Chronnectome fingerprinting: Identifying individuals and predicting higher cognitive functions using dynamic brain connectivity patterns

Abstract

The human brain is a large, interacting dynamic network, and its architecture of coupling among brain regions varies across time (termed the "chronnectome"). However, very little is known about whether and how the dynamic properties of the chronnectome can characterize individual uniqueness, such as identifying individuals as a "fingerprint" of the brain. Here, we employed multiband resting-state functional magnetic resonance imaging data from the Human Connectome Project (N = 105) and a sliding time-window dynamic network analysis approach to systematically examine individual time-varying properties of the chronnectome. We revealed stable and remarkable individual variability in three dynamic characteristics of brain connectivity (i.e., strength, stability, and variability), which was mainly distributed in three higher order cognitive systems (i.e., default mode, dorsal attention, and fronto-parietal) and in two primary systems (i.e., visual and sensorimotor). Intriguingly, the spatial patterns of these dynamic characteristics of brain connectivity could successfully identify individuals with high accuracy and could further significantly predict individual higher cognitive performance (e.g., fluid intelligence and executive function), which was primarily contributed by the higher order cognitive systems. Together, our findings highlight that the chronnectome captures inherent functional dynamics of individual brain networks and provides implications for individualized characterization of health and disease.

Unraveling reproducible dynamic states of individual brain functional parcellation

“…A proliferation of approaches also exist in the literature to study the dynamics of functional connectivity. These studies confirmed the spatio temporal reconfiguration of the brain networks [24][25][26][27] and associated it to dynamics of cognitive processing or mental states dictated by tasks [28][29][30][31][32] . The co-activation patterns [33,34] , spatial independent component analysis [35] were among the most widely applied techniques to consider brain dynamics (see [36] for a review).…”

“…Chen and colleagues modeled these states switching processes of resting state brain activities using a hidden markov model [33] . Therefore, neuroscientists mentioned there is a need to have neuroimaging tools to identify how brain parcels reconfigure spatially in the case of highly cognitive regions over time and to evaluate the variability of brain parcels across time and across individuals [32] . Even though dynamic functional connectivity is well studied, to the best of our knowledge, the only parcellation approach that considered dynamic changes of parcels was suggested by Salehi and colleagues.…”

Dynamic Functional Connectivity Reveals Abnormal Variability and Hyper‐connected Pattern in Autism Spectrum Disorder

“…Unlike SFC, DFC is a time-varying process. Therefore, we employed two common metrics to describe its dynamic characteristics: DFC-Str (strength) [Kung et al, 2019;Liu, Liao, Xia, & He, 2018] and DFC-SD (standard deviation). DFC-Str is the average of DFC, and DFC-SD is used to examine temporal variability.…”

Test–retest reproducibility of a multi‐atlas automated segmentation tool on multimodality brain MRI

“…This "fingerprint" problem has recently been explored in neuroimaging (Finn et al, 2015;Liu, Liao, Xia, & He, 2018;Mars et al, 2018). This "fingerprint" problem has recently been explored in neuroimaging (Finn et al, 2015;Liu, Liao, Xia, & He, 2018;Mars et al, 2018).…”

“…A problem that cannot be intuitively solved by looking at I2C2 or ICCs is whether the global individual pattern of image features is reproducible. This "fingerprint" problem has recently been explored in neuroimaging (Finn et al, 2015;Liu, Liao, Xia, & He, 2018;Mars et al, 2018). Assuming that the MRI data itself are reproducible (which is a valid assumption for structural data, such as T1-based volumes), if the postprocessing and quantification tool is reliable, a given individual will be closer to him/herself rather than to someone else in the space of the image features.…”