Bipolar disorder (BD) is a highly complex, chronic psychiatric illness marked by episodic disturbances of mood. While historically conceptualized as a disorder of neurotransmitter imbalance, converging evidence suggests it is better understood as a disorder of dynamic dysregulation across multiple interacting systems, including genetic and epigenetic vulnerability, HPA axis dysregulation, mitochondrial/metabolic dysfunction, immune activation, circadian disruption, and gut-brain axis disturbances. These systems comprise a hierarchically organized, bidirectionally coupled feedback network that is responsible for the oscillatory mood states that define the disorder. Here, it's proposed that a specific, testable, integrative model in which genetic/epigenetic vulnerabilities lead to destabilization of HPA axis regulation, which then leads to mitochondrial bioenergetic dysfunction (the central integrative node of the cascade), which then, through its interaction with immune activation and gut-brain signaling, causes the excitation-inhibition imbalance underlying the clinical presentation of mood-state transitions. From this model, three empirically tractable predictions are made: that co-occurring inflammatory/metabolic/circadian dysregulation will constitute a biological subtype with earlier onset, more rapid cycling, and a more negative response to lithium; that neuroendocrine/metabolic disturbances will prospectively precede rather than simply co-occur with mood-state transitions; and that interventions targeting the metabolic-inflammatory interface will produce mood-stabilizing effects whose magnitude correlates with baseline biological severity rather than symptom severity. Throughout, the importance of state-dependence, trait-level features, and medication confounds is discussed. Finally, translational implications for metabolic stratification, circadian-focused intervention, and gut-brain modulation are outlined.
