Do all roads lead to Rome? A comparison of brain networks derived from inter-subject volumetric and metabolic covariance and moment-to-moment hemodynamic correlations in old individuals

  • Xin Di
  • Suril Gohel
  • Andre Thielcke
  • Hans F. Wehrl
  • Bharat B. Biswal
  • The Alzheimer’s Disease Neuroimaging Initiative
Original Article

DOI: 10.1007/s00429-017-1438-7

Cite this article as:
Di, X., Gohel, S., Thielcke, A. et al. Brain Struct Funct (2017). doi:10.1007/s00429-017-1438-7

Abstract

Relationships between spatially remote brain regions in human have typically been estimated by moment-to-moment correlations of blood-oxygen-level dependent signals in resting-state using functional MRI (fMRI). Recently, studies using subject-to-subject covariance of anatomical volumes, cortical thickness, and metabolic activity are becoming increasingly popular. However, question remains on whether these measures reflect the same inter-region connectivity and brain network organizations. In the current study, we systematically analyzed inter-subject volumetric covariance from anatomical MRI images, metabolic covariance from fluorodeoxyglucose positron emission tomography images from 193 healthy subjects, and resting-state moment-to-moment correlations from fMRI images of a subset of 44 subjects. The correlation matrices calculated from the three methods were found to be minimally correlated, with higher correlation in the range of 0.31, as well as limited proportion of overlapping connections. The volumetric network showed the highest global efficiency and lowest mean clustering coefficient, leaning toward random-like network, while the metabolic and resting-state networks conveyed properties more resembling small-world networks. Community structures of the volumetric and metabolic networks did not reflect known functional organizations, which could be observed in resting-state network. The current results suggested that inter-subject volumetric and metabolic covariance do not necessarily reflect the inter-regional relationships and network organizations as resting-state correlations, thus calling for cautions on interpreting results of inter-subject covariance networks.

Keywords

Brain network Brain connectivity PET Gray matter volume Resting-state 

Funding information

Funder NameGrant NumberFunding Note
National Institutes of Health
  • R01 DA038895
  • R01 AG032088
BMBF
  • DLR-01GQ1415
DFG
  • WE 5795/2-1

Copyright information

© Springer-Verlag Berlin Heidelberg 2017

Authors and Affiliations

  • Xin Di
    • 1
  • Suril Gohel
    • 1
  • Andre Thielcke
    • 2
  • Hans F. Wehrl
    • 2
  • Bharat B. Biswal
    • 1
  • The Alzheimer’s Disease Neuroimaging Initiative
  1. 1.Department of Biomedical EngineeringNew Jersey Institute of TechnologyNewarkUSA
  2. 2.Department of Preclinical Imaging and Radiopharmacy, Werner Siemens Imaging CenterEberhard Karls University of TuebingenTübingenGermany

Personalised recommendations