Table 1.

Advantages and limitations of different lines of evidence

Lines of Evidence	Advantages	Limitations	Examples
Epidemiological/Cohort Studies	Large number of subjects Data on multiple factors Longitudinal follow-up over many years	Association not causation Uncertainty about the influence of unidentified factors Uncertainty about whether an intervention aimed at an identified factor would have an impact on clinical outcome	Physical activity Intellectually stimulating activity Diet/Nutrition Alcohol consumption
Basic Science/Animal Studies	Investigation under highly controlled conditions Determination of underlying mechanisms Identification of intervening variables (e.g., changes in brain structure, neurotransmitters, growth factors, inflammatory markers, etc.)	Reduction of complex issues into simple ones Need to translate findings from basic science or animal models to the study of humans Defining underlying mechanisms does not signify that manipulating an identified factor will alter clinical outcome	Physical activity Enriched/stimulating environment Pharmacological agents Statins DHA Resveratrol Ginkgo biloba
Human “Proof- of-Concept” Studies	Opportunity to examine markers of cerebral plasticity, reserve, efficiency, and neural compensation Utilization of the tools of cognitive neuroscience to elucidate underlying mechanisms Opportunity for further hypothesis testing and generation	Similar to the limitations associated with basic science/animal studies noted above	Structural Imaging Cognitive or fitness training and alterations in brain structure Functional Imaging Cognitive training and neurotransmitter binding or changes in correlates of neural activity Event- Related Potentials Aerobic Fitness and markers of resource allocation and processing speed Differences in resource allocation between cognitively high vs. average performers
Human Intervention Studies	Often considered the ‘gold standard’ Randomization helps control for selection bias and other variables that may influence clinical outcomes Outcome measures can include cognitive performance as well as biological or neuroimaging markers	Feasibility of conducting such studies for financial, logistical, or ethical reasons Uncertainty about the extent to which the observed effects can be generalized Uncertainty about whether statistically significant results are clinically relevant Uncertainty about whether a failed trial was due to the particular dose/duration of an agent, or the particular sample studied	Exercise/Physical training Cognitive training Pharmacologic agents Antihypertensive medication Statins Antioxidants DHA Ginkgo biloba

Key: DHA= docosahexaenoic acid