Excitatory-inhibitory mismatch shapes node recruitment in an epileptic network

Objective: Focal epilepsy is thought to be a network disease, in which epileptiform activity can spread noncontiguously through the brain via highly interconnected nodes, or hubs, within existing networks. Animal models confirming this hypothesis are scarce, and our understanding of how distant nodes are recruited is also lacking. Whether interictal spikes (IISs) also create and reverberate through a network is not well understood.

Methods: We injected bicuculline into the S1 barrel cortex and employed multisite local field potential and Thy-1 and parvalbumin (PV) cell mesoscopic calcium imaging during IISs to monitor excitatory and inhibitory cells in two monosynaptically connected nodes and one disynaptically connected node: ipsilateral secondary motor area (iM2), contralateral S1 (cS1), and contralateral secondary motor area (cM2). Node participation was analyzed with spike-triggered coactivity maps. Experiments were repeated with 4-aminopyridine as an epileptic agent.

Results: We found that each IIS reverberated throughout the network, differentially recruiting both excitatory and inhibitory cells in all connected nodes. The strongest response was found in iM2. Paradoxically, node cM2, which was connected disynaptically to the focus, was recruited more intensely than node cS1, which was connected monosynaptically. The explanation for this effect could be found in node-specific excitatory/inhibitory (E/I) balance, as cS1 demonstrated greater PV inhibitory cell activation compared with cM2, where Thy-1 excitatory cells were more heavily recruited.

Significance: Our data show that IISs spread noncontiguously by exploiting fiber pathways that connect nodes in a distributed network and that E/I balance plays a critical role in node recruitment. This multinodal IIS network model can be used to investigate cell-specific dynamics in the spatial propagation of epileptiform activity.