Section V
Alzheimer’s disease and natural cognitive aging may represent adaptive metabolism
reduction programs
Jared Edward Reser M.A.
Department of Psychology
University of Southern California
jared@jaredreser.com
Abstract
The present article examines several lines of converging evidence suggesting that the slow and
insidious brain changes that accumulate over the lifespan, resulting in both natural cognitive aging and
Alzheimer’s disease (AD), represent a metabolism reduction program. A number of such adaptive
programs are known to accompany aging and are thought to have decreased energy requirements for
ancestral hunter-gatherers in their 30s, 40s and 50s. Foraging ability in modern hunter-gatherers
declines rapidly, more than a decade before the average terminal age of 55 years. Given this, the
human brain would have been a tremendous metabolic liability that must have been advantageously
tempered by the early cellular and molecular changes of AD which begin to accumulate in all humans
during early adulthood. Before the recent lengthening of life span, individuals in the ancestral
environment died well before this metabolism reduction program resulted in clinical AD, thus there was
never any selective pressure to keep adaptive changes from progressing to a maladaptive extent.
Aging foragers may not have needed the same cognitive capacities as their younger counterparts
because of the benefits of accumulated learning and life experience. It is known that during both
childhood and adulthood metabolic rate in the brain decreases linearly with age. This trend is thought
to reflect the fact that children have more to learn. AD “pathology” may be a natural continuation of this
trend. It is characterized by decreasing cerebral metabolism, selective elimination of synapses and
reliance on accumulating knowledge (especially implicit and procedural) over raw brain power
(working memory). Over decades of subsistence, the behaviors of aging foragers became routinized,
their motor movements automated and their expertise ingrained to a point where they no longer
necessitated the first-rate working memory they possessed when younger and learning actively.
Alzheimer changes selectively and precisely mediate an adaptation to this major life-history transition.
AD symptomatology shares close similarities with deprivation syndromes in other animals including
the starvation response. Both molecular and anatomical features of AD imitate brain changes that have
been conceptualized as adaptive responses to low food availability in mammals and birds. Alzheimer’
s patients are known to express low overall metabolic rates and are genetically inclined to exhibit
physiologically thrifty traits widely thought to allow mammals to subsist under conditions of nutritional
scarcity. Additionally, AD is examined here in the contexts of anthropology, comparative neuroscience,
evolutionary medicine, expertise, gerontology, neural Darwinism, neuroecology and the thrifty genotype.
Background
Alzheimer’s disease (AD) is a central nervous system disorder that gradually increases in severity with
age resulting in memory loss, behavioral changes and a decline in thinking abilities. Connections
between individual neurons are destroyed in the AD brain due to the formation of neurofibrillary tangles
(i.e., tangled strands of neural fibers within the bodies of neurons) and senile plaques (i.e., masses of
dying neural material containing toxic protein called beta-amyloid) which obstruct, damage and lead to
the eventual death of many neural cells [1]. The most pronounced features of the AD state are thought
to be decreased cerebral blood flow and decreased cerebral metabolism [2,3] due to the fact that
dead neurons do not require oxygen or glucose [4]. The tangles and plaques responsible for this
accumulate in all adults advancing in age, but do so at an accelerated rate in those that will develop
AD [4]. All humans begin to develop the neurological markers of AD during their early 20s and continue
to do so throughout life, most towards clinically irrelevant degrees [5]. But why would these markers
present in everyone? How could natural selection have allowed them to become so invasive and
ubiquitous if they did not hold some sort of evolutionary significance?
Humans are not the only species that develop plaques and tangles and are also not the only species
whose behavior can be affected by them [6]. AD-like neuropathology is known to occur in the species
of several mammalian orders, including many species of primates, but has the potential to become
more severe in neuropathological presentation and in the degree of functional impairment in humans
[7]. The wide phylogenetic spread suggests that it responds to some selective pressure. Furthermore,
the fact that a wide variety of mammals exhibit AD-like neuropathology suggests that the first mutations
responsible for it may have arisen tens of millions of years ago. It will not be easy to determine exactly
when or how the genes responsible for AD evolved; however, the present article will focus on exploring
why. Framing AD as part of a strategy to save calories during the later stages of life may explain many
facets of the disease, such as (1) why it shares many similarities with known responses to starvation,
(2) why it resembles a neuroecological change, (3) why it selectively affects specific neuroanatomical
areas and not others, (4) why it presents comorbidly with the metabolic syndrome, (5) why it has been
tied genographically with the thrifty genotype, and (6) why such a psychiatrically conspicuous syndrome
was not eradicated by natural selection.
Presentation of the hypothesis
It is thought that analyzing disease states from an evolutionary perspective can ultimately do much to
inform and influence medical intervention strategies [8]. The present article will employ this approach
with Alzheimer’s disease (AD) emphasizing that, because susceptibility genes for AD are so prevalent
in human populations, the traits that characterize preclinical (or presymptomatic) AD must have been
naturally selected. Because the average life span of contemporary human hunter-gatherers is around
55 [9,10], it is thought that prehistoric man would have lived nearly as long. Given that clinical AD is only
rarely diagnosed before age 55 [11], it is clear that individuals would rarely have lived long enough to
develop AD in our evolutionary past. Natural selection then, could only have acted on the preclinical or
prodromal AD phenotype and, as in other disorders of aging [7], would not have had the opportunity to
curtail the detrimental effects, which of course appear late in life.
Some theorists assume tacitly that AD presents too late in reproductive life to have been exposed to
negative selection [12]. They realize that diseases that arise after individuals become infertile cannot
limit the total number of offspring produced, and that evolution has little way of excising such diseases.
However, many facts about preclinical AD indicate that the genes responsible for it must have been
exposed to blatant selective pressure, well before their bearers reached reproductive senescence.
Particularly, neuropathological changes begin in the early 20s and usually constitute a “heavy load” 10-
20 years before the first behavioral symptoms of marked cognitive decline surface [13]. Several forms
of neurophysiological and intellectual decline have been shown to begin in the first few decades in
individuals that will develop AD [14]. In fact, well controlled studies have shown that individuals that
carry susceptibility genes for AD exhibit lower levels of intellectual functioning throughout life and are
more likely to drop out of high school by age 15 when compared to their matched peers [15]. In other
words, because the genes that cause AD create conspicuous neurological and behavioral
characteristics that present during reproductive age, they could not have been invisible to evolutionary
forces.
Additionally, some evolutionary theorists find it hard to imagine that AD could have been adaptive
because of low ability for autonomy and self-sufficiency seen in AD patients. Despite some
impairment, individuals with preclinical AD are regularly capable of both domestic autonomy and
professional achievement [5]. Importantly, AD usually only creates conspicuously maladaptive behavior
in people that live far past the average terminal age, which is around 55, in foraging societies [16].
Moreover, medical advances in the last few centuries have artificially extended life by several decades,
and most experts believe that this has given AD neuropathology the time necessary to develop to an
injurious extent that would have been impossible in prehistoric times [17]. In other words, AD in its
advanced, debilitating form may simply be the unnatural progression of natural brain aging changes,
which would have rarely occurred in shorter-lived, ancestral populations.
It is important to emphasize that, on a cellular level, all of the changes seen in Alzheimer’s disease are
also found in normal aging adults [18]. The neurofibrillary tangles [19], senile plaques [16] and
neurotransmitter changes [20] are each present in aged-matched normals, with the same regional
preferences, but simply to a lesser extent [11]. It has been estimated that even those elderly individuals
that show the least cognitive impairment in old age would eventually be diagnosed with AD if they were
able to live to around 130 [21]. The genes responsible for this “inevitable transformation” exist in every
one. This frames AD-like neuropathology as a continuous, polymorphic cognitive strategy, where
clinical AD represents the extreme side of a continuum. Other researchers have also concluded that
AD is not a qualitatively distinct, abnormal entity, but is instead at one extreme end of a spectrum of
capability in old age within which a great deal of natural variation exists [22]. Like many other
polygenic, continuous traits the mutations responsible for pronounced AD were likely maintained by
balancing selection (specifically, environmental heterogeneity). In other words, they were kept in our
gene pool because, as environmental resource conditions fluctuated, different genetic polymorphisms,
or “multiple alternate alleles” were favored.
Today, the costs of AD are well-documented [23], but the defensive manifestations may be hidden
because of discrepancies between our modern and ancestral environments. Many traits that are clearly
maladaptive in the present are now thought to have been adaptive in the ancestral environment [24]
and this situation is known as an “environmental mismatch.” The science of evolutionary medicine
attempts to identify, analyze and explicate these traits [25]. Researchers have identified many
“pathological” conditions such as atherosclerosis, cardiovascular disease, cystic fibrosis, diabetes
mellitus, and obesity, and have helped to show that despite the stigma today, genetic susceptibility to
these diseases would have conferred adaptive benefits in prehistoric times [8]. Williams and Nesse
[26] suggest that certain criteria must be met to conclude that a “disease” may have been adaptive in
the past: it should be relatively prevalent, heritable, and susceptibility should vary within the population.
