Elsevier

Ageing Research Reviews

Volume 5, Issue 3, August 2006, Pages 354-369
Ageing Research Reviews

Review
Neuroprotective effects of cognitive enrichment

https://doi.org/10.1016/j.arr.2006.04.004Get rights and content

Abstract

Cognitive enrichment early in life, as indicated by level of education, complexity of work environment or nature of leisure activities, appears to protect against the development of age-associated cognitive decline and also dementia. These effects are more robust for measures of crystallized intelligence than for measures of fluid intelligence and depend on the ability of the brain to compensate for pathological changes associated with aging. This compensatory ability is referred to as cognitive reserve. The cognitive reserve hypothesis suggests that cognitive enrichment promotes utilization of available functions. Alternatively, late life cognitive changes in cognition may be linked to a factor, such as cholinergic dysfunction, that is also present early in life and contributes to the reduced levels of early life cognitive enrichment. Beneficial effects of environmental enrichment early in life have also been observed in rodents and primates. Research with rodents indicates that these changes have structural correlates, which likely include increased synapses in specific brain regions. Dogs also show age-dependent cognitive decline, and both longitudinal and cross-sectional studies indicate that this decline can be attenuated by cognitive enrichment. Furthermore, cognitive enrichment has differential effects, improving some functions more than others. From a neurobiological perspective, behavioral enrichment in the dog may act to promote neurogenesis later in life. This can be distinguished from nutritional interventions with antioxidants, which appear to attenuate the development of neuropathology. These results suggest that a combination of behavioral and nutritional or pharmacological interventions may be optimal for reducing the rate of age-dependent cognitive decline.

Introduction

Improved medical care and other factors leading to increased life span have a major downside, namely an increased incidence of neurodegenerative disorders, such as Alzheimer's disease (AD), associated with advanced age. The challenge of developing interventions to improving the quality of life of the elderly has become a major focus of the pharmaceutical industry, which has developed a wide variety of novel compounds, which typically are targeted against specific receptors linked to cognitive decline. This purely pharmaceutical strategy has been mainly directed at symptomatic treatment and has shown limited efficacy, although some compounds have achieved regulatory approval for treatment of AD. At best, these pharmaceuticals delay the normal progression of AD, and likely for months rather than years.

A second strategy based on nutritional interventions has shown some promise. However, this work is difficult to evaluate clinically, and is largely supported by retrospective data, which is inherently uncontrolled.

A third strategy involves application of behavioral interventions. Research on human subjects has focused on two categories, exercise and cognitive enrichment. Exercise should be equated with physical activity, and has been studied in both retrospective human studies as well as in animal studies, as an intervention. The human studies generally report beneficial effects of exercise on cognitive ability later in life (e.g., Christensen and Mackinnon, 1993), but there are also conflicting data (Anstey and Christensen, 2000, Churchill et al., 2002). Furthermore, positive effects of exercise may be linked to cognitive processes, such as the planning and other preparative activities (e.g., Colcombe and Kramer, 2003).

The present review focuses primarily on the utility of behavioral strategies, particularly those associated with cognitive experience, in protecting against age-dependent cognitive decline. Is the cliché “use it or lose it” valid in the case of cognitive ability and if so, why? To address these questions, we first provide an overview of studies with human subjects that point to beneficial long-term effects of cognitive activity early in life on cognitive ability at a later time. We then look at studies with animal models (rodents and primates), demonstrating that experiential effects do affect brain structure. Next, we will review work done in our laboratories on the beneficial effects of enrichment on cognition and brain pathology in the aged beagle dog. The last section of this review focuses on future directions, particularly in studies with animal models.

A substantial literature has considered the importance of cognitive enrichment earlier in life on cognitive ability at a later time. This work has been reviewed in detail (e.g., Kramer et al., 2004). Here, our goal is to provide a synopsis of the literature, and to outline some potential pitfalls in interpretation. Education is the most extensively studied measure of enrichment. Other indices of enrichment include job complexity and types of leisure activity.

