Chronic Stress Adaptively Remodels the Cortex and Hippocampus

In a new article published by Behavioral Processes I describe how the effects of chronic stress on the brain may be evolutionarily selected. The full open-access article can be found here:

http://www.sciencedirect.com/science/article/pii/S0376635716301334

Chronic Stress, Cortical Plasticity and Neuroecology

Abstract

Prolonged psychological stress and accompanying elevations in blood cortisol are known to induce hypometabolism and decreasing synaptic density in the hippocampus and the prefrontal cortex (PFC). This article evaluates and explores evidence supporting the hypothesis that these, and other, selective effects of prolonged stress constitute a neuroecological program that adaptively modifies behavior in mammals experiencing adverse conditions. Three complementary hypotheses are proposed: 1) chronic stress signifies that the prevailing environment is life-threatening, indicating that the animal should decrease activity in brain areas capable of inhibiting the stress axis; 2) stress signifies that the environment is unpredictable, that high-level cognition may be less effective, and that the animal should increase its reliance on defensive, procedural and instinctual behaviors mediated by lower brain centers; and 3) stress indicates that environmental events are proving difficult to systemize based on delayed associations, and thus the maintenance of contextual, task-relevant information in the PFC need not be maintained for temporally-extended periods. Humans, along with countless other species of vertebrates, have been shown to make predictive, adaptive responses to chronic stress in many systems including metabolic, cardiovascular, neuroendocrine, and even amygdalar and striatal systems. It is proposed in this article that humans and other mammals may also have an inducible, cerebrocortical response to pronounced stress that mediates a transition from time-intensive, explicit (controlled/attentional/top-down) processing of information to quick, implicit (automatic/preattentive/bottom-up) processing.

Keywords: cortisol, evolution, executive control, hippocampus, neuroecology, prefrontal cortex, stress cascade, top-down processing

1.0  Chronic Stress, Cortical Plasticity and Neuroecology

Organisms throughout the five kingdoms retain certain capacities to adaptively modify their phenotype in order to better conform to their environment (Auld et al., 2010). Some of these changes are transient and reversible, whereas some are comprehensive and permanent. The studies of phenotypic plasticity, polyphenism and “predictive, adaptive responses” have shown that virtually all species can be reprogrammed by portending environmental cues, that the morphological changes are brought about by alterations in gene expression, and that the changes allow conformation to occasional but regularly recurring environmental pressures (DeWitt & Scheiner, 2004). These alternate environments typically involve stressors which demand different body types, behaviors, reproductive tactics, and life-history strategies (Pigliucci, 2001). Often the adaptive response to stress is conserved within groups of closely related organisms that inhabit similar ecological niches (Via & Lande, 1985). For instance, even though all organisms respond plastically to nutrient/energy deprivation, mammals exhibit a unique suite of physiological changes aimed at lowering the metabolism of specific organ systems in the interest of continued survival (Wells, 2009). This article discusses phenotypic changes in mammalian brain structure and neurochemistry, known to be largely mediated by alterations in gene expression, that occur in response to chronically high levels of the stress hormone cortisol. Herein, well-documented brain changes, and their behavioral correlates, are characterized as potentially adaptive responses to adverse ecological scenarios. Different lines of converging evidence will be considered in an exploratory and expository manner.

The mature mammalian brain can be reshaped by chronic or prolonged stress in two primary ways: 1) metabolic activity, dendritic growth and implicit memory are enhanced in the amygdala and caudate nucleus; and 2) metabolic activity, dendritic growth, explicit memory and inhibitory functions are reduced in the hippocampus and prefrontal cortex (PFC) (Cohen et al., 2007; Sapolsky, 2003). Many of the effects of stress on neural circuitry are mediated by the stress hormone cortisol which activates the numerous cortisol receptors present in the amygdala, hippocampus, and PFC (Morales-Medina et al., 2009). Once activated, these receptors trigger pathways that result in the expression or silencing of particular genes, which are the molecular antecedents thought to be responsible for a large proportion of the neurological remodeling (Petronis, 2000, 2004; DeWitt & Scheiner, 2004). This remodeling, much of which has been shown to be epigenetic, may help stressed mammals to adapt to environmental adversity, with its particular set of recurrent and ecologically relevant threats and opportunities.

