From Monoamines to Systems Psychiatry: Rewiring Depression Science and Care (1960s–2025)

Major depressive disorder was, until recently, framed as a single entity arising from a linear stress–monoamine–hypothalamic–pituitary–adrenal (HPA) axis cascade, modeled by forced swim and learned helplessness tests and evaluated by brief Hamilton Depression Rating Scale (HAM-D)/Montgomery–Åsberg Depression Rating Scale (MADRS) symptom trials. This “unitary cascade” view has been dismantled by imaging, immune–metabolic and sleep profiling, and plasticity markers, which reveal divergent circuit, inflammatory, and chronobiological patterns across anxious, pain-laden, and cognitively weighted depression, with non-response and relapse common. Translationally, face-valid rodent assays that equated immobility with despair have yielded limited bedside benefit, whereas cross-species bridges—electroencephalography (EEG) motifs, rapid eye movement (REM) architecture, effort-based reward tasks, and inflammatory/metabolic panels—are beginning to provide mechanistically grounded, clinically actionable readouts. In current practice, depression care is shifting toward systems psychiatry: inflammation-high and metabolic-high archetypes, anhedonia- and circadian-dominant subgroups, formal treatment-resistant depression (TRD) staging, connectivity-guided neuromodulation, esketamine, selected pharmacogenomic panels, and early digital phenotyping, as endpoints broaden to functioning and durability. A central gap is that heterogeneity is acknowledged but rarely built into trial design or implementation. This perspective advances a plasticity-centered systems psychiatry in which a testable prediction is that manipulating defined prefrontal–striatal and prefrontal–limbic circuits in sex-balanced, chronic-stress models will reproduce human network-defined biotypes and treatment response, and proposes hybrid effectiveness–implementation platforms that embed immune–metabolic and sleep panels, circuit-sensitive tasks, and digital monitoring under a shared, preregistered data standard.