Abstract
Effective treatment of affective disorders (AD) requires a deep understanding of the underlying neurobiological mechanisms. However, in machine-learning-based analyses, cross-sectional studies have failed to identify robust individual-level biomarkers. Research Domain A of CRC/TRR393 addresses this gap by implementing longitudinal, multimodal studies using real-time mobile assessments. Central to this effort is the identification of "inflection signals"-clinically meaningful symptom changes marking transitions from euthymia to depressive or (hypo)manic episodes. These critical windows are captured through digital phenotyping and ecological momentary assessments and followed up by in-depth neurobiological profiling. Six projects examine the dynamic interplay of behavioral, cognitive-emotional, molecular, immune, and neural mechanisms during these transitions. Project A01 validates early-warning models using digital phenotypes and machine learning. Project A02 maps structural and functional brain changes in relation to disease course and risk factors. Project A03 investigates the role of microglial immune activation in recurrent depression. Project A04 investigates neurobiological alterations after inflection signals using intensive, multimodal data acquisition conducted both in laboratory settings and in the participants' personal environments. Project A05 adds molecular and immunological profiling and integrates findings from human and animal data. Project A06 studies trajectories from bipolar at-risk states to full-blown illness. Together, these projects form the empirical foundation for mechanism-based interventions (Domain C) and theoretical modeling of symptom trajectories (Domain B).