In the literature, the responses to stress in the amygdala and basal ganglia have been attributed adaptive significance (Sapolsky, 2003), but the responses of the hippocampus and PFC have mostly eluded the attention of evolutionary biologists (Reser, 2007). Increased activity in the amygdala is thought to help animals become more sensitive and responsive to threat (Radley & Morrison, 2005). Neural and dendritic hypertrophy in the basolateral amygdala potentiates the mechanisms dedicated to identifying stressors, and mobilizing the body to address them (Sapolsky, 2003). The amygdala stimulates the paraventricular nucleus of the hypothalamus (PVN) to release stress hormones, and hypertrophy of the amygdala increases its capacity to do this (Roozendaal et al., 2009). A different way to potentiate activity in the amygdala is to release it from the structures that tonically inhibit it (Mitra & Sapolsky, 2008). The PFC and hippocampus have long been identified in neurology as brain regions capable of inhibiting the autonomic and emotional responses to fear-inducing stimuli (Cannon, 1929; Papez, 1937; MacLean, 1949; see LeDoux, 1987, for a review). This circuitry ensures that mammals can override the fear response if they make the determination that the stimulus may appear threatening but is not actually threatening (Morgan et al., 1993). Diminishment of activity in the PFC and hippocampus may ensure that the areas that incite stress, the amygdala and PVN, can function unimpeded during stressful times.

Decreased activity in the PFC and hippocampus may also adaptively influence the animal to be less cerebral and more impulsive (Reser, 2007). When facing lasting adversity, it may be advantageous to suppress the PFC and hippocampus because these areas put inhibitory pressure on defensive, instinctual, and dominant responses. When an animal experiences extreme stress, it is probable that its high-order behavioral strategies are proving relatively ineffectual (Boonstra, 2005). It may benefit such an animal to be less reliant on learned behavior, and more reliant on genetically programmed and species-specific behaviors. Hence, the changes in the hippocampus and PFC may protectively disinhibit innate and instinctual urges (Reser, 2007).

The present article will elaborate on three complementary hypotheses: 1) stress signifies that the prevailing environment is antagonistic, and that the animal should not suppress the stress response or inhibit conditioned fears; 2) stress signifies that behaviors that the animal has learned may be inefficacious or deleterious and that it should increase its reliance on innate behaviors over learned behaviors; and 3) stress indicates that environmental events are proving difficult to systemize on long time scales (using delayed associations) and thus the maintenance of contextual, task-relevant information in the PFC need not be maintained for temporally-extended periods.

Several neurological changes to areas including the amygdala, the caudate nucleus, the hippocampus, the mPFC, and the PFC in general will be discussed. Table 1 describes the general psychological consequences of these changes, the implications that they have for modern people as well as hypothetical implications that they may have had for prehistoric foragers. This table attempts to highlight the disparity between the limiting repercussions of these changes in the modern “information age” and their potentially adaptive significance in the prehistoric past.

Table 1

The Neurological Effects of Stress, Then and Now

Neurological State	Psychological Consequences	Implications for Moderns	Implications for Foragers
Amygdala hyperactivity	Potentiation of conditioned fears	Anxiety, fear and excessive stress	Healthy caution, preparedness and mobilization
Caudate hyperactivity	Potentiation of procedural or habitual movements	Intrusion of habitual or procedural responses	Increased reliance on movements that have been proven effective
PFC hypoactivity	Behavioral disinhibition	Working memory and goal-setting problems	Increased reliance on instinctual and appetitive impulses
mPFC hypoactivity	Impaired inhibition of conditioned fears	Exaggerated stress responses to nonfatal threats	Enhanced awareness of potential threats
Hippocampal hypoactivity	Inaccessibility of contextual and episodic information	Explicit/ declarative memory problems	Increased reliance on dominant and procedural responses

Interestingly, prenatal and early-life stress cause a pattern of changes that is strikingly similar to the changes that occur in response to chronic stress in adulthood (Weinstock, 2008). When pregnant rodent or primate mothers are stressed, they program highly analogous changes in the amygdala, hippocampus, and PFC of their offspring (Francis et. al., 1999; Kapoor et al., 2006; Schneider et al., 1999). The behavioral changes in these offspring, which include increased vigilance, fearfulness and stress responsivity, have been interpreted by Michael Meaney and colleagues as constituting a predictive and adaptive response to early environmental adversity (Zhang et al., 2006). In this interpretation the amygdalar changes are attributed adaptive qualities. However, the role of the hippocampus and the PFC in contributing to this behavioral response has been neglected. Moreover, psychiatric disorders such as anxiety, depression, posttraumatic stress disorder and schizophrenia are associated with prenatal and postnatal stress, and involve the same pattern of changes to the hippocampus, PFC and amygdala (Axelson, 1993; Corcoran, 2001).