AD certainly meets the first three criteria [18], but now it is important to show that the fourth criterion
can be satisfied – that the benefits associated with the condition must have outweighed the costs.
A handful of articles have attempted to explain the existence of AD in terms of antagonistic pleiotropy
[12,17], however, AD and the underlying neuropathology have not been analyzed in terms of
evolutionary medicine and have not been firmly reconciled with evolutionary theory. Perhaps it can be
argued that AD is utterly pathological, has no advantageous qualities and no natural history. This article
will espouse an opposite view and explore the assertion that AD has been naturally selected and
represents a preservative ecological strategy. Many articles in the last few decades have analyzed
various forms of psychopathology (e.g., anxiety, bipolar disorder, depression, obsessive compulsive
disorder) in terms of evolutionary theory and evolutionary medicine [27,28]. This burgeoning area of
research is often referred to as “evolutionary psychopathology.” The present author has written articles
analyzing various forms of neuropathology using this approach and has called this area “evolutionary
neuropathology [29,30,31]”. Using this framework, the present article is prepared to explore many
congregating sources of evidence that strongly suggest that AD is highly explicable under evolutionary
theory.
Neural Darwinism, skill and specialization
Clearly, prehistoric foragers with preclinical AD would not have the same cognitive capacity to learn
from their environment that they had when they were younger. It is probable; however, that they would
no longer need this capacity. For reasons to be explored in this section, the cognitive symptoms of
preclinical AD may have complemented the aging process on the ancestral, African savannah.
A conspicuous trend has been documented for decades in the study of neuron number: many animals
overproduce neurons in their youth and then, later in life, lose the unused, inefficient ones [32]. This kind
of neuron loss is known to be controlled by programmed cell death, or apoptosis, which is also a major
contributing factor to the neuropathology in AD [33]. One function of neuron death in aging animals is to
remove neurons that have not made useful or sufficiently numerous connections [34,35,36,37]. For
example, in mammalian populations, neural degeneration in segments that control necessary functions
such as limb movement is very rare [38]. It is more common, however, in less behaviorally critical
neural tracts. A young mammal’s cortex is full of neural tracts that may never be critical, but it is not
clear which will be needed and which will not until the animal has acquired life experience [39]. Several
researchers have associated this process of “economization by elimination” with strong positive
increases in reproductive fitness [39,40,41,42]. A good deal of ethological evidence supports the idea
that as animals of many different phyla mature and age, intensive learning may become less important
with time [43]. In fact, one major neurodevelopmental trend seen throughout the human lifespan is a
compelling example of this. The metabolic rate found in the brains of children is far higher than in those
of young adults, which are, in turn, far higher than in those of middle-aged adults [44].
In the first decade of human life, the cerebral cortex undergoes dramatic fluctuations in energy
consumption. The metabolic rate of the brain increases rapidly from birth and begins to reach adult
levels by age two. From four years of age until nine, the amount of energy spent in a child’s brain far
exceeds that of an adult [45]. During this period, the brain’s neural architecture is characterized by
“hyperconnectivity,” where neurons of the cortex have formed many more connections (synapses) than
will be kept. Many of these will be selectively eliminated [46]. At ten years of age, the brain’s metabolic
rate begins to decline until the late teens, at which point the levels of glucose utilization have reached
adult values. This pattern of metabolic decline persists naturally, in everyone, until death [45].
Throughout the lifespan, neurons and synapses that are not used are “pruned,” limiting behavioral
plasticity and constraining what can be learned in the future [47]. Only the connections that are utilized
are maintained, and this inevitable process is responsible for the dramatic windows for learning (e.g.,
second languages and musical instruments) that close soon after adolescence [46].
The sharp diminishment in cerebral metabolism in young adulthood is currently conceptualized in the
literature as an evolutionarily mediated response to changes in life-history dynamics [46], but modern
AD researchers appear surprised that further reductions occur with advancing age. These reductions,
even in late life, should be seen as part of a natural process of continuing development. Young children
are small and their metabolic demands are met by their parents, yet they need to learn rapidly in order
to become ecologically competent [48]. They can afford a high cerebral metabolism because they
benefit so greatly from the incessant thinking and learning that accompanies it. Once the individual
becomes an adult though, they need not expend quite so much energy actively learning and analyzing.
The adaptive value of extracting large amounts of information from their environment and carrying it
with them through time has decreased. This is because, in adulthood, individuals should have already
internalized much of the cultural and ecological information that they will need.
As aging progresses, it is still necessary to be able to learn new mental maps, new language skills and
new techniques. However, because so much of what is learnable is already known, the usefulness of
this capacity must continue to diminish with increasing age. If an animal is able to survive to reach full
adulthood, then the functional conceptualizations that it made during development have been working
and the animal must have internalized information in a way that is conducive to survival. In this case it
should not be necessary to radically change a mental set that has proven to be effective. The profound
neuron loss with age to the nucleus basalis [49], the gateway module for new learning in the brain, may
help to ensure that a mature organism cannot reprogram its tried and true behaviors with new, untested
learning. It is accepted that elaborate, new memories have a tendency to interfere with old ones in both
laboratory and ecological contexts [50].
“Academic learning is often explicit: professors point out the things to be learned, and students try their
best to memorize them. But most ordinary human learning is probably implicit. A hunter may teach
young people how to track an animal, or how to kill and skin it. Most of the time such practical activities
can be taught more easily by modeling than by explicit labeling. Many of the subtleties of hunting and
gathering may not even have names”
-Moscovitch et al., 2007 [51]
As the quote above explains, in ancient times, a hunter-gatherer’s cognitive resources were probably
used (much more than they are in modern times) to analyze, refine and coordinate muscle activity. It is
an interesting observation that most AD patients can walk, gesticulate, inflect, intonate, gesture, and
move perfectly fine. This is due to the fact that their procedural memory, the ability to learn and
implement behavioral skills at an automatic, unconscious level, remains very much intact [52].
Interestingly, the structures responsible for procedural memory, the striatum, putamen and caudate
nucleus largely escape neuropathological load in old age [53]. As such, patients with AD can exhibit
severe episodic, semantic and working memory deficits, but because of preserved procedural
memory, are usually able-bodied and retain many of the habits, skills and implicit memories that they
have honed during their lifetime [5]. In fact, most individuals with AD who were athletes can continue to
perform admirably, usually only impaired by physical aging (e.g. muscle tone and cardiovascular
endurance) but not mental aging [54]. If reflexes, motor praxes and coordination of movement are
generally proficient in individuals with pronounced AD in their 90s, then surely they would have been
adequate in our ancient, preclinical ancestors in their 30s, 40s and 50s.
The changes during cognitive aging tend to preserve crystallized intelligence (i.e., the ability to use
known information to instruct behavior), over fluid intelligence (i.e., the ability to use new information to
instruct behavior) [55]. Moreover, it has been shown that even though general IQ score remains
somewhat stable, crystallized intelligence grows rapidly and fluid intelligence declines rapidly between
15 and 60 years of age [56]. The increases in crystallized intelligence are thought to correspond with
increased number of life experiences, and the decreases in fluid intelligence are thought to correspond
with decreasing metabolic output of the brain with age. For these reasons an aging individual may be
less likely to benefit from constant learning and concerted analysis and more likely to benefit from
being vigilant for those stimuli that elicit known behavior patterns. It seems clear that the accretion of
crystallized intelligence (knowledge) makes it so that fluid intelligence (deliberation) is less necessary.
After all, it is far less costly to inform behavior using neural connections that have already been made
than to go through the trouble of creating new connections. As Matt Ridley [57] states,
“Experience causes the unnecessary connections to wither away and thereby turns the brain from a
general to a specific device. Like a sculptor chipping away at a block of marble to find the human form
within, so the environment strips away the surplus neurons to sharpen the skills of the brain.”
Brain development happens throughout life and is characterized from even a very early stage by two
complementary processes, selection and instruction [58]. Both processes underlie learning and
memory and are fundamental to cognition as we know it [59]. In selection, the synapses and neurons
that are utilized most frequently are “selected” to be preserved, and those that are not used degenerate
[60]. This process of neural selection, also called “neural Darwinism,” is thought to be nearly as
important as “instruction,” which stands for the family of cellular and molecular processes that
accomplish long-term potentiation [57]. The main difference between selection and instruction is that
the latter necessitates biological resources. Preclinical AD can be viewed as a strategy that relies
heavily on selection rather than on instruction – emphasizing a cheap and metabolically thrifty way to
program behavior. The importance of metabolic thrift during human aging is what we turn to in the next
section.
