Converging lines of research indicate that complex mental activity is associated with reduced dementia risk. Thus, intense interest exists in whether different forms of cognitive exercise can help protect against cognitive decline and dementia. However, there is considerable confusion in terminology that is hindering progress in the field. We therefore introduce a concrete definition of cognitive training (CT) and make this the focus of our article. Clinical research that has evaluated CT in normal aging, mild cognitive impairment, and dementia is then critically reviewed. Despite many methodological shortcomings, the overall findings indicate that multidomain CT has the potential to improve cognitive function in healthy older adults and slow decline in affected individuals. Finally, practical issues, including the strengths and weaknesses of commercial products, are explored, and recommendations for further research and clinical implementation are made.