affidea Carouge Radiologic Diagnostic Center , Geneva, Switzerland .
6 Department of Surgical Sciences, Radiology, Uppsala University , Uppsala, Sweden .
7 Department of Neuroradiology, University Hospital Freiburg , Freiburg im Breisgau, Germany .


Resting-state functional magnetic resonance imaging (RS-fMRI) is based on the assumption that the vascular response and the blood oxygenation level-dependent response are homogenous across the entire brain. However, this a priori hypothesis is not consistent with the well-known variability of cerebral vascular territories. To explore whether the RS networks are influenced by varied vascular speed in different vascular territories, we assessed the time-shift maps that give an estimate of the local timing of the vascular response and checked whether local differences in this timing have an impact on the estimates of RS networks. Two hundred seventeen elderly (≥60 years), healthy participants (73.74 ± 4.41 years, 143 females, 203 right handed) underwent one MRI examination, including an RS-fMRI session. After preprocessing, statistical analyses included time-shift analyses and RS-fMRI analyses using as regressor the delay maps obtained from the time-shift analyses. The functional connectivity map of default mode network (DMN) of each participant was then calculated by using the seed-to-voxel analysis in the REST toolbox. Faster cerebrovascular responses were notably present in the primary motor and somatosensory and peri-insular cortex, while slower responses were present in various regions, including notably the posterior cingulate cortex (PCC). Moreover, significant changes notably in the DMN, including medial prefrontal cortex (t = 11.95), PCC (t = 11.52), right middle temporal lobe (t = 10.72), and right angular gyrus (t = 10.88), were observed also taking into account the cerebrovascular delayed maps. As the most prominent example of the RS networks, DMN activation patterns change as a function of the cerebrovascular delay. These data suggest that a group correction for vascular maps in RS-fMRI measurements is essential to correctly depict functional differences and exclude potential confounding effects, notably in the elderly with increasing prevalence of vascular comorbidity.


cerebral vascular territories; default mode network; iterative algorithm; resting-state functional connectivity magnetic resonance imaging; time-shift mapping

[Indexed for MEDLINE]

Supplemental Content

Full text links

Icon for Atypon
Loading ...
Support Center