Advancing age and the threat of starvation
Data from physical anthropology have shown that, in all foraging groups studied, elderly hunter-
gatherers implement low-yield foraging strategies and procure food (measured in calories) at a rate
that is far below that of younger foragers [61]. These aging individuals have more difficulty procuring
foodstuffs than their younger counterparts do because, due to natural senescence, they can no longer
meet the associated physical and athletic demands [62]. Unlike any other species, however, human
foragers engage in food sharing, which allows the elderly individuals to live longer than they would be
able to if they were forced to meet 100 percent of their own metabolic demands [62]. Traditionally, the
older individuals in foraging groups seek out food on their own, but also rely heavily on younger
individuals to supplement their diet [61]. This is an imprudent strategy because, during times of
hardship when food was scarce, they would most certainly be threatened by starvation. We know that
the Pleistocene era (roughly 2 million to 12 thousand years ago) was marked by frequent, prolonged,
dry spells [63] and consequent widespread nutritional scarcity that continually threatened our ancestors
[64]. It is clear that the demographic most susceptible to starvation during tough times, the aged, would
have benefited the most from metabolic penny-pinching.
Food production, in total calories attained, in male hunter-gatherers increases dramatically every year
between the ages of 16 and 25; this reflects the long learning curve for hunting ability [65]. The ratio of
production to consumption (i.e., catching versus eating) hits the maximum at 25. At this point the
individual is able to procure more food than it eats, and it shares the leftovers. This state of self-
sufficiency plateaus and remains stable as late as age 45, after which time it drops precipitously for
many years until death [61]. Food production ability eventually drops below food consumption resulting
in insufficiency. This race for calories against the clock- a very foreign concept today even in
developing nations- explains why aging humans desperately needed to conserve energy resources. In
contrast, chimpanzee subsistence during aging is very different. Indeed, chimpanzees live their entire
lives without food production ever dropping below consumption [61]. Because young chimpanzees don’
t share food with the older ones, old chimps, like most other wild animals in general, starve to death
relatively soon after their food production ability begins to decline [66]. Human hunter-gatherers on the
other hand exhibit a pronounced drop in food consumption – allowed of course by a diminishment of
metabolic rate – that helps them to survive the drop in production [62].
Because elderly foragers cannot gather enough calories to meet the energy requirements of a younger
forager, we can infer that they would have benefited from particular metabolic alterations that helped
them cut down on energy expenditure. Fittingly, decades worth of gerontological research shows that
elderly humans exhibit lower resting metabolic rates and that they accomplish this by down-regulating a
number of physiological systems resulting in lower muscle mass, lower thyroid levels, lower growth
hormone levels, lower testosterone levels and insulin resistance [67,68]. These physiological changes
that occur with age suggest that natural selection is responsible for selecting our species for
“advanced-age thrift,” allowing elderly individuals to survive on smaller amounts of food [67]. Again,
each of these thrifty metabolic changes that occur with age can be interpreted as pathological in the
present yet adaptive in the past, and clearly represent a type of “senile plasticity” that permitted
metabolic frugality [69]. This plasticity towards energy conservation presents so dependably in various
tissues and organs that we might expect it to generalize to the brain as well.
Testing the hypothesis
Neuroecology and Alzheimer’s
A growing body of literature has shown that neural tissue is highly metabolically expensive and that
species with large brains have been forced to negotiate tradeoffs leading to compensatory
adaptations in various organ systems [70]. Considering that humans spend 20-25 percent of their total
resting energy budget in their brains alone – most primates spend between 8-9 percent [71] – it is
clear that, as highly encephalized primates, humans must have been forced to make many inventive
compromises [72]. This line of reasoning makes it seem sensible that any excusable opportunity to
attenuate the number of calories consumed by the brain would increase an individual’s chances of
surviving periods of prolonged scarcity.
The emerging discipline of neuroecology has amassed a great deal of data (on an assortment of
species, avian and mammalian) that demonstrate that, because the brain is very costly to maintain,
specific brain regions can respond plastically to the environment, lowering energy consumption in
response to deprivation and adversity [73]. It has been shown that many species have this adaptive
ability to attenuate energy metabolism in the neocortex and hippocampus [74]. The animals have been
naturally selected to accomplish this despite the fact that, due to the associated cognitive
disadvantage, they are forced to employ less cognitively rigorous and lower yielding foraging
strategies [75,76]. In fact, studies have shown that the hippocampus can adaptively vary both metabolic
rate and volume with ecological rigor [73,76]. Birds and mammals have been shown to decrease
energy expenditure in the hippocampus in response to environmental deprivation [77], season of low
food availability, and decreased need to forage [76]. These findings have been interpreted widely and
emphatically as examples of protective changes that can significantly increase the animal’s ability to
survive periods of nutritional scarcity [73]. Fascinatingly, the hippocampus is known to show, by far, the
most severe decrement in metabolism when compared with other brain regions in AD [78,79,80].
Hippocampal atrophy and glucose hypometabolism in the neocortex are the two most well documented
findings in AD [81], even in the early stages [82]. In fact, the postmortem diagnosis of AD concentrates
primarily on the extent of neurodegeneration in the fronal cortex and hippocampus [83]. The fact that
the hippocampus and the neocortex are the most affected areas in both the naturally aging brain and
the AD brain [84] further suggests that the “neuropathology” represents an adaptive neuroecological
change. This seems especially likely given that old age is strongly associated with a diminished
foraging capacity in human hunter-gatherers [62]. The dentate gyrus is a subregion of the hippocampus
that is highly ecologically significant in that it is the main area in the brain that generates new neurons
throughout life. It is also the area of the hippocampus that is most damaged in normal aging and AD
[5]. Perhaps there are unexplored ecological implications here! Furthermore, the hippocampus and the
neocortex are known to have exhibited highly accelerated growth during human evolution [49] and thus
the age-associated atrophy they exhibit in AD may signify that aging individuals can benefit from
shifting towards more rudimentary, less cognitively rigorous, food extraction methods reminiscent of
those used by our evolutionary predecessors.
The neuroecological literature informs us that hippocampal hypometabolism in animals invariably
results in an alternate, less rigorous, ecological strategy and, like AD, is often accompanied by other
physiological modes of energy conservation [74]. Overall cerebral metabolism in AD patients is much
lower than in age-matched normal subjects [3], and the degree of severity of AD has been shown to be
positively correlated with decreases in cerebral metabolism [2]. All of these facts are consistent with
our characterization of AD “neuropathology” as a subtle but progressive neuroecological program that
has been engineered by natural selection.
The selectivity of neuropathological changes in Alzheimer’s
It is known that neurodegeneration in AD is not at all uniformly distributed in the brain. Specific
neuroanatomical structures are targeted by the disease relative to others [85], and if AD pathology
represents a form of plasticity that was adaptive in our evolutionary past, we should expect to find
evolutionary significance in the distinctively patterned distribution. The findings reported here are
consistent with the idea that AD selectively diminishes the metabolic costs of neural areas that might
prove extraneous to a deprived forager while sparing the essential areas.
AD pathology is most damaging to areas responsible for higher-order learning and explicit memory but
spares areas essential for sensing and moving. For example, plaques and tangles are known to be
much more prevalent within the anterior (learning) but not the posterior (sensory) cortex [85]. Scanning
studies demonstrate reduced glucose utilization in the frontal, parietal and temporal association
cortices of patients with Alzheimer’s disease, when compared with age matched controls, whereas the
primary sensory, motor and visual regions are preserved [49]. In fact, the sensorimotor and visual
cortex have been shown to suffer the least damage of any cortical structures [86] and their selective
preservation has been described as “perplexing.” The sensorimotor and visual regions are areas that,
if damaged, would certainly disrupt the abilities to see and move. In other words, within the cortex,
Alzheimer’s follows a genetic program which selectively spares the areas associated with the senses
and muscle movement – regions that any foraging animal would find indispensable. It primarily affects
the higher areas that are associated with lofty but expendable, abstract thought.
Within each of the major cortical structures, the higher-order “association cortices” are the most
vulnerable to plaque and tangle formation whereas the adjacent primary sensory areas are the least
vulnerable [87]. The association areas are responsible for functions that one can imagine might be
superfluous to a weathered forager that has already “seen it all” such as mental comparison,
conceptual integration and analytical thought [58]. The primary sensory areas though, which are highly
conserved, allow basic perception and stereotyped response to stimuli [87]- functions indispensable
for any animal. Genes that promote damage to such vital areas would be so damaging to reproductive
fitness that we can see why they would not be maintained and why they would not exist today. Genes
that promote damage to less vital, “metacognitive” areas would be more likely to persist, especially
given that these areas are responsible for abstract learning [58]– which is, of course, very important to
a youngster, but less important to the experienced aged. It is also interesting to mention that,
phylogenetically, association cortices expand at a faster rate than primary sensory cortices [88],
indicating that, in cortical function, the AD brain operates like an evolved throwback rather than in some
arbitrary fashion.