Anstey and Christensen (2000) identified 32 studies that examined various predictors of age-associated cognitive decline. Education was one of four positive predictors. These studies show, in general, that education has a protective effect with respect to development of dementia, with a decreased risk with increased education (e.g., Wilson et al., 2002). Positive effects of education have also been obtained on measures of mental vitality as assessed by the Mini-Mental State Exam (e.g., Butler et al., 1996), and on specific cognitive abilities including verbal and non-verbal memory and conceptualization (e.g., Albert et al., 1995). On the other hand, previous education has a much weaker relationship with measures of processing speed, indicted by rapidity of response (Christensen et al., 1997).

The effect of education earlier in life on cognitive ability has also been examined in cross-sectional studies. Ardila et al. (2000) studied 806 participants of various ages and educational background on a battery of neuropsychological tests. As expected, they found a strong association between test performance and level of education. This relationship, however, did not hold within all educational groups. Individuals with low levels of education tended to perform better at an older age on neuropsychological tests; higher-educated subjects, by contrast, showed just the reverse, performing more poorly with increasing age. Complicating the issue further, education affected performance on some tests more than it did others, which is consistent with evidence from longitudinal studies that beneficial effects of education vary among assessment measures.

This literature, therefore, suggests that education has beneficial effects on cognition later in life, but that these effects are process specific. Based on analysis of these studies, Kramer et al. (2004) suggest that education primarily affects performance on measures of crystallized intelligence, as opposed to fluid intelligence. Cognitive psychologists regard crystallized intelligence as a construct linked to acquired experience and knowledge and distinguish it from fluid intelligence, which is reasoning ability and assumed to be more dependent on biology than on experience. Processing speed, for example, is considered to reflect fluid intelligence, while verbal memory is an example of crystallized intelligence.

Both longitudinal and cross-sectional studies have limitations, and the results can be explained in more than one way. Neither type of study clearly distinguishes between cause and effect—decreased cognitive activity early in life may reflect factors causal to development of dementia or decreased cognitive activity early in life and dementia later in life may have separate causes. Riley et al. (2005) looked at linguistic ability early in life in participants of the Nun Study on both behavioral and neuropathological measures later in life. They found that high levels of linguistic ability at an early age were associated later in life with a lower incidence of cognitive impairment and with a lower incidence of neuropathological measures associated with AD. These findings suggest that the factors responsible for poor performance earlier in life contribute to both levels of education and to the subsequent development of impairment. Winnock et al. (2002) reported a much higher incidence of the APOE4 allele, one of the strongest biomarkers of dementia, in low educated subjects, raising the possibility of the APOE phenotype having an influence on cognition throughout the entire life span.

Sarter and Bruno (2004) argue that cholinergic abnormalities are a common underlying factor that accounts for both early life limitations in cognitive ability and also cognitive deficits later in life. The early life effect limits educational ability. This hypothesis is supported by evidence of cholinergic dysfunction in AD and in mild cognitive impairment. Cholinergic dysfunction early in life has been shown to limit cognitive ability of transgenic mice, but has not yet been linked to cognitive function in humans.

Mulatu and Schooler (1999), following up on earlier longitudinal studies, reported that complexity of work based on requirements of thought and independent judgment, has a positive influence on level of intellectual function later in life. Similar effects have also been reported for leisure activity and general lifestyle, such as participating in games such as bridge (Hultsch et al., 1999)

Bosma et al. (2003) provided further evidence of the importance of work experience. In a follow up of the Maastrict Aging Study, they examined a subset of the population that were 50 or older, and non-impaired when first tested. Work complexity was assessed, based on the subjects’ estimation of the extent that their work was mentally demanding, the amount of concentration and precision required, and whether they worked under time pressure. The follow up assessment occurred 3 years later and indicated that job complexity, independently of age, reduced the risk of developing cognitive impairment.

Wilson et al. (2002) attempted to assess the effect of participation of elderly subjects in cognitively demanding activities as part of the Chicago Health and Aging Project. They looked at seven measures: listening to the radio, reading newspapers, reading magazines, reading books, playing games like cards, checkers, etc. and going to museums. Four years later, the elderly subjects were assessed for AD. The risk of developing incident AD was found to be reduced by participation in demanding activities earlier in life. Furthermore, the greater the level of activity, the lower the risk. By contrast, physical activity was unrelated to the risk of developing AD.

