Dynamical blueprint reveals how brain states emerge from excitation-inhibition balance

A new paper from Kateryna Nechyporenko, Peter Ashwin, and Krasimira Tsaneva-Atanasova (University of Exeter) published on arXiv identifies a fundamental geometric structure—an organizing center—that governs how neuronal networks transition between different dynamic states. The brain constantly shifts between active (up) and quiescent (down) states, along with rhythmic oscillations critical for perception, memory, and information processing. Despite their importance, the dynamical principles enabling this repertoire of activity patterns have remained poorly understood. The authors derive the mathematical conditions for this organizing center and show it exists robustly across several canonical models of neuronal population dynamics. Near this center, switches between oscillations, bistability, and up/down states are orchestrated by the network's excitation-inhibition (E-I) balance. Crucially, the study demonstrates that E-I balance doesn't just modulate activity but defines the entire dynamical landscape from which distinct brain states emerge.

Furthermore, the team extended their model to include neuron-astrocyte interactions, revealing that astrocytes—glial cells traditionally seen as support cells—can actively tune the excitatory-inhibitory balance, thereby modulating transitions between different neuronal activity regimes. This suggests a previously underappreciated role for astrocytes in brain state control. The findings present a universal dynamical blueprint underlying brain state organization, with potential implications for understanding neurological disorders where E-I balance is disrupted, such as epilepsy or schizophrenia. For AI researchers, the work could inspire new architectures for neuromorphic computing that mimic these efficient state-switching dynamics.