In presurgical monitoring, focal seizure onset is visually assessed from intracranial electroencephalogram (EEG), typically based on the selection of channels that show the strongest changes in amplitude and frequency. As epileptic seizure dynamics is increasingly considered to reflect changes in potentially distributed neural networks, it becomes important to also assess the interrelationships between channels. We propose a workflow to quantitatively extract the nodes and edges contributing to the seizure onset using an across-seizure scoring. We propose a quantification of the consistency of EEG channel contributions to seizure onset within a patient. The workflow is exemplified using recordings from patients with different degrees of seizure-onset consistency.