This work collectively indicates that participating in cognitively taxing activities, both work and leisure, are associated with reduced risk of dementia and therefore supports the conclusions suggested from the research on education. However, this work also could be explained by a common factor hypothesis, linking development of dementia and participation in cognitive challenging activities early in life.

Our focus, thus far, is on enrichment early in life. Does cognitive enrichment also have beneficial effects when provided later in life?

Ball et al. (2002) provided volunteer elderly subjects (aged 65–94) with specialized cognitive training in either memory (verbal episodic), reasoning, or speed of processing (visual search). Cognitive skills in these subjects were subsequently followed over a 2-year period. Each of the interventions was found to improve the targeted cognitive ability, but not any of the others. The absence of transfer effects to other functions suggests the training is task specific. However, they did not use an impaired population of subjects, which may have introduced a ceiling effect.

Moore et al. (2001) examined the effects of specialized training on in a sample of 25 mild to moderate Alzheimer's disease patients. The participants were taught strategies to promote memory and recall of specific items, such as names and faces. They found significant improvement, which was maintained for at least 1 month following the intervention. Most of the effects were task specific, but there was also improvement in attention, a more generalized effect. Collectively, these results demonstrate that specialized cognitive training can improve performance on cognitive tests, but that these effects are largely limited to aspects of cognitive function directly linked to the training procedures.

We have previously distinguished between two types of behavioral plasticity linked to cognitive experience (Milgram, 2003). The first pertains to the use of higher-level cognitive abilities over an extended time, which appears to provide a buffer against cognitive decline. The second type of plasticity is more selective and pertains to improvement as a result of specialized training, which serves as procedural learning. These suggestions were largely based on our work with aged dogs. This improvement resulting from specialized training in human subjects, even those with dementia, is an example of this type of plasticity.

Brain imaging studies have provided clues as to the mechanisms of age-associated cognitive decline, and compensatory neural strategies. Anderson and Grady (2001) characterized two types of changes that occur with increased age. The first is an overall decrease in cerebral blood flow in functional networks associated with a given task. Thus, both young and old subjects show increased blood flow during performance of a Wisconsin Card Sorting Task, but the level of increase is higher in young subjects than in old. The second pertains to the spatial extent of activation, where old subjects show more extensive activation than young. These results have been obtained from normal successful agers, and suggest that even normal cognitive aging is associated with some kind of functional reorganization. This provides general support for a cognitive reserve hypothesis, which we now consider.

Both normal and pathological aging are accompanied by an increase in brain pathology, which includes but is not limited to development of amyloid-beta plaques (e.g., Keller, 2006). Further, there is a causal relationship between pathology and cognitive impairment, but in some instances normal cognition is present despite the existence of extensive pathology. The cognitive reserve hypothesis was formulated to account for the appearance of apparently normal cognition in individuals with clear neuropathology (e.g., LeCarret et al., 2005, Whalley et al., 2004). According to the cognitive reserve hypothesis, high educational levels are able to delay the clinical expression of dementia, presumably because of the brain's ability to utilize available neural structures as a backup or reserve.

The cognitive reserve hypothesis largely relies on the assumption that some kind of neuroplasticity linked to cognitive experience enables normal cognitive function despite the presence of brain pathology. This hypothesis, then suggests that cognitive experience early in life affects brain organization later in life. The evidence previously discussed showing functional reorganization MRI changes associated with aging directly supports the hypothesis.

Further evidence that education can influence the response to brain pathology was provided by Dufouil et al. (2003). They found that aged subjects, in general, had high frequencies of cerebral white matter hyperintensities (WMH), a pathological indication of the occurrence of vascular insults. In subjects with lower levels of education, WMH was correlated with lower cognitive performance on a battery of tests. By contrast, subjects with higher education did not show a significant cognitive reduction associated with WMH. They concluded that education was able to protect against cognitive deterioration associated with vascular insults.

