We found that quantitative abnormalities in interictal activity and connectivity were higher in regions which generated spikes and seizures compared to uninvolved brain and used a random forest model to predict epileptogenicity based on these two groups of features.