The cortex, which is proportionally very large in humans and tiny in the simplest vertebrates, is the area
of the brain that is damaged the most by AD [86]. The noncortical areas responsible for respiratory
and cardiac function, waking state, emotionality, sensory perception and motor capacity are relatively
spared by AD; the damage that does occur only begins to compromise these functions in very late,
advanced AD. All noncortical areas display proportionately far less damage than the frontal cortex and
hippocampus [81]. These two, phylogenetically new, areas are thought to have allowed humans the
patience and analytical ability to learn cognitively demanding food extraction techniques and social
conventions [89,90]. Interestingly, the noncortical brain regions that do exhibit neuropathology are
primarily regions that have undergone very recent “integrated phylogeny” (i.e., extensive elaboration
with association neocortices since our divergence with the other apes) [49]. These areas include the
posterior hippocampus, the entorhinal cortex, the basocortical amygdaloid complex and the nucleus
basalis of Meynert. The fact that these findings reveal reverse-integrated phylogeny has influenced
authors to posit that Alzheimer’s is, neurologically, a phylogenetic disease [49]. This characterization
fits neatly with the present hypotheses and further frames AD not as arbitrarily localized pathology, but
as an atavistic cerebrotype that recapitulates a fundamental, ancestral configuration.
The relative deficiencies of the preclinical AD brain may be seen as mimicking the mental faculties of a
lower primate. The facets of cognition that are struck the hardest by AD include: sustained attention,
prolonged concentration, inhibition of impulse and higher order capacities for deliberation, prudence
and forethought. Apes are generally seen as being less deliberate, prudent and inhibited, and as
having less capable working memory [91]. These “relative deficits” do not detract from the fact that
apes are exquisitely crafted, able-bodied animals whose minds are maximized for their ecological
niche. The main way that the ape brain is different from our own is in size of the cortex, and thus their
limbic (emotional) system is proportionately large [88]. Consistent with all of this, it has been found that
the cortical areas that are the least vulnerable to AD are the ones with the smallest connectional
“distance” (measured in synapses removed) from the limbic areas [87]. Interestingly, this same pattern
is also true of normal brain aging, indicating that all of us are designed to make very precise, very
deliberate neurological changes with age. Relative selectivity in the AD brain should be taken as a
record of our behavioral past, where distribution of neuropathology evinces relative adaptive value of
individual processing modules and, thus, also of individual cognitive processes.
Several experiments using a variety of complex tasks have shown that the level of metabolic activity in
the cortex drops with practice, familiarity and increasing automaticity [58]. Many different tasks that
require concentration, executive activity and recruitment of various brain areas when the task is new
can be automated with experience to a point where the task can be completed efficiently with only a
minimal expenditure of metabolic activity [92]. Many tasks, even ones that seem relatively complex and
unpredictable, such as playing the videogame Tetris, show this effect [93]. One might imagine that the
tasks associated with foraging could easily become automated in the same way and that the drop in
cortical metabolic activity with age represents a cunning move on the part of natural selection. Because
the metabolic requirements for the completion of familiar tasks decreases with age and experience,
aging individuals can afford to lower overall cerebral metabolic output and still maintain their
fundamental behaviors and their ability to interact adeptly within their environment.
Comparative physiology and Alzheimer’s
Definitive proof of the adaptive value of a physiological phenomenon can be very hard to attain, but the
comparative approach is thought to provide strong evidence [94]. This approach commonly compares
the same physiological process in (1) closely related species that live in strikingly different
environments, or (2) distantly related species that live in identical environments. If the process is
associated with adaptation to environment, it is expected that group 1 will show large variation in the
trait and group 2 will show little. First we will employ approach number one and examine the large
natural variation in susceptibility to AD within our species.
Two human groups that differ tremendously in way of life and natural environment from other humans
are the Khoi San hunter-gatherers and the pygmy foragers of sub-Saharan Africa. The APOE 4 allele
(currently the gene most strongly associated with AD susceptibility) has been shown to be more than
twice as common in Khoi San (.37) [95] and pygmies (.40) [96] as it is in other populations, where it is
usually less than 20 percent and often less than 12 percent. Why might this be so? These individuals
represent human groups that, because they never utilized agriculture, have been exposed to nutritional
shortages up until recent times [97]. They live and forage in much the same way that their ancestors
must have for tens of thousands of years. The fact that the APOE 4 gene is so frequent among these
groups suggests that there is a strong relationship between way of life, nutritional scarcity and selective
pressure on AD.
Modern molecular studies have shown that the Khoi San are more distantly related to the rest of
humanity than any other group, meaning that they may have changed the least since we all diverged
from a single, ancestral “stem” group around 180,000 years ago [98]. This suggests that the ancient
foraging ancestors to us all may have manifested preclinical AD to a larger degree than the modern
human population at large. Today, the incidence of APOE 4 remains highest among all populations
where an economy of foraging still exists (e.g., Pygmies, Khoi San, aborigines of Malaysia and
Australia, Papuans, European Lapps, and some Native Americans), or the food supply is known to
have been frequently scarce, or nutritionally poor [97]. This data shows that AD does in fact show high
variation between closely related but ecologically disparate groups. In the next section of this article
we will consider the observation that known “thrifty” genes, associated with adaptive energy
conservation, show this same geographical pattern, strengthening the association between AD and
metabolic thriftiness.
Now we will consider the second comparative approach; a commonality in two distantly related
species, rats and humans. Animals that have been subjected to starvation or nutritional deprivation
exhibit multiple physiological methods of energy conservation [99]. These changes involve lowering the
energy demands of the most expensive organs. Interestingly, some of these changes are known to
occur in the brain, and some appear highly analogous to the changes seen in AD. One of the
neurohistological hallmarks of AD, hyperphosphorylated tau proteins, form reversibly in the brains of
starving or stressed rats. After a number of hours of food restriction, hyperphosphorylated tau begins to
accumulate within neurons of the rat hippocampus [100]. This abnormal tau, a transport protein, is one
of the primary pathohistologicall hallmarks of AD, and is the cause of neurofibrillary tangles in humans
[101]. Kurt Heininger [102] has concluded that these molecular similarities shared between AD and
starving mammals suggests that AD constitutes a “rescue program” that “actively adapts to
progressive fuel deprivation.”
In rats, the accumulation of these proteins causes pronounced decrements in brain metabolism
allowing the rat more time to seek out food before it collapses from starvation. This phenomenon,
which has been replicated reliably, has been explicitly attributed adaptive significance [103].
Interestingly, after the rat is fed again, the tau dephosphorylates and the hippocampus regains normal
function [100]. This reversible, phenotypic change in rats is analogous to the permanent change seen
in AD because both may protect against starvation. Before these changes were observed in starving
lab animals, they were thought to be specific and exclusive to AD. Ironically, the literature today views
these changes as adaptive alterations to starvation in lab rodents [103,104] but as pathological in AD.
Mammals are well-known to demonstrate a suite of plastic responses to extreme hunger and starvation
that help to minimize energy expenditure. The most dramatic of these includes suppression of
metabolic rate, reduction of thyroid hormone levels and growth hormone levels, suppression of gonadal
function, and an increased activation of the hypothalamic-pituitary-adrenal axis [99,105]. As discussed
in the next section, aging individuals and especially individuals with AD exhibit increased incidence of
each of these physiological alterations [68,106]. Ostensibly aging humans in the environment of
evolutionary adaptedness would have benefited from the same physiological alterations as starving
animals. This further characterizes AD neuropathology as a response to nutritional scarcity.
Many species of mammals, most notably primates, have been shown to develop the neuropathological
hallmarks (both senile plaques and neurofibrillary tangles) of AD [7,107] It has been hard for
specialists to link changes in these animals’ behavior to the underlying pathology because they are
wild. However, our domesticated cats and dogs, whose lives have also been artificially extended in the
last few centuries, also exhibit Alzheimer’s like neuropathology. Cats, dogs and some other
mammalian pets diagnosed with cognitive dysfunction syndrome (CDS), which is thought to be highly
comparable behaviorally to Alzheimer’s in humans [108], have been shown to have specific molecular
and cellular changes that mimic the pathophysiology of AD [109]. In fact, the behavioral abnormalities
of CDS in cats and dogs are accompanied by the same hallmarks of AD in humans: high levels of
diffuse amyloid beta accumulation as well as hyperphosphorylated tau, predominantly in cortical and
hippocampal areas [110]. These findings suggest that AD-like neuropathology may have played an
adaptive role in the life histories of our pet’s non-domesticated ancestors. Perhaps most long-lived
mammals can benefit from diminishing the expenditure of resources on cognition during the aging
process. Very short-lived mammals, reptiles, amphibians and fish generally do not show clear signs of
AD pathology [7], and this may be because their brains are far smaller, their learning arcs are much
abbreviated, their life spans are typically much shorter and their life histories are much different.
Perhaps the degree of expression and the severity of the symptoms increased in humans (relative to
other mammals) because of life history parameters characteristic of our species, such as large
intergenerational resource flows, large brains and long lifespan.
