Digital Theses Archive


Tesi etd-06302022-093241

Type of thesis
Quantitative tools for the analysis of neurological and neuropsychiatric pathologies of the basal ganglia
Scientific disciplinary sector
Istituto di Biorobotica - BIOROBOTICS
relatore Dott. FALOTICO, EGIDIO
  • Artificial Neural Networks
  • Attention-Deficit/Hyperactivity Disorder
  • basal ganglia
  • biomarkers
  • computational modelling
  • EEG
  • motor control
  • Parkinson's disease
  • Sleep
  • Slow Waves
  • upper limb movements
Exam session start date
Neurological and neuropsychiatric disorders affect millions of people worldwide, resulting in a healthcare, social and economic burden in our times. On a clinical level, accurate diagnosis and prompt intervention are hampered by the lack of objective and reliable instruments of investigation and analysis. From the scientific perspective, understanding of our brain is challenging because of its inherent complexity and the different levels of computation at which it can operate. Biomedical engineering and computational sciences offer precious help in addressing these issues by providing clinicians with safe and consistent tools to extract quantitative information from the analysis of the Central Nervous System and its pathophysiological processes. Several aspects of the disease treatment have benefited from medical imaging technologies, Machine Learning algorithms and other computational approaches to neuroscience, for instance for extraction of biomarkers of disease and for in silico simulation of dysfunctional brain mechanisms and drug- or stimulation-induced perturbations. <br>In this work, I have investigated quantitative approaches to neurology and neuropsychiatry, by studying, developing and implementing computational tools for the analysis of brain diseases. In particular, the spotlight was turned on the pathologies of the basal ganglia, a complex brain structure that has been extensively studied so far but still represents a mystery object for its influence on many and various aspects of our basic and elaborate functions, both motor and non-motor ones. A first doctoral subproject focused on the neurological disorders of movement. Motor impairments in Parkinson&#39;s disease were studied by reviewing the scientific literature on upper limb aiming tasks. An extensive analysis of these neuroscientific findings allowed us to investigate the potential mechanisms of dysfunction in reaching and grasping that involve the interplay between the basal ganglia and the motor cortices, and thus the role of this interplay in the production of coordinated and effective motor behaviour. This knowledge was then used for devising and developing a computational model of the control and modulation of reaching movements, based on Artificial Neural Networks. This neural model simulated the influence of physiological and pathological conditions on the neural control of a realistic biomechanical model of the upper limb. The role of basal ganglia was modelled both on a functional level, exploiting the optimal feedback control theory, and on a neuroanatomical level, exploiting the known neural connections. A second subproject focused on the neuropsychiatric disorders that also see the engagement of basal ganglia and dopamine. It consisted of a retrospective study on hundreds of young patients diagnosed with Attention-Deficit/Hyperactivity Disorder that aimed at identifying electrophysiological markers of this disease for disentangling its heterogeneous diagnostic complexity. For this purpose, EEG recordings of these patients during sleep were analysed and processed, and EEG features were associated with patients&#39; clinical (diagnostic) scores. Consistent associations on the scalp were found between specific clinical profiles and microstructural characteristics of the sleep EEG, namely the maximum downward slope of slow waves, a parameter mostly unexplored in literature. The identified neurophysiological marker can be considered as a first step toward the construction of unambiguous phenotypes of this pathology, paving the way for a quantitative approach to complex neuropsychiatric diagnosis and patient-specific paths for intervention.