Objectives: Cryosurgery is an alternative technique for minimally invasive treatment of lesions. We have recently examined cryosurgery for epilepsy in animal models, and found that penicillin G (PG)-induced epileptiform discharges (EDs) mostly vanished after freezing. However, EDs were provoked again after insufficient freezing. Inadequate freezing is not visually detectable during and just after freezing and is not predictable beforehand. To manage this problem, we examined whether intraoperative monitoring of electrocorticograms (ECoGs) can predict recurrence of EDs after cryosurgery. Methods: A palm-sized cryoprobe system was applied to focal seizures in a Wistar rat model in which EDs were induced in advance by intracerebral injection of PG. During stable induction of EDs, the cryoprobe was carefully inserted into the epileptic region and this region was immediately frozen. After the series of prefreezing, freezing, and postfreezing, rats in which PG-induced EDs relapsed within 3 h were defined as the ED-relapsed (EDR) group, and other rats were defined as the ED-vanished (EDV) group. Time-frequency analysis was conducted on the ECoGs in each group through each freezing series. Results: Relapse of PG-induced EDs on ECoG after the freezing series was associated with the remaining power of the delta band in the freezing period more strongly in the EDR group than in the EDV group. Conclusions: Success or failure of the freezing procedure can be predicted by the specificity of the delta band of the ECoG obtained intraoperatively.