Effects of an aged garlic extract and its thioallyl components on rat brain ischemia were examined using a middle cerebral artery occlusion model and a transient global ischemia model. In focal ischemia, an aged garlic extract, S-allyl cysteine (SAC), Allyl sulfide (AS) or Allyl disulfide (ADS) was administered 30 min prior to ischemic insult. Three days after ischemic insult, water contents of both ischemic and contralateral hemispheres were measured to assess the degree of ischemic damage. The water content of the ischemic control (no drug treatment) group was 81.50 +/- 0.07% (mean +/- SEM). It was significantly reduced with the administration of 300 mg/kg of SAC; the water content was 80.66 +/- 0.11% (P < 0.001). The histological observation using 2,3,5-triphenyltetrazolium chloride staining demonstrated that the administration of SAC reduced infarct volume. Neither AS nor ADS was effective. In global ischemia, the production of reactive oxygen species (ROS) was measured ex vivo using a spin-trapping agent, alpha-phenyl-N-tert-butylnitrone, and electron paramagnetic resonance spectroscopy. The production of ROS had two peaks; first at 5 min and second at 20 min after reperfusion. Both SAC and 7-nitro indazole, a nitric oxide synthase inhibitor, did not attenuate the amount of ROS produced at the first peak, but did the amount of the second peak. A possible involvement of peroxinitrite, which may be formed from superoxide and nitric oxide and is known to be highly toxic in ischemia/reperfusion injury of the brain, was suggested.