Alzheimer’s and the thrifty genotype
The human brain is an extravagantly expensive organ [70]. In fact, the mass specific metabolic rate of
brain tissue is over 22 times that of skeletal muscle [111]. Such a ravenous organ can be viewed as an
unfavorable encumbrance because it is known that the Plio-Pleistocene was marked by recurring ice
ages where large, polar glaciations would frequently cause the African savannah to become arid, hot
and nutritionally scarce [64, 112]. Such unpredictability of food resources in our environment of
evolutionary adaptedness is widely thought to have seriously impacted the human genome [113]. In
fact, the blossoming science of “evolutionary medicine” has pointed out that many human diseases are
simply vestiges of our hunter-gatherer existence and some of the best characterized among these are
thought to have provided protection against starvation [8]. Genetic tendencies toward adiposity and
insulin resistance would have helped our metabolisms deal with low food availability in the ancestral
environment, but in modern times are responsible for the high incidence of obesity and diabetes [114].
Many metabolic disorders have been shown to represent mismatches – in the way our bodies are
regulated – between the austerity of the ancestral environment the cheap abundance of calories and
fats seen in modern times.
The thrifty genotype hypothesis [115] posits that certain human genes that are associated with
increased risk for metabolic disease today were naturally selected in the past because they helped
their bearers to be more ‘thrifty’ with energy stores. According to this hypothesis, phenotypes that
express low metabolic rates enjoy a survival advantage under deprived circumstances. However, they
face increased risk of negative health consequences when sugars and fats are artificially abundant
[116]. Thrifty genes, associated with a number of different metabolic traits, are thought to allow an
organism to conserve calories, increase fat deposition and adopt a sedentary nature [117].
Interestingly, AD patients are well-known to exhibit many of these well-established thrifty traits,
suggesting that genes that predispose to AD neuropathology may be part of a complement of thrifty
genes that have a tendency to present comorbidly.
The metabolic syndrome, a suite of thrifty traits that tend to present together including obesity, diabetes
and heart disease, has been strongly tied to AD [106]. Studies even show that AD patients have lower
levels of physical activity and energy expenditure than age-matched elderly individuals that do not have
AD [118,119]. Both old age and, especially, AD have been associated with lower testosterone levels,
lower muscle mass, lower thyroid levels, lower growth hormone levels, insulin resistance, increased
adiposity, upregulated hypothalamic pituitary adrenal axis and a significantly lower resting metabolic
rate [67,68]. Because modern day individuals that develop AD show a highly increased propensity for
developing these traits, we can expect that this was also the case in ancestral times, and that such
individuals would have enjoyed increased survival under deprived circumstances.
The thrifty phenotype hypothesis has been used widely to interpret distinct patterns of the geographic
distribution of genes that are associated with the metabolic syndrome. This widely accepted
interpretation emphasizes that, populations of foraging individuals that live in areas where, until
recently, food has been relatively unpredictable or frequently scarce have much higher prevalence of
thrifty genes [120]. The traits that these genes code for helped these individuals survive during
prolonged periods of scarcity by allowing them to maintain low basal metabolic rates [115]. Recent
research has shown that the geographic distribution of the gene most strongly related to AD
susceptibility – the APOE 4 allele – matches the geographic distribution for all the other major thrifty
genes, such as insulin resistance and increased adiposity. This has influenced some researchers to
hypothesize that APOE 4 may also be a thrifty gene [97]. These findings further indicate that AD may
represent a type of thrifty phenotype that would have been well-suited for nutritional scarcity.
Type 2 diabetes mellitus (T2DM) has been characterized as a disorder that is a prototypic example of
Darwinian medicine and the thrifty phenotype. It has been thought for decades now that, on a cellular
and organ system level, the disorder represents a thrifty condition – insulin resistance – that would
have only rarely manifested as disease in the ancestral environment because at that time individuals
had no access to pure sugar or processed foods [115]. The current literature holds that insulin
resistance, brought about by genes for T2DM, represents a finely tuned physiological state and that its
cellular and molecular pathways have been refined by natural selection over millions of years to help
the organism conserve blood sugar [117]. Not surprisingly, AD and T2DM share many common
biochemical and physiological pathways. The uncanny resemblances include widespread changes in
insulin signaling, changes in transforming growth factor systems, similarly high counts of
hyperphosphorylated proteins, and dense aggregations of amyloid peptides [121]. Fascinatingly, there
is a 90 percent structural similarity between the amyloid peptide in AD and the islet amyloid
polypeptide in T2DM; this close correspondence in amino acid sequence is thought to suggest similar
physiological roles [122]. In addition, other peptides such as APOE, heparin sulfate proteoglycans and
serum amyloid P play “crucial” roles in both T2DM and AD [123,124]. The literature on the adaptive
functions of diabetes has never been reconciled with the literature on the similarities between AD and
T2DM. A comprehensive comparison of pathology between the two should be informed by the
hypothesis that AD is also a thrifty condition.
Each of the thrifty conditions discussed here have inherent drawbacks. The organism is calibrated to
expend less energy, supply lines are narrowed and functionality is constrained. Every adaptation in
evolutionary science represents a tradeoff where an interest in one concern is emphasized at the
expense of another. The foremost, omnipresent concern for all life forms is “the struggle for
sustenance” and the brain changes that accompany aging are clearly a response to this concern.
Paradoxically though, these changes are being made at the expense of the human mind. The present
hypothesis, though simple and straightforward, may remain hostile or inconceivable to many because
of humanity’s anthropocentric perspective on the mind. That the metabolic output of internal organs can
decrease in response to starvation does not constitute a hostile idea. That the metabolic output in the
“organ of thought” might also decrease, however, is unfriendly for many philosophical (Cartesian
dualist) reasons. Perhaps the main reason that this is unfriendly is that many people have never
considered that their thoughts arise from, and in every sense are made possible by, ingested food. We
have come to revere our own minds so much that it is hard for us to see that advanced intelligence, like
every biological adaptation, can be maladaptive under the right circumstances.
“Paradoxically, sometimes losing neurons can improve brain function, as happens when synaptic
connections and neurons that have not been extensively used die off, in perhaps the most dramatic
case of use it or lose it. Keeping unused neurons supplied with blood, oxygen and energy is wasteful,
and getting rid of them keeps the brain more focused and efficient.”
-Norman Doidge, 2007 [47]
It has been concluded that millions of years of drastic fluctuations in food availability would have
exposed our close ancestors to periods of unavoidable hunger and starvation [64,112]. After sufficient
time passes without a meal, the body’s cells exhaust the supply of blood sugars and hypoglycemia
ensues. When the nutrients from the last meal are depleted, the body begins to catabolize stored
energy. Glycogen from the liver and muscles can sustain a moderately active human for several hours,
but when these fuel stores run out the body must begin to burn supplies of fat. The process of ketosis
liberates body fat for use in the muscles and brain; however, the brain has priority over these fuel
reserves. Many tissues in the body can down-regulate their metabolic rate in the absence of energy,
but the brain cannot [125]. Once the brain can no longer extract the fuel it requires from the blood,
unconsciousness, coma and death ensue [126]. This shows that the disproportionately gluttonous
human brain can potentially be an extravagant liability. Inability to temper this organ with age would
have created a burden that our aging forebears could often not afford.
Implications of the hypothesis
As longevity around the world continues to climb, the incidence of AD has been increasing steadily.
Today, it is estimated that about four million people in the United States suffer from AD and that
number is expected to triple before 2050 [127]. It has even been called the pandemic of the 21st
century [128]. In the environment of evolutionary adaptedness, our lives were far shorter and true cases
of AD would have been extremely rare throughout prehistory. It is commonly pointed out that evolution
has no foresight or predictive capacity and, unfortunately, it seems that an adaptation that was meant
to help our ancestors now makes us susceptible today to senile dementia since we have artificially
extended our lifespan.
It may not possible at this point to determine irrefutably if AD is a protective adaptation that served a
purpose during human evolution. Further analysis and experimentation should allow more definite
conclusions. For instance: (1) the formation of plaques and tangles should prove to be the most energy
and resource efficient way for neurons to lose connections; (2) the neuropathology should present only
in brain areas that are consistent with neuroecological rationale; (3) patterns of heritability and etiology
should reflect evolutionary trends; (4) onset of pathology should coincide with ages of reduced foraging
ability; and (5) molecular patterns and evolutionary signatures (such as relative intron-exon selection)
should reveal that the genes involved in AD were positively selected relative to alternative alleles. The
present hypotheses may also have major implications for medical research and ultimately for
treatment: (1) the pathways that allow the reversal of hyperphosphorylated tau in starving rats will hold
key clues to reversing or preventing AD (2) the close evolutionary association with diabetes mellitus
and other thrifty disorders should influence researchers to focus on pathocomparative analyses (3)
theorists now have a reason to reconcile the vast literature on thrifty genes, deprivation syndromes and
phenotypic plasticity with the literature on the pathophysiology of AD.
