Cross-sectional and longitudinal network analysis of bipolar depression and mania during early pharmacological...

This study examines the structure and early dynamic changes of symptom networks in patients with bipolar disorder (BD) during manic and depressive states under pharmacological treatment. We aim to explore how symptom networks initially evolve over four weeks with medication.

A total of 206 BD patients in the manic phase and 106 in the depressive phase were included. Symptom severity was assessed using the Bech-Rafaelsen Mania Rating Scale (BRMS), Hamilton Depression Rating Scale (HAMD), and Hamilton Anxiety Rating Scale (HAMA). Cross-sectional and longitudinal symptom networks were constructed based on baseline and four-week follow-up data. For manic-phase patients, network comparison tests (NCT) were performed to examine differences in baseline network characteristics between responders and non-responders to identify potential predictors of treatment response, as well as between male and female subgroups.

During the four-week medication period, the BRMS symptom network exhibited initial sparsification of core manic symptom clusters and intensification of the irritability cluster. The depressive phase symptom network showed a continuous co-activation mechanism of somatic symptom clusters, with depressed mood (Q6) consistently acting as a key bridging node. NCT results showed no significant difference in global network strength between the responding and non-responding groups of the BRMS baseline network (P = 0.07).

After four weeks of medication, associations between core symptom clusters generally weakened. In the BRMS symptom network, intensive treatment targeting the irritability cluster can be prioritized early. In the HAMA symptom network, depressed mood (Q6) remains a key intervention target during early treatment.