Intelligence is an important indicator of physical, mental and social well-being. In old age, intelligence is also associated with a higher quality of life and better health. Heritability studies have shown that there are strong genetic influences, yet unknown, on intelligence, including in old age. Other approaches may be useful to investigate the biological foundations of intelligence differences. Proteomics is a proven technique in revealing biomarkers for certain illnesses. In this pilot study, forty individuals were selected as the cognitive extremes from over 750 people in the Lothian Birth Cohort 1936 (age [approximately] 72 years) based on their high and low intelligence scores, as measured by a general cognitive ability factor. Urine samples were used as a stable, reliable and abundant source of proteins. Using capillary electrophoresis coupled to mass spectrometry (CE-MS) technology, the proteome of the high and low intelligence groups was determined. Data were calibrated and matched against the human urinary database, to enable comparative assessment. At a nominal significance level (P less than 0.05), there were several candidate proteins for association with intelligence, including a zinc finger protein (ZNF653) that has been associated with cognitive deficits, and complement C3 and collagen fragments that have been associated with Alzheimer's disease. Results are preliminary, do not survive multiple testing correction, and require validation. This pilot study shows the potential of this novel proteomics approach, and its applicability to understanding the biological foundations of intelligence differences. (Contains 1 table and 3 figures.)