Total RNA was extracted from 105 individual postmortem human brain samples representing a range of postmortem conditions. To improve upon parameters currently used to screen for RNA quality, electropherogram patterns generated by the Agilent Bioanalyzer 2100 were compared to the average score in random hexamer-primed reverse transcription real-time PCR for four housekeeping genes in each RNA sample. The ribosomal ratio (28S to 18S) was found to be unrelated to the housekeeping gene score (r = -0.06; P = 0.50), and there was no threshold value in the ratio that could be applied to effectively categorize the RNA degradation. Although the housekeeping gene score correlated significantly with the percentage of area in the electropherogram corresponding to moderate to high molecular weight intact mRNA (r = 0.41; P = 0.0001), the best discriminator was determined to be the ratio of the 18S peak height to the highest peak in the tRNA to 18S rRNA baseline. Applying a lower boundary of 2.12 for the ratio allowed for the screening out of samples with the lowest housekeeping gene scores without excluding better-quality samples. This measure represents a marked improvement over the 28S to 18S ratio, which proved to be a misleading indicator of the state of the mRNA for use in random hexamer-primed reverse transcription PCR.