Unravelling cognitive functioning in healthy and multiple sclerosis through the analysis of transiently bursting brain networks at milliseconds time scale.

Project Details


Lately, neuroimaging studies have shifted from ‘activation’ based studies (i.e. which brain region is active when?) to connectivity studies (e.g. does the correlation of fMRI signal between regions predict a certain behaviour?).
This new approach allows identifying large-scale networks describing brain activity patterns. Moreover, the dynamic recruitment of these networks underlies healthy cognitive functioning.
Most of the functional connectivity studies assume a static connectivity pattern, whereas recent research has revealed a temporally richer brain dynamics, where ‘microstates’ of 100–200ms activate and dissolve, the same timescale as cognitive functions evolve. Throughout this project, we will employ the Hidden Markov Model algorithm which allows the extraction of new promising parameters of network dynamics on a sub-second scale and trial level.
These new parameters will allow a novel and more extensive characterization of the brain’s functioning underlying both healthy and impaired cognition.

This project will demonstrate the unique potential of a non-static technique to investigate cognitive functioning in healthy and MS subjects, relying on an extensive and unique data set: a cohort of 100 MS patients and 50 HCs for which MEG, T1/T2/DWI MR and behavioural data have been collected.
Effective start/end date1/11/2031/10/24


  • Characterization of cognitive functioning in Healthy control and Multiple Sclerosis
  • Hidden Markovian Model – transient brain states dynamics
  • MEG neurophysiological measurements

Flemish discipline codes in use since 2023

  • Cognitive neuroscience
  • Computational biomodelling and machine learning
  • Neurophysiology
  • Engineering psychology


Explore the research topics touched on by this project. These labels are generated based on the underlying awards/grants. Together they form a unique fingerprint.