The occurrence of practice effects on repeated cognitive testing is well known, yet rarely systematically taken into account. Researchers may worry that it is too costly and time-consuming. Or perhaps it is just an esoteric measurement issue that will not be worth the effort. However, failure to account for practice effects can have a substantial negative impact in aging neuroscience. What practice effects are—usually thought of as simply improvement in performance at follow-up—tends to seem obvious, which may result in a tendency to assume that we are all talking about the same thing when we talk about practice effects. Thus, a common question is: What is the best approach to gauging practice effects? Rather than a one-size-fits-all approach, different approaches to gauging practice effects address different issues and serve very different purposes. The more appropriate question is: Which approach is best for the issue I want to address? Importantly, in older adults—except for very, very short test-retest intervals—age-related cognitive decline is the norm. As such, when there is no change or even when there is observed decline, there may still be practice effects because scores would have been even lower without prior test exposure. Moreover, practice effects in older adults have been demonstrated after intervals as long as 5 or more years. For this Research Topic, we present different approaches to practice effects and compare what purposes they can and cannot serve, and how well they gauge practice effects. Regression-based approaches can be used to predict cognitive trajectories. Examining practice effects after a very brief interval can be used to predict individuals who are most likely to decline and progress to mild cognitive impairment (MCI) or dementia. Doing so can be a substantial aid to subject selection for clinical trials aimed at early intervention of individuals at elevated risk. A replacement-subjects approach—comparing matched replacements who have taken tests once to returnees who have taken them before—can be used to revise cognitive scores based on estimated practice effects. This approach cannot aid in prediction of cognitive trajectories or who is likely to develop MCI/dementia in the future. By altering cognitive scores, it can, however, change when a person crosses the impairment threshold used in diagnosing MCI. In so doing, it provides for earlier detection of both incident MCI and progression from single- to multiple-domain MCI.
In turn, earlier detection means a better chance of slowing disease progression. Employing approaches to gauging practice effects for subject selection or earlier detection of MCI in clinical trials can also substantially reduce study time, subject/staff burden, and save literally millions of dollars. Alzheimer’s disease biomarker concordance is a key diagnostic validator, and accounting for cognitive practice effects can also increase the concordance between diagnosis and biomarker positivity, thereby increasing diagnostic accuracy. In sum, incorporation of practice effects into much aging neuroscience research has substantial real world clinical and research implications for understanding cognitive decline and disease progression, timely diagnosis, and cost savings in clinical trials.
The occurrence of practice effects on repeated cognitive testing is well known, yet rarely systematically taken into account. Researchers may worry that it is too costly and time-consuming. Or perhaps it is just an esoteric measurement issue that will not be worth the effort. However, failure to account for practice effects can have a substantial negative impact in aging neuroscience. What practice effects are—usually thought of as simply improvement in performance at follow-up—tends to seem obvious, which may result in a tendency to assume that we are all talking about the same thing when we talk about practice effects. Thus, a common question is: What is the best approach to gauging practice effects? Rather than a one-size-fits-all approach, different approaches to gauging practice effects address different issues and serve very different purposes. The more appropriate question is: Which approach is best for the issue I want to address? Importantly, in older adults—except for very, very short test-retest intervals—age-related cognitive decline is the norm. As such, when there is no change or even when there is observed decline, there may still be practice effects because scores would have been even lower without prior test exposure. Moreover, practice effects in older adults have been demonstrated after intervals as long as 5 or more years. For this Research Topic, we present different approaches to practice effects and compare what purposes they can and cannot serve, and how well they gauge practice effects. Regression-based approaches can be used to predict cognitive trajectories. Examining practice effects after a very brief interval can be used to predict individuals who are most likely to decline and progress to mild cognitive impairment (MCI) or dementia. Doing so can be a substantial aid to subject selection for clinical trials aimed at early intervention of individuals at elevated risk. A replacement-subjects approach—comparing matched replacements who have taken tests once to returnees who have taken them before—can be used to revise cognitive scores based on estimated practice effects. This approach cannot aid in prediction of cognitive trajectories or who is likely to develop MCI/dementia in the future. By altering cognitive scores, it can, however, change when a person crosses the impairment threshold used in diagnosing MCI. In so doing, it provides for earlier detection of both incident MCI and progression from single- to multiple-domain MCI.
In turn, earlier detection means a better chance of slowing disease progression. Employing approaches to gauging practice effects for subject selection or earlier detection of MCI in clinical trials can also substantially reduce study time, subject/staff burden, and save literally millions of dollars. Alzheimer’s disease biomarker concordance is a key diagnostic validator, and accounting for cognitive practice effects can also increase the concordance between diagnosis and biomarker positivity, thereby increasing diagnostic accuracy. In sum, incorporation of practice effects into much aging neuroscience research has substantial real world clinical and research implications for understanding cognitive decline and disease progression, timely diagnosis, and cost savings in clinical trials.
