Causal flow in brain activity: present and future of directed network inference.
Relazione su invito
V - Biofisica e fisica medica
Aula 32C-2 - Martedì 18 h 09:00 - 13:00
» Download abstract
Estimating the directionality of whole-brain connectivity from resting-state functional MRI (rs-fMRI) is an important and largely unaddressed issue. As opposed to heavily parametrized model comparison procedures such as dynamic causal modeling (DCM), Multivariate Granger causality (MVGC) based methods are considered to be statistically more intuitive, easier to implement and rely on less assumptions. However, MVGC applications in rs-fMRI are still controversial, mostly due to latency differences in neurovascular coupling across different brain regions, low-sampling rates, and noise confounds. With massive in-silico simulations of neuronal populations we propose a validation benchwork to test the performance of MVGC and to understand the confidence of the results when applying MVGC in-vivo.