The present evolutionary analysis of AD can be extended in far reaching ways. It is known that the most
conspicuous of the various changes that accompanied human evolution was the significant increase in
relative brain size, especially the size of the cerebral cortex. How human encephalization was
accomplished and why hominids were selected for bigger brains despite the energetic costs has been
a subject of much debate [129]. Many of the concepts invoked in this debate might help to elucidate the
evolutionary basis for AD. Weighing the ecological costs of AD with the benefits might prove to be very
informative, and future work should include discussion of concepts like carnivory vs. herbivory, dietary
quality, encephalization index, the expensive tissue hypothesis, food extraction techniques, life history
theory, optimal foraging theory, resting vs. maximal metabolic rate, total energy budget, the Kleiber
relationship, evolutionary theories of aging and others.
Many species of marine animals called urochordates, also known as tunicates and sea squirts, do an
amazing job of illustrating how intelligence and excessive nervous tissue can be an encumbrance that
natural selection will go to lengths to protect against. This small invertebrate is born with sense organs
and a brain (a cerebral ganglion) and resembles a tadpole. It needs its simple nervous system in order
to coordinate its behavior while searching for a home early in life. The sea squirt has a second stage
where it locates a suitable surface on a rock and attaches itself permanently. From then on, its life
history changes from a mobile to a sedentary strategy, and it actually spends the rest of its life
passively filter feeding – grabbing nutrients as they float by [130]. After attaching to the rock, the little
creatures digest their brains and large regions of their own nervous systems specifically in order to
minimize energy use [131]. They never grow them back.
Similarly, individuals in ancestral hunter-gatherer tribes in their 30s, 40s and 50s must have
experienced diminishing usefulness for higher order intelligence. They were no longer fastidiously busy
creating novel motor praxes, learning foreign spatial maps or developing original motor activities. The
pressure on them to internalize unfamiliar information and to see unprecedented relationships between
important variables dwindled every year of their life as their experience increased. They were not
actively developing untried social skills to create new alliances or find new partners for copulation, and
also were not directly responsible for caring for infants or inculcating children. Further, they could no
longer employ many of the high yield foraging techniques that they utilized in their youth, so they must
have been quite receptive to the benefits of adopting an alternate “thrifty cognitive” strategy.
A thoughtful Talmudic commentary states that “scholars become wiser as they age, but the
noneducated become foolish.” Specialists in technical fields acquire large amounts of knowledge over
the years that, during old age, allow them to converse fluently and intelligently on their subject despite
the cognitive morbidity that they may suffer. In a general sense, these sentiments reflect the idea that
whatever you spend much time doing, you become an expert at. Much unlike today, every individual in
prehistoric groups foraged throughout the day from a very early age until death. The expertise gained
from this would have made them very “wise” at hunting and gathering. It is the view espoused here that
this physical and mental wisdom, earned from decades of foraging, allowed foragers to reduce their
investment in the ‘fixed cost’ of energy consumption within the brain. Moreover, it is clear that
preclinical AD would not have been disruptive to the implicit techniques, procedural knowledge or
motor praxes necessary for foraging, even though clinical AD is disruptive to the explicit techniques,
semantic knowledge and declarative cognition necessary for both occupational employment and
autonomous living today.
Competing interests
The author declares that he has no competing interests.
References
1. Selkoe DJ. Alzheimer’s disease: From genes to pathogenesis. Am J Psychiatr 1997, 154:1198.
2. Erickson CA, Barnes CA. The neurobiology of memory changes in normal aging. Exp Gerontol
2003, 38: 61-69.
3. Troller JN, Sachdev PS, Haindl W, Brodaty H, Wen W, Walker BM. Regional cerebral blood flow
deficits in mild Alzheimer’s disease using high resolution single photon emission computerized
tomography. Psychiatr Clin Neurosci 2005, 59:280-290.
4. Niedermeyer E. Alzheimer disease: Caused by primary deficiency of the cerebral blood flow. Clin
EEG Neurosci 2006. 37(3):175-177.
5. Hedden T & Gabrieli JDE. Insights into the aging mind: A view from cognitive neuroscience. Nat
Neurosci 2004, 5:87-96.
6. Rosen RF, Farberg AS, Gearing M, Dooyema J, Long PM, Anderson DC, Davis-Turak J, Coppola
G, Geschwind DH, Pare J, Duong TQ, Hopkins WD, Preuss T & Walker LC. Taupathy with paired
helical filaments in an aged chimpanzee. Journal Compar Neurol 2008, 509:259-270.
7. Finch CE, Sapolsky RM. The evolution of Alzheimer disease, the reproductive schedule, and apoE
isoforms. Neurobiol Aging 1999, 20: 407-428.
8. Nesse R, Williams G. Evolution and the origins of disease. Sci Am 1998, 279: 58-65.
9. Blurton Jones N, Smith L, O’Connell J, Hawkes K, Samuzora CL. Demography of the Hadza, an
increasing and high density population of savanna foragers. Am J Phys Anthropol 1992, 89:159-181.
10. Hill K, Hurtado AM. Ache life history: the ecology and demography of a foraging people.
Hawthorne, NY: Aldine de Gruyter 1996.
11. Morris JC, Storandt M, McKeel DW Jr, Rubin EH, Price JL, Grant EA, Berg L. Cerebral amyloid
deposition and diffuse plaques in “normal” aging: evidence for presymptomatic and very mild
Alzheimer’s disease. Neurology 1996, 46:707-719.
12. Bufill E, Blesa R. Alzheimer’s disease and brain evolution: Is Alzheimer’s disease an example of
antagonistic pleiotropy? Rev Neurol 2006, 42(1): 25-33.
13. Wick G, Berger P, Jansen-Durr P, Grueck-Loebenstein B. A Darwinian-evolutionary concept of
age-related diseases. Exp Gerontol 2003, 38:13-25.
14. Twamley EW, Ropacki SA, Bondi MW. Neuropsychological and neuroimaging changes in
preclinical Alzheimer's disease. J Int Neuropsychol Soc 2006, 12 (5):707–735.
15. Hubacek JA, Pitha J, Skodova Z, Adamkova V, Lanska V & Poledne R. A possible role of
apolipoprotien E polymorphism in predisposition to higher education. Neuropsychobiology 2001, 43:
200-203.
16. Adams RD: Morphological aspects of aging in the human nervous system. In Handbook of Mental
Health and Aging. Edited by Birren JE, Sloane RB. Englewood Cliffs, NJ: Prentice Hall; 1980:149-160.
17. Finch CE, Stanford CB. Meat adaptive genes and the evolution of slower aging in humans. Q Rev
Biol 2004, 79:3-50.
18. Cavanaugh JC & Blanchard-Fields. Adult Development and Aging (Fifth Edition). Belmont CA:
Wadsworth/Thompson Learning; 2002
19. Scheibel AB. 1996. Structural and functional changes in the aging brain. In Handbook of the
psychology of aging. Edited by Birren JE, Schaie KW. San Diego: Academic Press; 1996:105-128..
20. Whitbourne SK. The aging individual: Physical and psychological perspectives. New York:
Springer; 1996.
21. Terry RD & Katzman R. Life span and synapses: Will there be a primary senile dementia?
Neurobiol Aging 2001, 22: 347-348.
22. Plomin R, McGuffin P. Psychopathology in the postgenomic era. Ann Rev Psychol 2003, 54:205-
228.
23. Ernst RL, Hay JW. The US economic and social costs of Alzheimer’s disease revisited. Am J
Public Health 1994, 84(8): 1261-1264.
24. Hood E & Jenkins KP. Evolutionary Medicine: A Powerful Tool for Improving Human Health.
Evolution: Education and Outreach 2008, 1:114-120.
25. Nesse R, Williams G. Why We Get Sick. New York: Random House; 1995.
26. Williams G. Nesse R. The Dawn of Darwinian Medicine. Q Rev Biol 1991, 66:1-22.
27. Nesse, R. What Darwinian medicine offers psychiatry. In Evolutionary Medicine. Edited by
Trevathan W, Smith E, McKenna J. Oxford: Oxford University Press; 1999.
28. Baron-Cohen S. The maladapted mind: Classic readings in evolutionary psychopathology. Hove,
East Sussex: Psychology Press. 1997.
29. Reser JE. Evolutionary neuropathology and congenital mental retardation: Environmental cues
predictive of maternal deprivation influence the fetus to minimize cerebral metabolism in order to
express bioenergetic thrift. Med Hypotheses 2006; 67(3): 529-544.
30. Reser JE. Evolutionary neuropathology and Down syndrome: an analysis of the etiological and
phenotypical characteristics of Down syndrome suggests that it may represent an adaptive response
to severe maternal deprivation. Med Hypotheses 2006; 67(3):474-481.
31. Reser JE. Schizophrenia and phenotypic plasticity: Schizophrenia may represent a predictive,
adaptive response to severe environmental adversity that allows both bioenergetic thrift and a
defensive behavioral strategy. Med Hypotheses 2007, 69:383-394.
32. McLoon SC. Alteration in precision of the crossed retinotectal projection during chick development.
Science 1982, 215:1418-1420.
33. Perry G, Nunomura A, Lucassen P, Lassmann H, Smith MA. Apoptosis and Alzheimer’s Disease.
Science 1998, 282(5392):1265.
