A method has been developed to differentiate between non-muscle tissues such as liver, kidney and heart and that of muscle in meat samples using tissue specific DNA detection. Only muscle tissue is considered meat from the point of view of labelling (Food Labelling [Amendment] (England) Regulations 2003) and Quantitative Ingredient Declaration (QUID), and also certain parts of the carcass are prohibited to be used in raw meat products (Meat Products [England] Regulations 2003). Included in the prohibited offal are brain and spinal cord. The described methodology has therefore been developed primarily to enforce labelling rules but also to contribute to the enforcement of BSE legislation on the detection of Central Nervous System (CNS) tissue. The latter requires the removal of Specified Risk Material (SRM), such as bovine and ovine brain and spinal cord, from the food chain. Current methodologies for detection of CNS tissue include histological examination, analysis of cholesterol content and immunodetection. These can potentially be time consuming, less applicable to processed samples and may not be readily adapted to high throughput sample analysis. The objective of this work was therefore to develop a DNAbased detection assay that exploits the sensitivity and specificity of PCR and is potentially applicable to more highly processed food samples. For neuronal tissue, the DNA target selected was the promoter for Glial Fibrillary Acidic Protein (GFAP), a gene whose expression is restricted to astroglial cells within CNS tissue. The promoter fragments from both cattle and sheep have been isolated and key differences in the methylation patterns of certain CpG dinucleotides in the sequences from bovine and sheep brain and spinal cord and the corresponding skeletal muscle identified. These have been used to design a PCR assay exploiting Methylation Specific PCR (MSP) to specifically amplify the neuronal tissue derived sequence and therefore identify the presence of CNS tissue in an assay sample.