Banked tissue is essential to the study of neurological disease but using postmortem tissue introduces a number of possible confounds. Foremost amongst these are factors relating to variation in postmortem interval (PMI). Currently there are conflicting reports on how PMI affects overall RNA integrity, and very few reports of how gene expression is affected by PMI. We analyzed total RNA extracted from frozen cerebellar cortex from 79 deceased human subjects enrolled in the Banner Sun Health Research Institute Brain and Body Donation Program. The PMI, which ranged from 1.5 to 45 h, correlated with overall RNA quality measures including RNA Integrity Number (RIN) (r = -0.34, P = 0.002) and RNA quantitative yield (r = -0.25, P = 0.02). Additionally, we determined the expression of 89 genes using a PCR-based gene expression array (RT(2) Profiler™ PCR Array: Human Alzheimer's Disease; SABiosciences™, Frederick, MD). A greater proportion of genes had decreased rather than increased expression with increasing PMI (65/89 vs. 20/89; P < 0.0001). Of these, transcripts from the genes ADAM9, LPL, PRKCG, and SERPINA3 had significantly decreased expression with increasing PMI (P < 0.01). No individual gene transcripts had significantly increased expression with increasing PMI. In conclusion, it is apparent that RNA degrades progressively with increasing PMI and that measurement of gene expression in brain tissue with longer PMI may give artificially low values. For tissue derived from autopsy, a short PMI optimizes its utility for molecular research.