The cognitive reserve hypothesis is also supported by evidence that rate of cognitive decline, once AD has been diagnosed, is directly related to educational background and occupational complexity. The greater the educational background, the more rapid the decline; the greater the occupation complexity, the more rapid the decline (Andel et al., 2006). The cognitive reserve hypothesis is also consistent with evidence from brain imaging studies showing recruitment of brain regions, which are not utilized in young subjects, in older subjects.

Jacobs et al. (1993) examined neuron structure derived from Golgi stained materials in subjects differing in amount of education. Dendritic measures showed an increase in length with increased levels of education, suggesting a possible structural basis for the beneficial effects of education.

To summarize, a large literature indicates a relationship between cognitive experience early in life, and cognitive decline associated with aging. The results suggest that cognitive enrichment, as assessed by education and job complexity, has protective effects against the development of dementia. The literature further suggests that these effects are at least partially function specific, with crystallized intelligence and memory processes showing greater protection than fluid intelligence. There is also typically a major confound, namely the mechanisms accounting for the differences in cognitive ability early in life. One study indicated a link between the presence of APOE4 allele and level of education, suggesting that factors linked to intelligence early in life can also contribute to development of dementia. Another related factor that could be important both early and late in life is cholinergic dysfunction.

The ideal way to demonstrate a critical role of cognitive enrichment early in life would be with an intervention experiment in which the early life experience of equivalent groups could be controlled. That is, the experiment would include an enriched group provided with extensive education and complex jobs, while the control group would have no education and menial jobs. Assessment later in life would then indicate the utility of the enrichment procedure. It is not feasible to do this type of study in human subjects. Nor is it possible to do the type of invasive studies that would enable us to best understand the underlying mechanisms. We are limited, therefore, to the use of animal models. This section will discuss the two most widely used animal models, rodents and primates. We will then look at our work on another animal model, beagle dogs.

Rodent studies, as well as studies in other animals, typically do not distinguish between cognitive stimulation and other types of enrichment. For the purpose of this review, we will use cognitive stimulation to refer to protocols that involve demonstrable learning and memory. This includes associative learning tasks as well as mental stimulation such as object exploration or acrobatic training. These should be distinguished from environmental enrichment paradigms, which typically consist of a combination of complex social and sensorimotor stimuli including larger cages, cage mates, toys, tunnels and running wheels (Lambert et al., 2005). This experimental design precludes isolation of any single element, such as cognitive stimulation or exercise, and few studies have examined the effects of cognitive stimulation in the absence of exercise on the brain or on learning and memory.

There is evidence, however, that cognitive experience early in life can have beneficial effects on cognition at later time points. Vicens et al. (1999) found that water maze training at 6 months of age in mice led to improved water maze learning in at 10 months of age. The extent of improvement varied between strains. In a subsequent study, they found similar effects in mice tested at 18 months of age, suggesting that experience might prevent development of age-related learning deficits Vicens et al. (2002). Their results did not explore underlying mechanisms. Lambert et al. (2005) found that cognitive stimulation produced by the presence of cage mates and novel objects increased synaptophysin levels compared to controls in the neocortex, hippocampus and cerebellum but did not improve memory in the water-escape motivated radial arm maze when young mice were used. This study, however, used only young animals. Simply placing novel objects in the cage may not be sufficient to benefit the animal's cognitive function. Rats exposed to a novel toy show a significant increase in hippocampal acetylcholine efflux but only if the animal actively manipulates the object (Degroot et al., 2005).

Acrobatic training consists of repeated trials through a course designed to encourage problem solving and coordination (Lambert et al., 2005). Acrobatic training may be a more effective paradigm for stimulating mental function since the animals need to learn to cross various types of bridges (chains, ropes, wires) and other obstacles to reach a series of platforms. Acrobatic training did not affect memory on the water-motivated radial arm maze or synaptophysin levels in the cerebellum or neocortex in young mice. Buitrago et al. (2004) found that the type of training affects performance on motor skill learning tests. Rats tested using the accelerated rotarod task showed improvements when trained on the rotarod but not when trained with a running wheel. The running wheel improves general locomotor activity but the rotarod training showed that the rats developed a motor strategy specific to the rotarod task by modifying their gait pattern during training.

