Introduction Responsive Neurostimulation (RNS) provides an alternative treatment for medication-refractory epilepsy cases which are not a candidate for resective epilepsy surgery. It has been shown effective with complication rates comparable to similar procedures. This study reports the data on responsive neurostimulation patients from Massachusetts General Hospital (MGH) for the first time. Methods We retrospectively studied all patients who were implanted with RNS at MGH, which was from October 2006 to July 2017. We collected data on age of implantation, sex, and the localization of epileptogenic zone. The seizure reduction was calculated by subtracting the frequency of seizures per month in the last visit after implantation (report of past 6 months) from the seizure frequency before implantation. This was divided by seizure frequency before RNS implantation and reported in percentage. For calculating the responder rate, we excluded the three cases that received stimulation less than 6 months. Also we reviewed the rate of different complications associated with RNS. Results 25 patients (12 female and 13 male) ranging from 20 to 65 years old (mean ± std: 37.4 ± 12.6) were implanted with RNS from October 2006 to July 2017. One patient was explanted after 2 months due to wound infection so it was excluded from seizure outcome analyses. In terms of epileptogenic zone, 11 cases were neocortical, 9 cases were bilateral mesiotemporal, 2 cases were left mesiotemporal, 2 cases were left mesiotemporal and neocortical and 1 case was thalamic. The seizure reduction percentage ranged between 0% and 100% with median reduction of 88%. The responder rate (patients with >50% seizure reduction) was 76.19% (16 out of 21 patients) and the super-responder rate (patients with >90% seizure reduction) was 57.14% (12 out of 21 patients). There was a positive correlation between seizure reduction and time from implantation which means patients who had the RNS for longer showed better outcome (Spearman correlation: r = 0.43, p = 0.067).There was no significant difference in average seizure reduction percentage between neocortical (mean: 45.75 %) and bilateral mesiotemporal (mean: 70.00%) cases (p: 0.351). Seventeen patients had no post-operative complication (68%), 4 patients had intracranial hemorrhage (16%), 2 patients had wound infection (8%), 1 patient had headache (4%) and 1 patient had wound dehiscence (4%). Conclusion This study shows that RNS has been an effective treatment modality in medication refractory epilepsy cases at MGH, with minimal complications. The improvement of outcome over time supports the hypothesis that RNS may function through seizure network modulation. Future works are needed to find optimal frequency and length of stimulation to achieve faster and more efficacious network modulation.