INTRODUCTION Acetone has been shown to have broad-spectrum anticonvulsant actions in animal seizure models and has been hypothesized to play a role in the anticonvulsant mechanism of the ketogenic diet (KD). The present study examined the ability of a KD to elevate amygdaloid afterdischarge thresholds (ADT) in fully kindled rats. The effects of the KD were studied in the presence and absence of diallyl sulfide (DAS), an inhibitor of acetone metabolism. METHODS Twenty-four adult male rats were kindled to 30 stage 5 seizures. Afterdischarge thresholds (ADT) were determined. Subjects were then administered one of the following diets: (1) KD+V (vehicle; KD+V); (2) KD+DAS; (3) control diet+V (CD+V); (4) CD+DAS. They were stimulated every second day. Blood sampling was performed every second day--on non-stimulating days--to determine levels of glucose, beta-hydroxybutyrate, acetoacetate, and acetone. After 20 days, ADTs were re-determined. RESULTS Blood acetone concentrations were significantly higher in the KD+DAS group as compared to the other groups, although they did not reach "therapeutic levels". None of the treatments, however, elevated ADTs. CONCLUSIONS The KD was unable to elevate amygdaloid ADTs in fully kindled rats. Although subjects in the KD+DAS group achieved significant elevations of blood acetone, these concentrations (e.g. 0.2 mM) were much lower than those (>2.0 mM) previously shown to confer anticonvulsant activity. There appears to be large difference between humans and rats in their ability to produce elevated blood acetone levels on the KD. These data suggest that adult rats are not ideal subjects for modeling the anticonvulsant actions of the KD.