This type of motor learning (as opposed to the motor activity involved in voluntary or forced exercise) significantly increases synapse formation in the cerebellar cortex of female rats (Black et al., 1990, Kleim et al., 1997, Kleim et al., 1998). Acrobatic training in rats also enhances synaptogenesis in the motor cortex in male rats (Jones et al., 1999). It is also difficult to determine whether morphological brain changes are a result of learning or a consequence of performing the behaviors required to perform the task. Black et al. (1990) reported that motor learning is necessary to induce new synapse formation in the cerebellar cortex and not simply the repetitive use of the synapses during the physical activity. Rats provided with complex visuomotor learning (acrobatic training) form substantial numbers of new synapses in cerebellar cortex, whereas animals given extensive locomotor exercise formed new blood vessels but no more new synapses compared to animals in an inactive control group. Both complex motor learning (acrobatic training) and motor activity affect the expression of BDNF in the motor cortex and cerebellum (Klintsova et al., 2004). The expression of its receptor TrkB in the cerebellum changes in response to both motor conditions but in the motor cortex TrkB expression was only affected by the acrobatic training (Klintsova et al., 2004). These changes persisted longer in the acrobatic training condition. BDNF is involved in synaptogenesis (Vicario-Abejon et al., 1998). Acrobatic training also leads to an increase in Fos-positive cells in the motor cortex of rats, which may underlie synaptogenesis (Kleim et al., 1996).

Another factor that can also affect brain structure is social grouping. Housing rats in groups produce significantly greater brain weights and brain enzymes (Welch et al., 1974). The addition of inanimate stimulation to social grouping produced even greater effects on brain weight and nucleic acids and enzymes than social stimulation alone (Rosenzweig et al., 1978).

Rodents are particularly useful for these studies because their short life span readily allows longitudinal studies. However, rodents do not naturally develop the kinds of pathologies associated with human aging. A further problem in using laboratory rodents to model the effect of cognitive stimulation stems for differences in baseline levels of enrichment. Laboratory rats, of necessity, in control experiments live a sedentary existence, very different from humans. Are the changes seen a result of enrichment modifying brain structure, or are the changes a result of degenerative effects of living in an impoveraged environment? Future experiments should be able to take this problem into consideration. In an ideal experiment, an enriched environment would be provided to both the controls and the treatment group. We could then look at possible beneficial effects of adding a program of cognitive enrichment.

Several studies have attempted to develop enrichment toys and playgrounds incorporating perches, toys, mirrors, music, grassy areas, and social groups (Bayne et al., 1992, Bayne et al., 1993, Howell et al., 2003, Brent et al., 1991, Line et al., 1990, Eichberg et al., 1991, Bryant et al., 1988, Videan et al., 2005, Harris and Edwards, 2004) for use in non-human primates. Enriched environments in non-human primates generally reduce behavioral pathology such as stereotypes, agitation, aggression, abnormal behaviors, self-directed behaviors, and self-abusive behaviors. These behaviors typically reappear when the animal is removed from the enrichment area (Bryant et al., 1988). Environmental enrichment studies typically do not use non-human primates. Attempting to group house previously single-caged rhesus monkeys can lead to aggressive conflicts and considerable risks and only works if the cohabitants are compatible (Reinhardt, 1991, Washburn and Runbaugh, 1992). Simple enrichment toys are associated with a rapid decline in use by non-human primates as well (Bayne et al., 1993).

Most enrichment is associated with improved physical and psychological well-being (Eichberg et al., 1991). Washburn and Runbaugh (1992) found that training individually housed monkeys to manipulate a joystick on a computerized test system is a sufficient means of providing environmental enrichment. Performance levels are not affected by pair-housing and this cognitive challenge is as preferred by rhesus monkeys as social opportunities (Washburn et al., 1994).

