Mapping effective connectivity by virtually perturbing a surrogate brain

TL;DR

This paper introduces Neural Perturbational Inference (NPI), a novel, data-driven framework that uses a surrogate neural network to map whole-brain effective connectivity non-invasively, capturing directionality and causality in brain interactions.

Contribution

NPI is the first framework to infer whole-brain effective connectivity by perturbing a neural surrogate, overcoming limitations of traditional invasive and limited spatial coverage methods.

Findings

NPI outperforms Granger causality and DCM in validation tests.

NPI reveals consistent, structurally supported EC in resting-state fMRI.

NPI's inferred EC correlates significantly with real stimulation pathways.

Abstract

Effective connectivity (EC), indicative of the causal interactions between brain regions, is fundamental to understanding information processing in the brain. Traditional approaches, which infer EC from neural responses to stimulations, are not suited for mapping whole-brain EC in humans due to being invasive and having limited spatial coverage of stimulations. To address this gap, we present Neural Perturbational Inference (NPI), a data-driven framework designed to map EC across the entire brain. NPI employs an artificial neural network trained to learn large-scale neural dynamics as a computational surrogate of the brain. NPI maps EC by perturbing each region of the surrogate brain and observing the resulting responses in all other regions. NPI captures the directionality, strength, and excitatory/inhibitory properties of brain-wide EC. Our validation of NPI, using models having ground-truth EC, shows its superiority over Granger causality and dynamic causal modeling. Applying NPI to resting-state fMRI data from diverse datasets reveals consistent and structurally supported EC. Further validation using a cortico-cortical evoked potentials dataset reveals a significant correlation between NPI-inferred EC and real stimulation propagation pathways. By transitioning from correlational to causal understandings of brain functionality, NPI marks a stride in decoding the brain's functional architecture and facilitating both neuroscience research and clinical applications.