Infarct size is a good predictor of the neurological outcome following stroke. Estimation of infarct size in the early phase following experimental stroke depends on the availability of reliable techniques that can distinguish ischemic from nonischemic tissue. The objective of this study was to provide a simple and robust method for reliable delineation of the ischemic infarct area in fresh frozen cryosections from mice subjected to focal cerebral ischemia. Mice were subjected to permanent middle cerebral artery (MCA) occlusion and euthanised after 30 min, 1, 2, 4, 6, 12 and 24 h. The size of the developing infarct was compared in parallel series of sections in situ hybridized for mRNA encoding the enzyme glyceraldehyde-3-phosphate dehydrogenase (GAPDH) or stained with toluidine blue (TB). The infarct was clearly delineated in GAPDH mRNA in situ hybridized sections as soon as 4 h after MCA occlusion. Infarct size was similar at 4 and 6 h in GAPDH mRNA in situ hybridized sections. Sections hybridized for GAPDH mRNA showed significantly larger infarcts than sections stained with TB after 6 h but not after 24 h of ischemia. Analysis of in situ hybridized sections revealed changes in neuronal GAPDH mRNA in areas prone to undergo degeneration 30 min to 1 h after MCA occlusion, thereby preceding visible pycnosis in TB-stained sections. The results showed that in situ hybridization for GAPDH mRNA was a reliable method and superior to TB staining for precise infarct delineation prior to 6 h of permanent MCA occlusion.