Over the past 16 years, understanding cognitive aging in the dog has been the main focus of our lab. Our interest in dogs stems largely from parallels between canine and human aging and AD, including specific, progressive and early cognitive deficits in the domains of memory, complex learning and executive function. The progression of the cognitive deficits eventually spans several cognitive domains, similar to late-stage dementia. In addition, aged dogs develop neuropathological changes, including an accumulation of amyloid-beta protein with a deposition pattern and peptide sequence similar to that seen in humans (Cummings et al., 1996). Aged dogs also show progressive increases in oxidative damage (Head et al., 2002), cortical atrophy, ventricular enlargement (Tapp et al., 2004) and age-associated decreases in neuronal numbers and cholinergic receptors. Like humans, cognitive decline associated with aging in the dog is also sensitive to cognitive enrichment earlier in life. This evidence is reviewed here, along with data suggesting possible underlying mechanisms.

We have recently completed a 3-year longitudinal study of cognitive aging in dogs that focused on two interventions, behavioral enrichment and antioxidant supplementation (see Milgram et al., 2005). At the start of the study the dogs were divided into four cognitively equivalent treatment groups: (1) controls; (2) antioxidant supplementation only; (3) behavioral enrichment only; (4) a combination of antioxidant supplementation and behavioral enrichment. There were two findings relevant to this review. The first was behavioral, and the second neuroanatomical.

The behavioral enrichment condition consisted of pair-housing, providing extra stimulation in the form of toys and increased opportunity for behavioral activity. The enriched group also received a program of regular cognitive testing over the duration of the 3-year period. The control subjects, by contrast, were provided cognitive testing over a limited period at the beginning of each year of the study. Because the housing environment consisted of indoor/outdoor runs, all animals were given the opportunity for extra exercise, and because most control animals were also pair housed, it is likely that the primary difference between the groups was in cognitive enrichment.

The behavioral assessment protocol carried out after 1 and 2 years, involved testing the dogs on tasks of comparable difficulty, including discrimination and reversal learning and also on a delayed non-matching to position task (DNMP), which provides measures of both visuospatial learning and memory (Chan et al., 2002).

On the discrimination and reversal learning tasks, the groups, with the combined food supplementation and cognitive enrichment showed the highest level of performance. The cognitive enrichment alone group and the antioxidant supplementation alone followed this group, with the control groups showing the poorest overall performance, followed this group. On the DNMP task, by contrast, cognitive enrichment alone had a smaller non-significant effect. We attribute the differential cognitive effects of the interventions to the nature of the cognitive enrichment provided, which largely focused on various kinds of discrimination learning paradigms.

More recently, we have also obtained cross-sectional evidence. We typically obtain aged dogs from either of two types of sources. The first are class A breeders, which keep these dogs for breeding until they become old. The second are smaller breeders, who buy the dogs from their owners as retired hunters. The retired breeders have a known background, having spent their entire lives in the same facility. It is very likely that these animals have very limited cognitive experience, and may never have had to work for food in their lives. By contrast, the hunters almost certainly have had experience learning to work for reward, and likely have lived in varied environments.

When initially trained on our cognitive testing paradigm, the retired breeders performed significantly poorer than the hunters on a reward approach learning task, involving learning to approach food. Reward approach learning is a procedural learning test (Adams et al., 2000) and can be considered analogous to measures of crystallized intelligence in humans. Presumably, the retired hunters, show better performance because of they have not always relied on their master for food. These differences, however, proved to be task specific. After completing our standard pretraining protocol, dogs from both groups were tested on a reversal learning task, which provides a measure of executive function and subsequently on a delayed-non-matching to place task, which tests visuospatial learning and memory. On these tests, as illustrated in Fig. 1, the differences between the two groups were small and non-significant. These data can be construed to support the suggestion arising from the human research that cognitive experience early in life improves measures of crystallized intelligence, to a greater extent than it does fluid intelligence.

