In this paper, we review studies that have investigated whether neuropsychological, neuropsychiatric, and neuroimaging measures predict decline to Alzheimer's disease (AD). Prospective neuropsychological studies indicate that cognitive performance may be an excellent indicator of future progression from mild cognitive impairment (MCI) to AD, particularly when episodic memory is combined with tasks relying on executive control and language tasks. Research on neuropsychiatric symptoms reveal that depression, apathy, anxiety, and sleep disturbances can contribute to predictive models, though their sensitivity is typically lower than that found with cognitive measures. Finally, different structural brain imaging markers reveal excellent predictive accuracy. The paper discusses issues that will have to be addressed in future studies. First, it will be necessary to increase the evaluation of combined markers, as this may considerably improve predictive accuracy. Second, it will be necessary to move to earlier stages than MCI in order to expand the detection window. Third, processes of compensation and plasticity will have to be better investigated as research moves into earlier stages. The Consortium for the early identification of AD-Quebec (CIMA-Q) is presented as an instance of this approach, and potential batteries of measures are proposed.