EDITORIAL Brain and Cognitive Reserve Perminder S. Sachdev, M.D., Ph.D., F.R.A.N.Z.C.P., Michael Valenzuela, B.Sc. Hons., M.B.B.S. Hons., Ph.D. he concept of reserve is well-recognized in med- related greater head size, measured as head circumfer- T icine. Various organs of the body, for example ence or intracranial volume, to a reduced risk of de- the kidney and the liver, manifest reserve such that mentia.4 The argument presented was that larger significant damage can occur to the organ without brains were able to tolerate a greater extent of pathol- impacting on clinical function. The very fact that an ogy before they reached the functional threshold at individual can donate one kidney and yet remain which symptoms became clinically manifest, that is, healthy attests to this fact. The glomerular filtration they had further to fall before reaching the disorder rate (GFR) is generally considered to be a measure of threshold. Head size is a proxy measure for brain size the excretory function of the kidney, with the normal before the onset of pathology, but its appeal is limited GFR being Ͼ90 mL/min/1.73 m2. Significant kidney by the fact that its association with intelligence or cog- damage can occur without affecting GFR, and the nitive capacity is generally poor. Further studies GFR is generally Ͻ30 mL/min/1.73 m2 before fea- showed that the relationship between head circumfer- tures of chronic renal failure are manifest.1 ence and dementia was restricted to those in the low to The brain is recognized to demonstrate a similar very low range of head girth,5 or to an interaction with reserve capacity. It is known that degeneration of do- the polymorphism on the apolipoprotein E gene.6 It pamine cells in the substantia nigra starts long before was also not clear what constituted the basis of a large the symptoms of Parkinson’s disease develop, and the brain. Was it large neuronal numbers, increased den- absolute number of pigmented neurons has been re- dritic proliferation, greater synaptic density, more ported to be reduced to about one-third in patients.2 white matter connections, or simply large ventricles? Individuals with brain injury appear to have different A major limitation of the “large brain” concept of thresholds at which they will develop clinical symp- reserve is that it is a static and passive viewpoint. An toms. This reserve capacity has received the greatest approach that Perneczky et al.7 take in this issue is to attention in the case of Alzheimer disease. An early relate the brain’s reserve capacity to its metabolic report showed that some individuals with high levels activity rather than the size. They use 18fluoro-2- of Alzheimer’s type pathology in their brains escaped deoxy-glucose positron emission tomography to ex- dementia during life, and that this was related to amine relative metabolic activity in different brain higher counts of large pyramidal neurons in their ce- regions. In previous studies, metabolic activity has rebral cortices.3 This led to the more commonly held been shown to be better related to functional capacity conceptualization of brain reserve, that is, the so-called than brain size,8 and Perneczky et al. have again dem- “hard” or “neurological” brain reserve. onstrated a significant association between glucose hy- According to the hard brain reserve concept, individ- pometabolism and impaired activities of daily living, uals differ in their reserve capacity depending upon especially for the right temporoparietal cortex. some characteristics of the brains they are endowed Furthermore, their study takes the concept of brain with. Early support for this came from studies that reserve as being dynamic and active, and subject to From the School of Psychiatry, Faculty of Medicine, University of New South Wales, Sydney; and Neuropsychiatric Institute, Prince of Wales Hospital, Randwick, Australia. Send correspondence and reprint requests to Perminder S. Sachdev, M.D., Ph.D., F.R.A.N.Z.C.P., UNSW School of Psychiatry, NPI, Euroa Centre, Prince of Wales Hospital, Barker Street, Randwick, NSW 2031, Australia. e-mail: [email protected] © 2009 American Association for Geriatric Psychiatry Am J Geriatr Psychiatry 17:3, March 2009 175 Editorial modification by experience. In fact, this is the dom- authors modeled a neural system as having small-world inant theme of the set of papers included in this architecture with “nodes” connected by “edges.” Ran- issue; that brain reserve is modifiable and the brain is dom node deletion, representing neuronal death, a highly plastic organ. Experience-related modifica- does not result in loss of local or global efficiency. tion of brain structure is well-illustrated by the study This aligns with the notion of neuronal plasticity. of Nithianantharajah et al.9 Previous research has Targeted deletion of central nodes leads to the loss of shown that environmental enrichment can produce a efficiency, which can be compensated by simulated number of positive brain effects in rodents, which neurogenesis. This article illustrates how