Divergent Excitability of GABAergic Neurons Derived from Bipolar Disorder Patients Shapes Energy Shifts of Network Dynamics, possibly mimicking mania and depression

Bipolar disorder (BD) is characterized by fluctuating mood states, yet the cellular and circuit-level mechanisms distinguishing lithium-responsive (LR) from non-responsive (NR) patients remain elusive. We derived dentate gyrus granule neurons and GABAergic interneurons from induced pluripotent stem cells (iPSCs) of BD patients stratified by lithium response, and from healthy controls. Using patch-clamp electrophysiology, we assessed intrinsic excitability. We further developed a computational model simulating large-scale neuronal networks based on patient-derived electrophysiological properties and ion channel conductance distributions. Granule neurons from both LR and NR patients exhibited hyperexcitability compared to controls. However, GABAergic neurons showed a striking divergence: LR neurons were hyperexcitable, while NR neurons were hypoexcitable. Transcriptomic profiling revealed distinct molecular signatures between NR and LR neurons, including dysregulation of GABA receptor genes (GABRR1). Computational simulations over 10,000 iterations, mimicking long-term network activity, revealed that dentate granule neuron changes alone failed to recapitulate netwrok shifts between global hyperexcitability and global hypoexcitability. Networks incorporating only granule neuron phenotypes entered persistent hyper- or hypoactive states, depending on lithium response. Remarkably, when GABAergic neuron phenotypes were added to the model, both LR and NR networks exhibited spontaneous transitions between high and low activity states, that may be associated with the mood episodes that the patients exhibit. Control networks did not show such bistability. We, therefore, conclude that GABAergic neuronal excitability is a key determinant of lithium responsiveness in BD and critically shapes the emergence of state-shift dynamics in neural networks. These findings suggest that restoring excitatory/inhibitory (E/I) balance via targeted modulation of interneuron function may offer novel therapeutic avenues for BD.