The reduction of seizures achieved by open-loop neurostimulation beyond the stimulation intervals is attributed to adaptations in brain network dynamics. Our study investigates whether global network characteristics derived from high-density EEGs change over time of a novel focal neurostimulation therapy and whether these changes differentiate between responders and non-responders.

In the patients, either high-frequency or direct current (DC)-like stimulations were applied via electrode arrays placed epicranially above the individual epileptic focus region.

The connectivity analysis study included two groups of patients. In the first group, high-density EEGs were recorded before the stimulator was switched on and after 8-16 months of open-loop stimulations. In the second group, a further session of high-density recordings was performed 20 minutes after the initial start of the stimulation procedure. From 10 minutes of resting-state EEG, 10 artefact-free segments of equal length were selected for the analysis. Global graph metrics, including average path length, average cluster coefficient, synchronizability, assortativity, a measure which reflects the tendency of edges to connect vertices with similar properties, and average degree, were estimated from these segments.

Due to the still small sample size, no statistical tests were applied between the groups or across the different time points. On a qualitative level, direct after the start of the initial stimulation, network analysis showed a shorter average path length and a higher synchronizability. Responders to the stimulation presented lower assortativity prior to neurostimulation than non-responders and a higher average degree after 8-16 months of neurostimulation.

The analysis of epileptic network properties in epilepsy patients may help identify suitable candidates for focal neurostimulation therapy and provide deeper insights into its effects on the neuronal network.