Investigation of the human specimens is an essential element for understanding the pathogenesis of neurodegenerative disorders, such as Alzheimer's disease, Parkinson's disease, and multiple sclerosis. The studies hold promise for identifying biomarkers for diagnosis and prognosis, elucidating disease mechanisms, and accelerating the development of new strategies for therapeutic intervention. Here, we review proteomics studies of human brain samples in light of recent advances of mass spectrometry, focusing on the general strategies for experimental design and analysis (e.g., sample pooling and replication, selection of proteomics platforms, and false discovery rate in data processing), because quantitative analysis of clinical samples is confounded by a number of variables, including genetic differences, antemortem and postmortem factors, and experimental errors. Diverse proteomics platforms are also discussed with respect to sensitivity, throughput, and accuracy. Regarding the enormous complexity of the human brain and the limitation of current proteomics technologies, it may be more practical to analyze a subset of proteome in a functional context, in order to facilitate the identification of important disease-related proteins in the substantial noise reflecting biological and technical variances.