neuroana- include greater dendritic length and branching, an tomical changes associated with brain reserve might increase in the number and size of synapses, forma- translate into maintenance or loss of functional net- tion and maturation of new neurons and circuits, works. It is likely that individuals with high reserve increased expression of neuronal signaling mole- have greater neurocomputational flexibility and cules, and greater synaptic plasticity.10 In the current might use a range of information processing path- study,9 the authors have shown that effects of envi- ways for solving complex problems.14 ronmental enrichment are region-specific. They have This brings us to another approach to brain reserve demonstrated decreased spine density in neurons of that emphasizes functional rather than structural re- the dentate gyrus, but no effect on neuronal density serve, the so called “cognitive reserve”14 or “intellec- in the CA1 region or the somatosensory and motor tual reserve.”15 According to this conceptualization, cortices. In addition, enrichment increased dendritic individuals with high brain reserve perform better diameter in the dentate gyrus and CA1 regions of the on cognitive tasks, and this provides them with a hippocampus but decreased soma area of layer V greater buffer in the process of their decline before pyramidal cells in the somatosensory cortex. This they reach a threshold for diagnosis of dementia. The work is further support for the concept of an adap- superior performance may be related to an efficient tive brain, and underlies the dynamic nature of what set of neural networks or a wider repertoire of con- was considered to be hard brain reserve. scious and preconscious cognitive strategies. These The morphological consequences of environmental may be innate and/or enriched by environmental enrichment raise the possibility that this might produce exposure. This repertoire arguably permits the high disease modification itself. Previous research in a reserve individuals to compensate for loss more ef- mouse model has shown that an enriched environment fectively than those with a limited repertoire.16 can delay the onset of Huntington’s disease.11 The In the dementia and ageing literature, various study reported in this issue9 showed different results, proxy measures have been used for cognitive re- suggesting variability in response, possibly due to dif- serve, and these include measures of intelligence, ferent enrichment methods or other methodological education, and occupation.17 The inference would be differences. This is important to remember when trans- that those with high education would be able to lating these findings to humans—rigorous human accumulate more pathology before clinical deficits studies are necessary before these findings are to be are manifest. A meta-analysis of the literature relat- accepted and translated into clinical intervention. The ing brain reserve to incident dementia,18 which in- implications are of great importance, as the morpho- cluded 22 studies comprising 29,000 individuals fol- logical changes reported may be sufficient to rescue lowed-up over a median 7.1 year, concluded that dysfunctional circuits and restore deficient memories higher brain reserve was associated with a lowered or other cognitive functions.12 There is even the possi- risk for incident dementia (summary odds ratio, 0.54; bility of slowing the progress of disease, that is, indi- 95% confidence interval, 0.49–0.59). Consistent with viduals with high brain reserve might have slower previous literature, Perneczky et al.7 very elegantly disease progression in the presence of similar patho- showed that well-educated patients with dementia mechanisms. with Lewy bodies could offset more brain damage, Activity or enrichment-dependent neuronal plas- as measured by glucose metabolic rates, until they ticity demonstrated in the mice experiments can be reached the same level of impairment on activities of modeled on a computer using network theory, as is daily living as the less educated comparison subjects. demonstrated by Rubinov et al.13 in this issue. These On the other hand, Christensen et al.19 did not find 176 Am J Geriatr Psychiatry 17:3, March 2009 Sachdev and Valenzuela an association between level of education indicators cognitive function,22 consistent with the Karp report.20 of brain reserve, such as brain volume and the extent Of the 54 published intervention studies, seven met of white matter hyperintensities, and high education strict criteria for inclusion. The authors21 conclude that was not a protective factor for cognitive decline. This intervention with cognitive exercises produces a posi- discrepant finding may be related to a number of tive effect with a strong effect size in healthy older methodological issues, including the fact that this people compared to wait-and-see control conditions, was a relatively young sample, being 60–64 years at and this effect does not appear to be diluted over 2 index assessment, most had high levels of education years.