Cognitive reserve (CR) is a concept meant to account for the frequent discrepancy between an individual’s measured level of brain pathology and her expected cognitive performance. It is particularly important within the context of aging and dementia, but has wider applicability to all forms of brain damage. As such, it has intimate links to related compensatory and neuroprotective concepts, as well as to the related notion of brain reserve. In this article, we introduce the concept of cognitive reserve and explicate its potential cognitive neural implementation. We conclude that cognitive reserve is compatible and complementary with many related concepts, but that each much draw sharper conceptual boundaries in order to truly explain preserved cognitive function in the face of aging or brain damage.

We introduce and describe the Reference Ability Neural Network Study and provide initial feasibility data. Based on analyses of large test batteries administered to individuals ranging from young to old, four latent variables, or reference abilities (RAs) that capture the majority of the variance in age-related cognitive change have been identified: episodic memory, fluid reasoning, perceptual speed, and vocabulary. We aim to determine whether spatial fMRI networks can be derived that are uniquely associated with the performance of each reference ability. We plan to image 375 healthy adults (50 per decade from age 20 to 50; 75 per decade from age 50 to 80) while performing a set of 12 cognitive tasks. Data on 174 participants are reported here. Three tasks were grouped a priori into each of the four reference ability domains. We first assessed to what extent both cognitive task scores and activation patterns readily show convergent and discriminant validity, i.e. increased similarity between tasks within the same domain and decreased similarity between tasks between domains, respectively. Block-based time-series analysis of each individual task was conducted for each participant via general linear modeling. We partialled activation common to all tasks out of the imaging data. For both test scores and activation topographies, we then calculated correlations for each of 66 possible pairings of tasks, and compared the magnitude of correlation of tasks within reference ability domains to that of tasks between domains. For the behavioral data, globally there were significantly stronger inter-task correlations within than between domains. When examining individual abilities, 3 of the domains also met these criteria but memory reached only borderline significance. Overall there was greater topographic similarity within reference abilities than between them (p<0.0001), but when examined individually, statistical significance was reached only for episodic memory and perceptual speed. We then turned to a multivariate technique, linear indicator regression analysis, to derive four unique linear combinations of Principal Components (PC) of imaging data that were associated with each RA. We investigated the ability of the identified PCs to predict the reference domain associated with the activation of individual subjects for individual tasks. Median accuracy rates for associating component task activation with a particular reference ability were quite good: memory: 82%; reasoning: 87%; speed: 84%; vocabulary: 77%. These results demonstrate that even using basic GLM analysis, the topography of activation of tasks within a domain is more similar than tasks between domains. The follow-up regression analyses suggest that all tasks with each RA rely on a common network, unique to that RA. Our ultimate goal is to better characterize these RA neural networks and then study how their expression changes across the age span. Our hope is that by focusing on these networks associated with key features of cognitive aging, as opposed to t...

Recent advances in neuroimaging have identified a large number of neural measures that could be involved in age-related declines in cognitive functioning. A popular method of investigating neural-cognition relations has been to determine the brain regions in which a particular neural measure is associated with the level of specific cognitive measures. Although this procedure has been informative, it ignores the strong interrelations that typically exist among the measures in each modality. An alternative approach involves investigating the number and identity of distinct dimensions within the set of neural measures and within the set of cognitive measures prior to examining relations between the two types of measures. The procedure is illustrated with data from 297 adults between 20 and 79 years of age with cortical thickness in different brain regions as the neural measures, and performance on 12 cognitive tests as the cognitive measures. The results revealed that most of the relations between cortical thickness and cognition occurred at a general level corresponding to variance shared among different brain regions and among different cognitive measures. In addition, the strength of the thickness-cognition relation was substantially reduced after controlling the variation in age, which suggests that at least some of the thickness-cognition relations in age-heterogeneous samples may be attributable to the influence of age on each type of measure.

Cognitive Reserve and Brain Maintenance have traditionally been understood as complementary concepts: Brain Maintenance captures the processes underlying the structural preservation of the brain with age, and might be assessed relative to age-matched peers. Cognitive Reserve, on the other hand, refers to how cognitive processing can be performed regardless of how well brain structure has been maintained. Thus, Brain Maintenance concerns the "hardware," whereas Cognitive Reserve concerns "software," that is, brain functioning explained by factors beyond mere brain structure. We used structural brain data from 368 community-dwelling adults, age 20-80, to derive measures of Brain Maintenance and Cognitive Reserve. We found that Brain Maintenance and Cognitive were uncorrelated such that values on one measure did not imply anything about the other measure. Further, both measures were positively correlated with verbal intelligence and education, hinting at formative influences of the latter to both measures. We performed extensive split-half simulations to check our derived measures' statistical robustness. Our approach enables the out-of-sample quantification of Brain Maintenance and Cognitive Reserve for single subjects on the basis of chronological age, neuropsychological performance and structural brain measures. Future work will investigate the prognostic power of these measures with regard to future cognitive status.

Analyses of large test batteries administered to individuals ranging from young to old have consistently yielded a set of latent variables representing reference abilities (RAs) that capture the majority of the variance in age-related cognitive change: Episodic Memory, Fluid Reasoning, Perceptual Processing Speed, and Vocabulary. In a previous paper (Stern et al., 2014), we introduced the Reference Ability Neural Network Study, which administers 12 cognitive neuroimaging tasks (3 for each RA) to healthy adults age 20–80 in order to derive unique neural networks underlying these 4 RAs and investigate how these networks may be affected by aging. We used a multivariate approach, linear indicator regression, to derive a unique covariance pattern or Reference Ability Neural Network (RANN) for each of the 4 RAs. The RANNs were derived from the neural task data of 64 younger adults of age 30 and below. We then prospectively applied the RANNs to fMRI data from the remaining sample of 227 adults of age 31 and above in order to classify each subject-task map into one of the 4 possible reference domains. Overall classification accuracy across subjects in the sample age 31 and above was 0.80 ± 0.18. Classification accuracy by RA domain was also good, but variable; memory: 0.72 ± 0.32; reasoning: 0.75 ± 0.35; speed: 0.79 ± 0.31; vocabulary: 0.94 ± 0.16. Classification accuracy was not associated with cross-sectional age, suggesting that these networks, and their specificity to the respective reference domain, might remain intact throughout the age range. Higher mean brain volume was correlated with increased overall classification accuracy; better overall performance on the tasks in the scanner was also associated with classification accuracy. For the RANN network scores, we observed for each RANN that a higher score was associated with a higher corresponding classification accuracy for that reference ability. Despite the absence of behavioral performance information in the derivation of these networks, we also observed some brain-behavioral correlations, notably for the fluid-reasoning network whose network score correlated with performance on the memory and fluid-reasoning tasks. While age did not influence the expression of this RANN, the slope of the association between network score and fluid-reasoning performance was negatively associated with higher ages. These results provide support for the hypothesis that a set of specific, age-invariant neural networks underlies these four RAs, and that these networks maintain their cognitive specificity and level of intensity across age. Activation common to all 12 tasks was identified as another activation pattern resulting from a mean-contrast Partial-Least-Squares technique. This common pattern did show associations with age and some subject demographics for some of the reference domains, lending support to the overall conclusion that aspects of neural processing that are specific to any cognitive reference ability stay constant across age, while aspects that are common to all...