Metabolic constraints shape hypersynchronous dynamics in spiking cortical microcircuit models

Neuronal activity depends on ATP-consuming ion homeostasis, yet the circuit-level consequences of impaired energy availability remain difficult to isolate experimentally. Existing seizure models often represent metabolic effects indirectly or at single-cell scale, leaving unclear how cellular energetic stress can alter structured cortical network dynamics. Here, we tested whether coupling intracellular energy availability to neuronal excitability is sufficient to destabilize baseline cortical activity and generate seizure-like synchronization. We extended the Adaptive Exponential Integrate-and-Fire model with a normalized energy variable governed by explicit production and consumption terms, fitted layer-specific parameters to human cortical current-clamp recordings, and embedded the model in a laminar cortical microcircuit. Reduced ATP production shifted the network from stable asynchronous baseline activity into a low-ATP burst-synchronized state characterized by reduced mean firing rates, increased Fano factor, and high-amplitude LFP-like oscillations. Increasing inhibitory synaptic conductance during persistent metabolic stress suppressed burst synchrony without restoring the metabolic state variable, producing an inhibition-stabilized low-energy state. Forward-backward parameter sweeps suggested history-dependent changes in burst synchrony, most consistently during inhibitory-conductance modulation, although these effects were readout- and dwell-time dependent. Together, these results provide a scalable spiking-network framework for studying how metabolic constraints shape cortical stability and for distinguishing suppression of seizure-like electrical activity from recovery of the underlying metabolic state.