Brain network dynamics fingerprints are resilient to data heterogeneity

Abstract

The advent of extremely large data repositories of brain activity recordings comes with no free lunch. While multi-site datasets have significantly boosted the quest to unravel the brain's mechanics, inevitable differences in physical parameters and recording protocols may lead to erroneous conclusions. In this study, we investigate from a perspective -the viewpoint of dynamical models -which factors tend to have non-negligible effects on the recordings of hemodynamic signatures of brain activity. To this end, we make use of data-driven models to capture the dynamical wandering of the brain between large-scale networks of activity. We first confirm the emergence of robust, subject-specific dynamical patterns of brain activity in resting-state fMRI data. Next, we exploit these fingerprints to appraise the effect of an array of scanning factors in a multi-site dataset. We find that scanning sessions belonging to different sites and days tend to induce high variability in such fingerprints, while other factors affect the same metrics to a minor extent. These results concurrently indicate that each subject has its own unique trajectory of brain activity changes, but also that our ability to infer such patterns is affected by how, where and when we try to do so.

Functional Control of Oscillator Networks

“…Furthermore, we show that the Kuramoto model represents an accurate and interpretable mapping between the brain anatomical organization and the functional patterns of frequency-synchronized neural co-fluctuations. Once a good mapping is inferred, it can be used to define a generative brain model to replicate in silico how the brain efficiently enacts large-scale integration of information, and to develop personalized intervention schemes for neurological disorders related to abnormal synchronization phenomena [16,36,62].…”