Bipolar disorder presents with coupled disturbances of affective regulation, cognition, and psychomotor activity, yet neuroimaging and stimulation studies commonly examine these dimensions separately. A clinically useful circuit account must identify neural elements that recur across dimensions, distinguish depression from mania, and remain stable when influential studies are removed. A study-level bipolar circuit evidence matrix was constructed from 32 investigations comprising 2,614 reported participant records: 10 emotion-connectivity cohorts, 12 cognition-network cohorts, and 10 non-invasive stimulation trials. Regional terms were harmonized into ten circuit elements. Each study received a design coefficient and a logarithmic sample contribution. A cross-domain control-priority score combined evidential prevalence, domain reach, stimulation alignment, and mood-phase specificity. Signed phase loadings quantified depression–mania polarity. Stability was tested through 32 leave-one-study-out runs and three complete stream-deletion runs. The dorsolateral prefrontal cortex (dlPFC) ranked first with a normalized priority of 100 and remained among the leading three elements in every perturbation. The amygdala ranked second (50.31) and retained a leading-three position in 97.1% of perturbations. Salience and sensorimotor networks followed (32.56 and 32.04), whereas default-mode-network priority (31.43) was strongly dependent on the cognition stream. Eight of ten stimulation trials reported symptomatic or cognitive improvement; among depression trials, six of eight were positive. Phase coding separated depression-dominant internal-network loading from mania-dominant sensorimotor loading, while prefrontal–amygdalar coupling showed large opposing values in both phases. The research question is answered by a hierarchical result: bipolar circuit evidence converges most reliably on a prefrontal–amygdalar control axis, with salience, default-mode, and sensorimotor systems determining phase expression. The dlPFC is therefore the most defensible cross-domain perturbation target, but phase-sensitive network state should guide when and how it is engaged.
Bipolar disorder presents with coupled disturbances of affective regulation, cognition, and psychomotor activity, yet neuroimaging and stimulation studies commonly examine these dimensions separately. A clinically useful circuit account must identify neural elements that recur across dimensions, distinguish depression from mania, and remain stable when influential studies are removed. A study-level bipolar circuit evidence matrix was constructed from 32 investigations comprising 2,614 reported participant records: 10 emotion-connectivity cohorts, 12 cognition-network cohorts, and 10 non-invasive stimulation trials. Regional terms were harmonized into ten circuit elements. Each study received a design coefficient and a logarithmic sample contribution. A cross-domain control-priority score combined evidential prevalence, domain reach, stimulation alignment, and mood-phase specificity. Signed phase loadings quantified depression–mania polarity. Stability was tested through 32 leave-one-study-out runs and three complete stream-deletion runs. The dorsolateral prefrontal cortex (dlPFC) ranked first with a normalized priority of 100 and remained among the leading three elements in every perturbation. The amygdala ranked second (50.31) and retained a leading-three position in 97.1% of perturbations. Salience and sensorimotor networks followed (32.56 and 32.04), whereas default-mode-network priority (31.43) was strongly dependent on the cognition stream. Eight of ten stimulation trials reported symptomatic or cognitive improvement; among depression trials, six of eight were positive. Phase coding separated depression-dominant internal-network loading from mania-dominant sensorimotor loading, while prefrontal–amygdalar coupling showed large opposing values in both phases. The research question is answered by a hierarchical result: bipolar circuit evidence converges most reliably on a prefrontal–amygdalar control axis, with salience, default-mode, and sensorimotor systems determining phase expression. The dlPFC is therefore the most defensible cross-domain perturbation target, but phase-sensitive network state should guide when and how it is engaged.