The results of this study are only suggestive, and should ideally be replicated under more controlled conditions. One factor that might have impacted the data is the age of the two different cohorts of dogs: the retired breeder group was younger than the hunters, which would be expected to have benefited the breeders, as we have previously shown that the tests of executive function and working memory are highly sensitive to age (Tapp et al., 2003). Like all cross-sectional studies, cohort effects potentially confound the results. We have also examined a group of retired breeders from the same source in a second facility with lower levels of non-cognitive enrichment. The group in the facility with less enrichment performed significantly poorer on all of the aforementioned tests, which supports the claim above that animal models of human cognitive function should be tested in high levels of enrichment, which is standard for humans.

Neuron loss within the hippocampus and underlying cortex occurs as a function of age in humans and may contribute to age-associated cognitive decline. We first tested the hypothesis that neuron loss occurs in the canine model of brain aging. The total unilateral number of neurons in the canine entorhinal cortex and subdivisions of the hippocampus were estimated using stereological methods and systematic sampling techniques. We focused on the hippocampus and entorhinal cortex of five old and five young dogs because previous studies show these regions to be vulnerable to aging and the progressive accumulation of neuropathology. There was a significant loss of neurons with age in the hilus (∼30%) of the hippocampus but not in the remaining hippocampal subfields or entorhinal cortex (Siwak et al., 2005).

This loss could be a result of decreased neurogenesis in the hippocampus, increased cell death with age or a combination of both factors. We recently completed another stereological study and found an age-related decrease in the number of doublecortin positive neurons, a marker for newly formed neurons, in the dentate gyrus of the canine hippocampus. This suggests that decreased neurogenesis may contribute to the loss of neurons with age but we have not yet examined the rate of cell death in these regions. The hippocampus is involved in learning and memory processes that decline with age and neuron loss could be a contributing factor in cognitive dysfunction. Variability in the extent of neuron loss could account for the severity of the decline or the individual variability in cognitive abilities.

In the longitudinal study previously described, we were able to obtain brain tissue from half the animals, and examine measures of both cell death and brain pathology. We initially focused on the effects of both the enrichment and antioxidant interventions on neuron number in the hippocampus of the aged dogs. The number of hilar neurons in aged dogs was modified by the behavioral enrichment paradigm but not with the antioxidant supplementation compared with age-matched controls (Siwak et al., 2005); enriched dogs had more hilar neurons (∼18%) compared to the non-enriched group. The enrichment could be increasing neurogenesis or reducing cell death in the dog brain. We found no difference in the number of doublecortin positive neurons between the enriched dogs and age-matched controls. This suggests that the enrichment is neuroprotective, but further research is necessary to confirm this hypothesis. Hilar neuron number also correlated with performance on some of the cognitive tests. This could be one mechanism through which the enrichment paradigm was able to improve cognitive function in the aged dogs.

In the analysis of neural pathology, one of our main focuses was on the development of amyloid-beta pathology. By contrast, we found that an antioxidant enriched diet did not modify neuron numbers in the hilus of the hippocampus but did lead to a reduction in amyloid-beta accumulation in the parietal cortex (Pop et al., 2003). However, beta-amyloid pathology in the frontal cortex, which typically accumulates at ages prior to when we implemented the treatment, remained relatively unaffected. This suggests that the antioxidant diet can slow new amyloid deposition but not reverse existing pathology. One mechanism underlying reduced amyloid-beta accumulation in antioxidant treated dogs is by increasing the activity of alpha-secretase, which cleaves the amyloid precursor protein but prevents the production of beta-amyloid (Pop et al., 2005). Interestingly, the behaviorally enriched animals showed no evidence of amyloid reduction relative to untreated controls suggesting that neurons in the brains of these treated animals are healthy despite being surrounded by pathology. This has significant implications for the human clinic; individuals with Alzheimer's disease with clinical symptoms most likely already have extensive beta-amyloid pathology. Thus, behavioral enrichment improves neuronal function independently of amyloid pathology and may therefore be very useful for improving cognition in this population. In combination, our data in dogs suggests that an antioxidant diet and behavioral enrichment each improve neuronal function but through independent molecular pathways.

Collectively, this evidence suggests that behavioral enrichment does not act to prevent the development of brain pathology, at least amyloid pathology, but rather through other means that might allow the animals to better compensate for the occurrence of pathology.

