Ageing of the brain may lead to a slow degradation of cognitive function, which ultimately leads to learning and memory deficits in the elderly. These processes are complex and their molecular mechanisms are still incompletely known. The hippocampus is a key region involved in several types of learning including spatial learning. Since both ageing and learning processes are known to involve changes in gene expression, we have studied gene expression changes during ageing in the event that these perturbations may be responsible for the learning deficits in the elderly and those suffering from neurodegenerative disease. To identify key genes involved in these processes, we measured gene expression in the dorsal hippocampus (dHPC) of 6-, 15- and 24-month-old mice following a hippocampus-dependent spatial learning protocol, the Repeated Acquisition and Performance Chamber (RAPC). Behavioural data showed a significant learning impairment in middle- and old-aged mice when compared with young mice, as assessed by an increased number of mistakes made by the subjects in the RAPC protocol. Gene expression changes in the dHPC were measured using high-density oligonucleotide microarrays. A number of statistical and supervised learning approaches have been employed to analyse the data set. One data analysis approach employs the Independent Consistent Expression Discriminator (ICED), which was designed to provide a more biologically relevant search criterion during predictor selection by embracing the inherent variability of gene expression in any biological state. The search criteria employed by ICED is designed to identify not only genes that are consistently expressed at one level in one class and at a consistently different level in another class but identify genes that are variable in one class and consistent in another. The result is a novel approach to accurately selecting biologically relevant predictors of differential biological states from a small number of microarray samples. Using both statistical and pattern recognition approaches, we found 175 genes significantly changed with ageing (59 % of which decreased), 305 with learning (58 % of which decreased) and 325 that showed interactive effects with both ageing and learning. In addition, we have generated class predictors from the analysis which can be used classify the memory and learning deficits as a function of ageing and disease. This study links gene expression to behaviour and shows the possibility of using microarray technology to profile gene expression in behavioural protocols. Our large-scale gene expression analysis after learning in young, middle and old aged animals provides for the first time an overview of genes that may be responsible for the spatial learning deficit in aged mice. Genes identified in this study and their proteins are potential candidates for therapeutic interventions that could revert or slow down age- and disease-dependent cognitive decline.

All procedures accord with current UK legislation.