• autism spectrum disorder
• brain network model
• computational neuroscience
• digital twin
• personalized connectivity

Autism Spectrum Disorder (ASD) is an extremely heterogeneous neurodevelopmental
disorder on various levels (genetic, neurobiological, cognitive) and the underlying
causes are still open to debate. Given this characterizing heterogeneity, the
construction of a personalized model of the brain provides the optimal approach
both for investigating underlying causes of ASD and for the design of
personalized therapies.
In this thesis, I implemented this personalized approach simulating the dynamics of
cortical regions with a brain network model through the neuroinformatics platform The
Virtual Brain. I coupled mesoscopic models of neural masses representing different
areas of the cortex through a network of connections reconstructed from the anatomy
of one pediatric ASD subject starting from MRI data.
I was able to reconstruct the spectral properties of
each area and the functional connectivity between them, which will allow validation of
the model via EEG recordings.
In parallel, a quantitative analysis of the empirical data revealed
significant differences in both activity and connectivity of the ASD subject compared to
one control pediatric patient. This allowed me to investigate the role that dynamical
parameters andì local and global coupling play in determining ASD-related cortical
activity.
This work marks a major advance in brain modeling by introducing
methodological novelties and paving the way to personalized approaches for
neurodevelopmental diseases.