Poor patient outcomes in resective epilepsy surgery remain an area of clinical concern. Regardless of epilepsy type, a third or greater of drug-resistant epilepsy (DRE) patients who undergo resection do not report seizure freedom at a two-year follow-up. One of the likely causes is that current methods for identifying which tissue to resect are subjective and potentially inaccurate. Much of resective surgery planning is based on identification of initial ictogenic tissue--the seizure onset zone (SOZ). However, complete resection of the SOZ, as currently identified, is a poor prognosticator of seizure freedom. We developed an algorithm using Granger Causality in the time domain to create an objective and reproducible measure of the SOZ. Intracranial electroencephalographic DRE patient data (N = 58) from the Hospital of the University of Pennsylvania was pre-processed, and time domain Granger Causality analysis was performed on one minute of recording centered around the clinically identified ictal onset. Graph analysis was used to evaluate the weighted out-degree of each channel in the network, and normalized values above 0.5 were considered candidate SOZ channels. These candidate SOZ channels were then compared with the current gold-standard, clinically identified SOZ. Our method had a sensitivity and specificity of 74% and 62%, respectively, with Cohen’s Kappa 0.14. While further algorithm development is expected to increase the performance, these results demonstrate the potential utility of effective functional connectivity as an objective, reproducible biomarker of the SOZ. Future work will evaluate the potential for this biomarker to improve resective surgery outcomes.