Section snippets

Conclusions and future directions

Thus far, the most compelling and comprehensive indication of the neuroprotective effects of cognitive enrichment comes from studies of human subjects, which lead to several general conclusions. First, cognitive enrichment, which is manifested in education, job complexity, and/or leisure activities, has beneficial effects, which helps preserve several cognitive functions, and can delay the development of dementia. Second, these positive effects, impact crystallized intelligence to a greater

Acknowledgements

This research was supported by grants to the authors from National Institute of Aging (grant AG12694), Natural Sciences and Research Council of Canada (NSERC A7659) and by CanCog Technologies of Toronto, Ontario, Canada.

References (62)

  • K.P. Riley et al.

    Early life linguistic ability, late life cognitive function, and neuropathology: findings from the Nun Study

    Neurobiol. Aging

    (2005)
  • M.R. Rosenzweig et al.

    Social grouping cannot account for cerebral effects of enriched environments.

    Brain Res.

    (1978)
  • M. Sarter et al.

    Developmental origins of the age-related decline in cortical cholinergic function and associated cognitive abilities

    Neurobiol. Aging

    (2004)
  • P. Vicens et al.

    Previous training in the water maze: differential effects in NMRI and C57BL mice

    Physiol. Beh.

    (1999)
  • P. Vicens et al.

    Effects of early spatial training on water maze performance: a longitudinal study of mice

    Exp. Gerontol.

    (2002)
  • B.L. Welch et al.

    Isolation, restrictive confinement or crowding of rats for one year. I. Weight, nucleic acids and protein of brain regions

    Brain Res.

    (1974)
  • L.J. Whalley et al.

    Cognitive reserve and the neurobiology of cognitive aging

    Aging Res. Rev.

    (2004)
  • M.S. Albert et al.

    Predictors of cognitive change in older persons

    Psychol. Aging

    (1995)
  • R. Andel et al.

    The effect of education and occupational complexity on rate of cognitive decline in Alzheimer's patients

    J. Int. Neuropsychol. Soc.

    (2006)
  • N.D. Anderson et al.

    C.L. Functional imaging in cognitively intact aged people

  • K. Anstey et al.

    Education, activity, health blood pressure and apolipoprotein E as predictors of cognitive change in old age: a review

    Gerontol.

    (2000)
  • K. Ball et al.

    Effects of cognitive training interventions with older adults: a randomized controlled trial

    JAMA

    (2002)
  • K.A. Bayne et al.

    Kong toys for laboratory primate: are they really an enrichment or just fomites?

    Lab. Anim. Sci.

    (1993)
  • K.A. Bayne et al.

    Evaluation of the preference to and behavioral effects of an enriched environment on male rhesus monkeys

    Lab. Animal Sci.

    (1992)
  • J.E. Black et al.

    Learning causes synaptogenesis, whereas motor activity causes angiogenesis, in cerebellar cortex of adult rats

    Proc. Ntl. Acad. Sci.

    (1990)
  • H. Bosma et al.

    Mental work demands protectg against cognitive impairment: MAAS prospective cohort study

    Exp. Aging Res.

    (2003)
  • L.D. Brent et al.

    Evaluation of a chimpanzee enrichment enclosure

    J. Med. Primat.

    (1991)
  • C.E. Bryant et al.

    Effects of different environmental enrichment devices on cage stereotypies and autoaggression in captive cynomolgus monkeys

    J. Med. Primat.

    (1988)
  • S.M. Butler et al.

    Age, education and changes in the mini-mental state exam scores of older women: findings from the Nun Study

    J. Am. Geratric. Soc.

    (1996)
  • A.D.F. Chan et al.

    Visuospatial Impairments in aged canines. Canines: the role of cognitive-behavioral flexibility

    Behav. Neurosci.

    (2002)
  • H. Christensen et al.

    Education and decline in cognitive performance: compensatory but not protective

    Int J. Geriatr. Psychiatry

    (1997)
  • Cited by (0)

    View full text