From concept to computation On modeling psychological phenomena

Psychological disorders are inherently complex, arising from an interplay of biological, psychological, and environmental factors that evolve over time. The network approach to psychopathology acknowledges this complexity. Rather than conceptualizing psychological disorders as arising from a single cause, this approach conceptualizes them as emerging from dynamic causal relations among symptoms and related factors.
This dissertation explores two methods of addressing this complexity within network theory. The first part takes a data-driven approach and focuses on the methodological challenges of statistical network modeling. Network models reveal the structure of relationships among many variables and investigate how symptoms and external factors relate to each other. The second part takes a theory-driven approach, using computational models to formalize explanatory principles based on network theory and simulate resulting behaviors.
Though these approaches are presented separately in the dissertation, they are inextricably linked. Robust empirical findings from statistical network models constrain the explanatory principles formalized in computational models and serve as benchmarks for evaluating them. Formal models make explicit which empirical findings a theory can explain and generate new predictions that subsequent empirical (network) studies can test. Together, these approaches foster an iterative cycle between theory and empirical research, which is crucial for advancing our understanding of psychological phenomena.