34. Jacobs DS, Perry VH, Hawken MJ. The postnatal reduction of the uncrossed projection from the
nasal retina in the cat. J Neurosci 1984, 4:2425-2433.
35. Lanser ME, Fallon JF. Development of the lateral motor column in the limbless mutant chick
embryo. J Neurosci 1984, 4:2043-2050.
36. Lanser ME, Fallon JF. Development of the branchial lateral motor column in the wingless mutant
chick embryo: Motorneuron survival under varying degrees of peripheral load. J Comp Neuro 1987,
261: 423-434.
37. O’Leary DDM, Fawcett JW, Cowan WM. 1986. Topographic targeting errors in the retinocollicular
projection and their elimination by selective ganglion cell death. J Neurosci 1986, 6:3692-3705.
38. Hamburger, V. Cell death in the development of the lateral motor column of the chick embryo. J
Comp Neurol 1975, 160:535–546
39. Finlay BL, Wikler KC, Sengelaub DR. Regressive events in brain development and scenarios for
vertebrate brain evolution. Brain Behav Evol 1987, 30:102-17.
40. Katz MJ, Lasek RJ. Evolution of the nervous system: Role of ontogenetic mechanisms in the
evolution of matching populations. Proc Natl Acad Sci 1978, 75:1349-1352.
41. Finlay BL, Slattery M. Local differences in the amount of early cell death in neocortex predict adult
local specializations. Science 1983, 219:1349-51.
42. Williams RW, Herrup K. The control of neuron number. Ann Rev Neurosci 1988, 11:423-453.
43. Baltes PB, Staudinger UM, Lindenberger U. Lifespan psychology: Theory and application to
intellectual functioning. Annu Rev Psychol 1999, 50: 471-507.
44. Sylvester, R. What the biology of the brain tells us about learning. Educational leadership 1993, 51
(4):46-51.
45. Epstein, H. Stages of increased cerebral blood flow accompany stages of rapid brain growth.
Brain Dev 1999, 21(8): 535-539.
46. Greenough WT, Black JE & Wallace CS. Experience and Brain Development. Child Dev 1987, 58
(3):539-559.
47. Doidge N. The brain that changes itself. New York: Penguin Books; 2007, p.82.
48. Smith H, Tompkins RL. Towards a life history of the hominidae. Annu Rev Anthropol 1995, 24:257-
279.
49. Rapoport SI. Brain evolution and Alzheimer’s disease. Rev Neurol (Paris) 1988, 144(2): 79-90.
50. Brainerd CJ, Reyna VF. Memory independence and memory interference in cognitive
development. Psychol Rev 1993, 100(1): 42-67.
51. Moscovitch M, Chein JM, Talmi D & Cohn M. Learning and memory. In Cognition, brain, and
consciousness: Introduction to cognitive neuroscience. Edited by BJ Baars & NM Gage. London, UK:
Academic Press; 2007, p.234.
52. Budson AE, Price BH. Memory Dysfunction. N Engl J Med 2005, 352: 692-699.
53. Herholz K, Carter SF, Jones M. Positron emission tomography imaging in dementia. Br J Radiol
2007. 80(2):S160-167.
54. Burton, RA. On Being Certain. New York, NY: St. Martin’s Press; 2008.
55. Carlesimo GA, Oscar Berman M. 1992. Memory deficits in Alzheimer’s patients: a comprehensive
review. Neuropsychologia 1992, 3:119-169.
56. Horn JL. Organization of data on life-span development of human abilities. In Life-span
developmental psychology: Research and theory. Edited by Goulet LR, Baltes BP. London: Academic
Press; 1970: 146-159.
57. Ridley M. Genome: The autobiography of a species in 23 chapters. New York: HarperCollins
Publishers; 2000, p.96.
58. Baars BJ & Gage NM: Cognition, brain, and consciousness: Introduction to cognitive
neuroscience. London, UK: Academic Press. 2007
59. Thompson RF. In search of memory traces. Annu Rev Psychol 2005. 56:1-23.
60. Edelman, G. Neural Darwinism. The Theory of Neuronal Group Selection. New York: Basic Books;
1987.
61. Kaplan H, Hill K, Lancaster J, Hurtado AM. The evolution of intelligence and the human life history.
Evol Anthrop 2000, 9:156-184.
62. Kaplan H, Lancaster J, Robson A. Embodied Capital and the Evolutionary Economics of the
Human Life Span. Population Dev Rev 2003, 29:113-129.
63. Ravelo AC, Andreasen DH, Lyle M, Lyle AO and Wara MW. Regional climate shifts caused by
gradual global cooling in the Pliocene epoch. Nature 2004, 429:263–267.
64. Reed KE. Early hominid evolution and ecological change through the African Plio–Pleistocene. J
Hum Evol 1997, 32:289–322.
65.Gurven M, Kaplan H, Gutierrez M. How long does it take to become a proficient hunter?
Implications for the evolution of extended development and long life span. J Hum Evol 2006, 51(5):454-
470.
66. Kaplan H & Robson AJ. The emergence of humans: The coevolution of intelligence and longevity
with intergenerational transfers. PNAS 2002, 99:10221-10226.
67. Heininger Kurt. Aging is a deprivation syndrome driven by a germ-soma conflict. Ageing Res Rev
2002, 1:481-536.
68. Ricklefs RE & Finch CE. Aging: A natural history. New York: W.H. Freeman and Company; 1995.
69. Zafon, C. Spend less, live longer. The “thrifty aged” hypothesis. Med Hypotheses 2006, 67:15-20.
70. Aiello LC, Wheeler P: The Expensive Tissue Hypothesis: the brain and the digestive system in
human and primate evolution. Curr Anthropol 1995, 36:199-221.
71. Mink JW, Blumenschine RJ, Adams DB. Ratio of central nervous system to body metabolism in
vertebrates: Its constancy and functional basis. Am J Physiol 1981, 241: R203-212.
72. Fish JL & Lockwood CA. Dietary constraints on encephalization in primates. Am J Phys Anthropol
2003, 120(2):171-181.
73. Kempermann G. Why New Neurons? Possible Functions for Adult Hippocampal Neurogenesis. J.
Neurosci 2002, 22(3) 635-638.
74. Garamszegi LZ, Eens M. The evolution of hippocampus volume and brain size in relation to food
hoarding in birds. Ecology Letters 2004, 7:1216.
75. Gould E, Beylin A, Tanapat P, Reeves A, Shors TJ. Learning enhances adult neurogenesis in the
hippocampal formation. Nat Neurosci 1999, 2:260-265.
76. Dukas R. Evolutionary Biology of Animal Cognition. Annu Rev Ecol Evol Syst 2004, 35:347-374.
77. Nilsson M, Perflilieva E, Johansson U, Orwar O, Eriksson P. Enriched environment increases
neurogenesis in the adult rat dentate gyrus and improves spatial memory. J Neurobiol 1999, 39:569-
578.
78. Brun A, Englund E. Regional pattern of degeneration in Alzheimer’s disease: neuronal loss and
histopathological grading. Histopathology 1981, 5:549-564.
79. Meguro K, LeMestric C, Landeau B, Desgranges B, Eustache F, Baron JC. Relations between
hypometabolism in the posterior association neocortex and hippocampal atrophy in Alzheimer’s
disease: a PET/MRI correlative study. J Neurol Neurosurg Psychiatry 2001, 71:315-321.
80. Pearson RC, Esiri MM, Hiorns RW, Wilcock GK, Powell TP. Anatomical correlates of the
distribution of the pathological changes in the neocortex in Alzheimer disease. Proc Natl Acad Sci
USA 1985, 82:4531-4534.
81. De Leon MJ, Convit A, de Santi S. Structural neuroimaging: early diagnosis and staging of
Alzheimer’s disease. In Alzheimer’s disease and related disorders. Edited by Iqbal K, Winblad B,
Nishimura T. New York: John Wiley; 1999: 105-126.
82. Minoshima S, Giordani B, Berent S. Metabolic reduction in the posterior cingulate cortex in very
early Alzheimer’s disease. Ann Neurol 1997, 42:85-94.
83. Killiany RJ, Gomez-Isla T, Moss M. Use of structural magnetic resonance imaging to predict who
will get Alzheimer’s disease. Ann Neurol 2000, 47:430-439.
84. Friedland RP, Budinger TF, Ganz E. Regional cerebral metabolic alterations in dementia of the
Alzheimer’s type: positron emission tomography with 18F-fluorodeoxyglucose. J Comput Assist
Tomog 1983, 7:590-598.
85. Ohnishi T, Matsuda H, Tabira T, Asada T, Uno M. Changes in Brain Morphology in Alzheimer
Disease and Normal Aging: Is Alzheimer Disease an Exaggerated Aging Process? Am J Neuroradiol
2001, 22: 1680-1685.
86. Thompson PM, Hayashi KM, de Zubicaray G, Janke AL, Rose SE, Semple H, Herman D, Hong
MS, Dittmer SS, Doddrell DM & Toga AW. Dynamics of Gray Matter Loss in Alzheimer’s Disease.
