The present study was designed to evaluate the anticonvulsant effects of a high-fat ketogenic diet (KD) in rats. Animals were maintained on one of four experimental diets: (1) calorie-restricted ketogenic (KCR); (2) calorie-restricted normal (NCR); (3) ad libitum ketogenic (KAL); or (4) ad libitum normal (NAL). The calorie-restricted diets were fed in quantities such that they were calorically equivalent. All animals began diet treatment at age P37 and each was subjected to one of five chemically-induced seizure tests: bicuculline (BIC; s.c.), picrotoxin (PIC; s.c.), kainate (KA, i.p. or s.c.) and gamma-butyrolactone (GBL, i.p.), strychnine (s.c.). Bipolar epidural electrodes were implanted under ketamine/xylazine anesthesia to permit recording the spike and wave discharges (SWD) characteristic of electroencephalograms during absence seizures. Ketonemia was assayed by measuring blood levels of beta-hydroxybutyrate (BHB) spectrophotometrically prior to induction of seizures in each experiment. Animals fed ketogenic diets (i.e. either calorie restricted or ad libitum) exhibited greater blood levels of BHB compared to control groups. Seizure results show that treatment with a KD: (1) reduced the incidence of bicuculline-induced convulsions; (2) diminished the number of picrotoxin-induced seizures (KCR group only); (3) increased latency to GBL-induced SWD and reduced both the number and duration of SWD; but (4) conferred no protection from strychnine-induced seizures; and (5) made KA-induced seizures more severe. Together these results indicate a spectrum of anticonvulsant action for the KD in rats that includes threshold seizures induced via GABA receptors (BIC, PIC, GBL) but not those induced at glycine (strychnine) or the KA-subclass of glutamate receptors. Uniquely, the KD is the only treatment described that protects against both convulsive and non-convulsive (absence) seizures in rats.