• Computational neuroscience
• Complex systems modeling
• Parkinson's disease

This thesis aims to simulate a multiscale brain model able to describe
the peculiar changes in brain dynamics which characterize Parkinson’s disease (PD),
the most common basal ganglia (BG) related movement disorder and to test possible DBS configurations.
My goal was to simulate a model which combines
different scales of details. A previously validated data-driven spiking
network model for BG has been placed in a whole-brain framework that uses a
mean-field approach specifically tuned from neuroimaging data of some PD patients.
The results have been analyzed by comparing the most relevant state
variables which characterize the network dynamics between the PD case and the
healthy case (HC). My results from the resting-state simulations are coherent with
important findings from literature, especially the well-known hypotheses regarding
PD patients’ decreased activity level in thalamic region is identified as the
major cause of some relevant PD symptoms (bradykinesia or akinesia). Additionally,
the model shows increased neural activity of the subthalamic nucleus
according to the literature and many experimental findings.
Some differences between PD and HC states in the resting
state firing activity also emerged for some cortical regions.
Interestingly, the validated models for PD and HC
are used to explore different stimuli injection configurations which are assumed to
simulate different kinds of brain perturbations for PD patients, clinically used to
attenuate PD symptomatology. Starting from a subject-specific model, a method to
evaluate the stimuli locations. The potential of this work is the possibility to
evaluate stimulation placement and configuration before surgery and to improve the
settings for individual patients.