Journal Neurosci 2003, 23(3): 994.
87. Arnold SE, Hyman BT, Flory J, Damasio AR and Van Hoesen GW: The Topographical and
Neuroanatomical Distribution of Neurofibrillary Tangles and Neuritic Plaques in the Cerebral Cortex of
Patients with Alzheimer's Disease. Cereb Cortex 1991, 1:103-116.
88. Lynch G, Granger R. Big Brains: The origins and future of human intelligence. New York: Palgrave
MacMillan; 2008.
89. Clutton-Brock TH, Harvey PH. Primates, brains and ecology. Journal of the Zoological Society of
London 1980, 190:309-23.
90. Gibson KR. Cognition, brain size and the extraction of embedded food resources. In Primate
Ontogeny, Cognition and Social Behavior. Edited by Else JGF, Lee PC. Cambridge: Cambridge
University Press; 1987.
91. Pennisi E. Psychology. Nonhuman primates demonstrate humanlike reasoning. Science 2007, 317
(5843):1308.
92. Schneider W & Chein JM. Controlled and automatic processing: behavior, theory and biological
mechanisms. Cogn Sci 2003, 27:525-559.
93. Haier RJ, Siegel BV Jr, MacLachlan A, Soderling E, Lottenberg S, Buchsbaum MS. Regional
glucose metabolic changes after learning a complex visuospatial/motor task: a positron emission
tomographic study. Brain Res 1992, 570:134-143.
94. Miles DB, Dunham AE. Historical perspectives in ecology and evolutionary biology: The use of
phylogenetic comparative analyses. Ann Rev Ecology Systematics 24: 587-619.
95. Sandholzer C, Delport R, Vermaak H, Utermann G. High frequency of the apo epsilon 4 allele in
Khoi San from South Africa. Hum Genet 1995. 95(1):46-48.
96. Zekraoui L, Lagarde JP, Raisonnier A, Gerard N, Aouizerate A, Lucotte G. High frequency of the
apolipoprotein E 4 allele in African pygmies and most of the African populations in sub-Saharan Africa.
Hum Biol 1997. 69(4):575-581
97. Corbo RM & Scacchi R. Apolipoprotein E (APOE) allele distribution in the world. Is APOE 4 a
‘thrifty’ allele? Ann Hum Genet 1999, 63:301-310.
98. Semino O, Santachiara-Benerecetti AS, Francesco F, Cavalli-Sforza LL, Underhill PA. Ethiopians
and Khiosan share the deepest clades of the human y-chromosome phylogeny. Am J Hum Genetics
2002, 70:265-268.
99. Flier JS. Clinical review 94: What’s in a name? In search of leptin’s physiologic role. J Clin
Endocrinol Metab 1998, 83:1407–1413.
100. Planel E, Yasutake K, Fujita SC, Ishiguro K. Inhibition of protein phosphatase 2a overrides tau
protein kinase i/glycogen synthase kinase 3β and cyclin-dependent kinase 5 inhibition and results in
tau hyperphosphorylation in the hippocampus of starved mouse. J Biol Chem 2001, 276(36), 34298-
34306.
101. Yanagisawa M, Planel E, Ishiguro K, Fujita SC. Starvation induces tau hyperphosphorylation in
mouse brain: implications for Alzheimer’s disease. FEBS Letters 1999, 461(3):329-333.
102. Heininger K. A unifying hypothesis of Alzheimer's disease. IV. Causation and sequence of events.
Rev Neurosci 2000, 11:213-328.
103. Ikeda Y, Ishiguro K, Fujita SC. Ether stress-induced Alzheimer-like tau phosphorylation in the
normal mouse brain. FEBS Letters 2007, 581:891-897.
104. Arendt T, Stieler J, Strijkstra AM, Hut RA, Rudiger J, Van der Zee EA, Harkany T, Holzer M and
Hartig W: Reversible paired helical filament-like phosphorylation of tau is an adaptive process
associated with neuronal plasticity in hibernating animals. J Neurosci 2003, 23: 6972-6981.
105. Schwartz MW, Dallman MF, Woods SC. Hypothalamic response to starvation: implications for the
study of wasting disorders. Am J Physiol 1995, 269:R949–957.
106. Razay G, Vreugdenhil A, & Wilcock G. The metabolic syndrome and Alzheimer disease. Arch
Neurol 2007, 64:93-96.
107. Pawlik M, Fuchs E, Walker LC, Levy E. Primate-like amyloid-beta sequence but no cerebral
amyloidosis in aged shrews. Neurobiol Aging 1999.20:47-51.
108. Landsberg G & Araujo JA. Behavior problems in geriatric pets. Vetrinary clinics of North America:
Small animal practice. 2005, 35(3):675-698.
109. Head E, Moffat K, Das P, Sarsoza F, Poon WW, Landsberg G, Cotman CW, Murphy MP. Beta-
amyloid deposition and tau phosphorylation in clinically characterized aged cats. Neurobiol Aging
2005, 26(5):749-763.
110. Gunn-Moore D, McVee J, Bradshaw JM, Pearson G, Head E, Gunn-Moore F. Ageing changes in
cat brains demonstrated by β-amyloid and AT8-immunoreactive phosphorylated tau deposits. Journal
of Feline Medicine & Surgery 2006, 8(4):234-242.
111. Aschoff J, Günther B, Kramer K: Energiehaushalt und Temperaturregulation. Munich: Urban and
Schwarzenberg. 1971.
112. Bobe R, Behrensmeyer AK & Chapman RE. Faunal change, environmental variability and late
Pliocene hominin evolution. J Hum Evol 2002, 42:475–497.
113. Wynn JG. Influence of Plio–Pleistocene aridification on human evolution: evidence from paleosols
of the Turkana Basin, Kenya. Am J Phys Anthropol 2004, 123:106–118.
114.Hales CN, Barker DJ. Type 2 (non-insulin-dependent) diabetes mellitus: the thrifty phenotype
hypothesis. Diabetologia 1992, 35:595-601.
115. Neel JV. Diabetes mellitus: a ‘thrifty’ genotype rendered detrimental by ‘progress’? Am J Hum
Genet 1962, 14: 353-362.
116. Bateson P, Barker D, Clutton-Brock T, Deb D, D’Udine B, Foley R, Gluckman P, Godfrey K,
Kirkwood T, Mirazon Lahr M, McNamara J, Metcalfe N, Monaghan P, Spencer H, Sultan S.
Developmental Plasticity and Human Health. Nature 2004, 430: 419-
117. Hales CN, Barker DJ. The thrifty phenotype hypothesis. Br Med Bull 2001, 60:5-20.
118. Poehlman ET, Toth MJ, Goran MI, Carpenter WH, Newhouse P, Rosen CJ. Daily energy
expenditure in free-living non-institutionalized Alzheimer’s patients: a doubly labeled water study.
Neurology 1997, 97:997-1002.
119. Poehlman ET, Dvorak RV. Energy expenditure, energy intake, and weight loss in Alzheimer
disease. American Journal of Clinical Nutrition 2000, 71(2): 650s-655s.
120. Williams DR. Epidemiological and geographic factors in diabetes. Eye 1993, 7(2): 202-204.
121. Li L, Holscher C. Common pathological processes in Alzheimer disease and type 2 diabetes: A
review. Brain Res Rev 2007, 56(2):384-402.
123. Kahn SE, Andrikopoulos S, Verchere CB. Islet amyloid: a long-recognized but underappreciated
pathological feature of type 2 diabetes. Diabetes 1999, 48:241-253.
124. Neumann KF, Rojo L, Navarrete LP, Farias G, Reyes P, Maccioni RB. Insulin resistance and
Alzheimer’s disease: molecular links & clinical implications. Curr Alzheimer Res 2008. 5(5):438-437.
125. Owen OD, Caprio S, Reichard GA Jr, Mozzoli MA, Boden G, Owen RS. Ketosis of starvation: a
revisit and new perspectives. Clin Endocrinol Metab 1983, 12(2): 359-379.
126. Clark WR. A means to an end: The biological basis of aging and death. New York: Oxford
University Press; 1999.
127. Leifer BP. Early diagnosis of Alzheimer’s disease: Clinical and economic benefits. J Am
Geriatrics Society 2003, 51:S281-S288.
128. Jellinger KA. Alzheimer 100 – highlights in the history of Alzheimer research. J Neural Trasm
2006, 113:1603-1623.
129. Foley RA, Lee PC, Widdowson EM, Knight CD, Jonxis JHP. Ecology and Energetics of
Encephalization in Hominid Evolution. Philos. Trans. R. Soc. London, Ser. B 1991, 334:223-232.
130. Wada H. Evolutionary history of free-swimming and sessile lifestyles in urochordates as deduced
from 18S rDNA molecular phylogeny. Mol Biol Evol, 1998 15:1189-1194.
131. Dennett DC. Consciousness Explained. Boston: Little, Brown and Company; 1991.
To continue to section VI click here.