Author Summary Network neuroscience applied to epileptic brains seeks to identify pathological neural connections that promote and maintain seizures, and holds promise to guide surgical planning in patients with intractable epilepsy. However, sampling of the epileptic network in intracranial EEG recording is limited by the choice of where to place intracranial electrodes, which is highly variable within and between epilepsy centers. The effect of incomplete spatial sampling generated by sparse electrode placement on network statistics is unknown. Here, we determine the sensitivity of several network statistics to incomplete spatial sampling, and we propose a method using electrode subsampling to determine patient-specific confidence intervals in network model predictions. Abstract Network neuroscience applied to epilepsy holds promise to map pathological networks, localize seizure generators, and inform targeted interventions to control seizures. However, incomplete sampling of the epileptic brain because of sparse placement of intracranial electrodes may affect model results. In this study, we evaluate the sensitivity of several published network measures to incomplete spatial sampling and propose an algorithm using network subsampling to determine confidence in model results. We retrospectively evaluated intracranial EEG data from 28 patients implanted with grid, strip, and depth electrodes during evaluation for epilepsy surgery. We recalculated global and local network metrics after randomly and systematically removing subsets of intracranial EEG electrode contacts. We found that sensitivity to incomplete sampling varied significantly across network metrics. This sensitivity was largely independent of whether seizure onset zone contacts were targeted or spared from removal. We present an algorithm using random subsampling to compute patient-specific confidence intervals for network localizations. Our findings highlight the difference in robustness between commonly used network metrics and provide tools to assess confidence in intracranial network localization. We present these techniques as an important step toward translating personalized network models of seizures into rigorous, quantitative approaches to invasive therapy.