Neurological diseases are primarily defined by their clinical phenotype, with imaging traditionally serving to identify specific features and support diagnosis. In this branch of our research, we aim to shift the paradigm toward imaging-centered phenotyping, advancing the understanding of neurological disorders such as autoimmune encephalitis and developmental and epileptic encephalopathies..
From neuroradiological assessment to normative modelling
Imaging-based phenotyping relies on normative modeling, where individual scans are compared to large population-based datasets. Deviations in imaging features, such as cortical thickness, metabolic profiles, and microstructural markers, define an individual’s unique imaging phenotype.
Comparison against normative cohorts
Deep phenotyping
In larger groups of individuals with neurological or neurodevelopmental disorders, pattern recognition algorithms are used to identify distinct subgroups based on shared imaging phenotypes. By analyzing these imaging patterns alongside clinical data, we can uncover common and distinguishing clinical features, thereby expanding imaging phenotypes to clinical phenotypes.
Deep Phenotyping
Vision
We envision a future where imaging becomes a cornerstone of data-driven diagnosis and treatment, complementing the clinical heuristics that guide today’s medical care.