The brain tissue obtained after death is subjected to several circumstances that can affect RNA integrity. The present study has been directed to reveal possible pitfalls and to control RNA normalization in post-mortem samples in order to recognize the limitations and minimize errors when using TaqMan PCR technology. This has been carried out in samples of the frontal cortex in a series of control and diseased cases covering Parkinson's disease, dementia with Lewy bodies pure form and common form, and Alzheimer's disease. Special attention has been paid to the value of the agonal state, post-mortem delay and pH of the nervous tissue as approximate predictors of the quality of RNA, as well as to the use of the Bioanalyzer to confirm RNA preservation. In addition, since possible disease-modified mRNAs have to be normalized with ideal unaltered RNAs, TaqMan® human endogenous control plates have been used to determine the endogenous control most appropriate for the study. β-glucuronidase (GUS) and β-actin were good endogenous controls because their expression levels showed a small variation across a representative number of control and pathological cases. RNA stability was also analysed in a paradigm mimicking cumulative delay in tissue processing. GUS mRNA levels were not modified although β-actin mRNA levels showed degradation at 22 h. Finally, the control of RNA degradation for the normalization of genes of interest was also tested. mRNA expression levels for superoxide dismutase 1 (SOD1) and metalloproteinase domain 22 (ADAM22) were examined at several artificial post-mortem times, and their expression levels compared with those for putative controls β-actin and GUS. In our paradigm, the expressions of SOD1 and ADAM22 were apparently not modified when normalized with β-actin. Yet their expression levels were reduced with post-mortem delay when values were normalized with GUS. Taken together, these observations point to practical consequences in TaqMan PCR studies. Short post-mortem delays and acceptable pH of the brain are not sufficient to rule out RNA degradation. The selection of adequate endogenous controls is pivotal in the study. β-actin and GUS are found to be good endogenous controls in these pathologies, although GUS but not β-actin expression levels are preserved in samples with long post-mortem delay.