Identify functional connectivity differential graphs in DLB patients: From network based statistics to functional graphical models by Sara Sommariva, Assistant Professor (RTDa) in Statistics at the Department of Mathematics, Università di Genova
When: 17 November 2025 at 3:00 pm
Where: Sala Seminari VIMM (Fondazione Ricerca Biomedica Avanzata, Via Orus 2, Padova)
Abstract: Neurodegenerative diseases such as Dementia with Lewy Bodies (DLB) are heterogeneous diseases characterized by multiple clinical phenotypes. Understanding the relationship between these phenotypes and the underlying brain activity and functional connectivity is a key step for a timely diagnosis and a personalized treatment.
In this talk, I will discuss some recent methodological approaches for comparing brain functional connectivity graphs in groups of patients with different clinical features starting from electroencephalographic (EEG) data.
Specifically, I will first describe a classical approach based on network based statistics which aims at identifying connected subgraphs in coherency-based functional connectivity matrices that show significant between-group differences. In a recent work we showed that this approach was able to highlight an increased connectivity in DLB patients with rem sleep behaviour and in those with visual hallucinations.
Then I will present a more recent approach based on conditional functional graphical regression models for comparing functional connectivity graphs when accounting for multiple covariates.