Abstract
Functional magnetic resonance imaging has revealed correlated activities in brain regions even in the absence of a task. Initial studies assumed this resting-state functional connectivity (FC) to be stationary in nature, but recent studies have modeled these activities as a dynamic network. Dynamic spatiotemporal models better model the brain activities, but are computationally more involved. A comparison of static and dynamic FCs was made to quantitatively study their efficacies in identifying intrinsic individual connectivity patterns using data from the Human Connectome Project. Results show that the intrinsic individual brain connectivity pattern can be used as a ‘fingerprint’ to distinguish among and identify subjects and is more accurately captured with partial correlation and assuming static FC. It was also seen that the intrinsic individual brain connectivity patterns were invariant over a few months. Additionally, biological sex identification was successfully performed using the intrinsic individual connectivity patterns, and group averages of male and female FC matrices. Edge consistency, edge variability and differential power measures were used to identify the major resting-state networks involved in identifying subjects and their sex.
Highlights
A common assumption made in many past studies is that the functional connectivity (FC) does not change over the data acquisition time period the brain wanders through a state of connectivities and the FC is non-stationary in nature
The subject misidentified by partial static functional connectivity (sFC) was the same individual over trials, and this individual was misidentified across methods
The FC matrices were for data acquired several months apart
Summary
A common assumption made in many past studies is that the FC does not change over the data acquisition time period the brain wanders through a state of connectivities and the FC is non-stationary in nature. The most commonly used approach is that of a windowed analysis[5] in which the repeated states[6] are determined using some clustering algorithms[7] While these methods are more involved and better model the dynamic brain activity, it is not clear how advantageous or what new information can be gleaned by considering the short-term, defined here to be in the order of a few minutes, non-stationarities in the correlation values or dynamic functional connectivity www.nature.com/scientificreports/. Studies are needed to determine any medium- to long-term changes in the intrinsic individual brain connectivity due to aging or other environmental factors. The commonly used sFC and emerging dFC methods were quantitatively compared in this study by evaluating the accuracy of identifying subjects and their sex using their intrinsic individual brain connectivity patterns or ‘fingerprint’. Results are presented that show the intrinsic individual connectivity patterns do not significantly change in the medium term and brain connectivity ‘fingerprinting’ is possible even with data acquired several months apart
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
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 call
