Abstract
IntroductionGraph metrics have been proposed as potential biomarkers for diagnosis in clinical work. However, before it can be applied in a clinical setting, their reproducibility should be evaluated.MethodsThis study systematically investigated the effect of two denoising pipelines and different whole‐brain network constructions on reproducibility of subject‐specific graph measures. We used the multi‐session fMRI dataset from the Brain Genomics Superstruct Project consisting of 69 healthy young adults.ResultsIn binary networks, the test–retest variability for global measures was large at low density irrespective of the denoising strategy or the type of correlation. Weighted networks showed very low test–retest values (and thus a good reproducibility) for global graph measures irrespective of the strategy used. Comparing the test–retest values for different strategies, there were significant main effects of the type of correlation (Pearson correlation vs. partial correlation), the (partial) correlation value (absolute vs. positive vs. negative), and weight calculation (based on the raw (partial) correlation values vs. based on transformed Z‐values). There was also a significant interaction effect between type of correlation and weight calculation. Similarly as for the binary networks, there was no main effect of the denoising pipeline.ConclusionOur results demonstrated that normalized global graph measures based on a weighted network using the absolute (partial) correlation as weight were reproducible. The denoising pipeline and the granularity of the whole‐brain parcellation used to define the nodes were not critical for the reproducibility of normalized graph measures.
Highlights
Graph metrics have been proposed as potential biomarkers for diagnosis in clinical work
This study systematically investigated the effect of two denoising pipelines and different whole-brain network constructions on reproducibility of subjectspecific graph measures
We investigated the use of the two functional connectivity measures, three different approaches to handle negative correlations, two types of network, and two types of calculations of the weights of the network (based on the correlation values or based on a transformation of Z-values obtained after a Fisher r-to-Z transformation)
Summary
Resting-state fMRI (rs-fMRI) is a task-free and an easy-to-use tool for neuroscientific data acquisition. Node definition may affect reproducibility and the way functional connectivity between nodes is calculated (Liang et al, 2012) The latter is typically based on a Pearson correlation and partial correlations can be used since it can remove the influences of other nodes (Marrelec et al, 2006; Smith et al, 2011). Most studies only focused on the robustness of resting-state fMRI based graph measures at the group level (Braun et al, 2012; Du et al, 2015; Paldino, Chu, Chapieski, Golriz, & Zhang, 2017) It can provide interesting information about brain functioning in normal or pathological conditions, it is not sufficient if we want to introduce these techniques in a clinical setting in which subject-specific graphs have to be constructed and quantified (Gordon et al, 2017; Poldrack et al, 2015). The standard preprocessing steps included: (1) The first four dummy scans of each run were removed; (2) all resting-state functional images were realigned to correct for head movement; (3) slice timing; (4) coregistration of the structural and mean functional image; (5) segmentation of the structural image which provides deformation fields to warp the data to MNI space; and (6) warping of the functional images to MNI space using the deformation field of each subject obtained in the previous step
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callDisclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
