Individual Discriminative Ability of Resting Functional Brain Connectivity is Susceptible to the Time Span of MRI Scans. (1st February 2022)

• Record Type: Journal Article
• Authors: Ni, Huangjing; Song, Ming; Qin, Jiaolong; Jiang, Tianzi
• Is Part Of: Neuroscience. Volume 482(2022)
• Page Start: 43
• Page End: 52
• Publication Date: 2022-02-01
• Subjects: BN Brainnetome site -- DMN default mode network -- FPN fronto-parietal network -- HCP Human Connectome Project -- kNN k nearest neighbor -- LOSO leave-one-subject-out -- MCFS multi-cluster feature selection -- NMI normalized mutual information -- ROI region of interest -- RSFC resting-state functional connectivity -- rs-fMRI resting state functional magnetic resonance imaging -- UESTC University of Electronic Science and Technology of China; individual identification -- brain connection -- individual differences -- imaging time span; Neurochemistry -- Periodicals; Neurophysiology -- Periodicals; Neurology -- Periodicals; Neurochimie -- Périodiques; Neurophysiologie -- Périodiques; Neurochemistry; Neurophysiology; Electronic journals; Periodicals; Electronic journals; 612.8
• DOI: 10.1016/j.neuroscience.2021.12.008
• Languages: English
• ISSNs: 0306-4522

Abstract

Highlights: Individual discriminative accuracies of resting functional brain connectivity decline with the increase of time spans. Imaging time spans are explored ranging from within one day to across three years. The most discriminative brain connectivities from one day to across three years remains in the DMN and FPN. Abstract: Recent studies have suggested that resting-state brain functional connectivity (RSFC) has the potential to discriminate among individuals in a population. These studies mostly utilized a pattern of RSFC obtained from one scan to identify a given individual from the set of patterns obtained from the second scan. However, it remains unclear whether the discriminative ability would change with the extension of the time span between the two brain scans. This study explores the variations in the discriminative ability of RSFC on eight time spans, including 6 hours, 12 hours, 1 day, 1 month, 3–6 months, 7–12 months, 1–2 years and 2–3 years. We first searched for a set of the most discriminative RSFCs using the data of 200 healthy adult subjects from the Human Connectome Project dataset, and we then utilized this set of RSFCs to identify individuals from a population. The variations in the discriminative accuracies over different time spans were evaluated on datasets from a total of 682 unseen adult subjects acquired from four different sites. We found that although the accuracies were detectable at above-chance levels, the discriminative accuracies showed a significant decrease ( F = 17.87, p < 0.01) along with the extension of brain imaging time span, from over 90% within one month to 66% at 2–3 years. Furthermore, the decreasing trend was robust and not dependent on the training set or analysis method. Therefore, we suggest that the discriminative ability of RSFC in identifying individuals should be susceptible to the length of time between brain scans.