Elevated levels of noradrenaline and dopamine, such as occur during acute yet transient stress, impair PFC and hippocampus-dependent abilities such as working memory and attention regulation, but strengthen amygdala, caudate and subcortical-dependent functions such as fear conditioning, habitual behaviors and reflexes (Elliot & Packard, 2008; Packard & Teather, 1998). Thus, acute stress, chronic stress, prenatal stress and a number of major psychiatric disorders have all been shown to engineer a switch from thoughtful “top-down” control based on task-relevance to bottom-up control based on salience (Buschman & Miller, 2007; Hermans et al., 2014). This article focuses on these cortical corollaries of pronounced stress, and attempts to interpret them in terms of their ecological utility to mammals, from wild rodents to prehistoric humans. If the neurological changes that respond to stress were diffuse or only degenerative this might indicate that they do not represent adaptation. That the alterations are very selective, that they completely spare critical cortical and subcortical regions, that there are dozens of documented molecular pathways that converge toward these changes (Stankiewicz et al., 2013), and that arborization and neural activity in the amygdala (Francis et al., 1999; Radley & Morrison, 2005; Vyas et al., 2003) and caudate (Kim et al., 2001; Schwabe et al., 2008) is actually enhanced, suggests that these changes may not be pathological. To further explore the proposed evolutionary rationale for why these changes, in adulthood and in utero, might constitute an adaptive response we turn to the neurobiology of stress perception.

The rest of the article can be found here:

http://www.sciencedirect.com/science/article/pii/S0376635716301334




Here are some passages from an earlier version of the manuscript that I removed:

The PFC and hippocampus are probably targeted by the stress cascade because of their subtle contributions to behavior. They are both multimodal convergence areas that integrate information from multiple sensory areas to allow sensory reconciliation and ultimately self-control and self-guided action. These two areas do not control any basic unitary ability though. Most other areas of cortex, if rendered hypometabolic, would result in the loss of motor skills, or sensory processing. For instance, pronounced dendritic shrinkage in the occipital lobe would adversely affect or even abolish certain visual abilities. Yet the neuropathology in the stress cascade spares early sensory areas and even most association areas and is confined to the more expendable, more abstract areas...

Cortical dendritic spines are highly motile structures that are capable of responding quickly to physiological and environmental input - responses that are integral to learning and behavioral plasticity (Kirov et al., 2004). Restricting their growth and actually retracting them should limit behavioral plasticity and also reduce the accessibility of information (memories) that is encoded by the synaptic weights. That the apical, but not the basal dendrites are affected (Brown et al., 2005) suggests that this response is designed to reduce the contribution of higher cortical layers. The superficial layers, II and III, are responsible for sending and receiving recurrent (feedback/ topdown/ modulatory) signals that encode subtle distinctions of learned information (Calvin, 1995). This epigenetic program spares the lower layers which receive inputs from and produce outputs to subcortical areas. Such apical degeneration could make it so that cortical areas are accessible to lower areas, only they are being used to a lesser extent. In other words, during stress it is still important to draw vertically from various cortical columns, or clusters of cells with similar receptive fields, however; it is less important to draw from the higher cortical lamina (layers II and III). This makes the content of memories accessible, but minimizes the horizontal cross-talk within higher layers, perhaps resulting in simplified answers to behavioral questions. It emphasizes bottom up at the expense of top down processing. Cognition without the horizontal information-sharing decreases the temporal persistence of information processing, potentially contributing to the documented diminishments in working memory and inhibition. It should be very interesting to better resolve how the debranching of apical dendrites affects the processing within and between individual cortical columns...

The dendritic changes in the medial PFC follow another definite pattern where larger spines are affected to a greater extent. Larger dendritic spines which are less plastic, have larger synapses, a high proportion of AMPA receptors, are stable, and are thought to represent “memory spines” (Kasai et al., 2003). Small spines are more plastic, labile, have smaller synapses with a high proportion of NMDA receptors, and can either retract or become larger and stable and these are thought to represent “learning spines.” During stress the larger spines are the ones that retract the most (Cerqueira et al., 2007), hence it should be interesting to determine how this affects the cell’s receptive field, and why. Clearly there is a biological favoritism here. It should be important to understand why larger spines should retract. This seems to be consistent with the second hypothesis described in the introduction, suggesting that stress signifies that learned behavior may be ineffectual, and that the animal should increase its reliance on instinctual and procedural behaviors...

Perhaps the stress cascade makes an animal hesitant to stray far from home to keep it safe. And perhaps it also makes it so that the animal’s inclination to go out and explore, to be an infovore and search out new areas, because it has mastered the old ones, is diminished.

It will be interesting to see how reflexes are potentiated by stress aside from hyperactivity in the caudate nucleus. I imagine that it is possible that the corticospinal system (which is controlled by the frontal cortex) is downregulated relative to the extrapyramidal motor systems. It is possible that the primary motor cortex (which commands the cortiospinal (pyramidal) system is unchanged but the areas that regulate and inhibit its activity (like the premotor, supplementary motor and presupplementary motor areas) are downregulated. In fact, the presupplementary motor cortex but not the primary motor cortex is damaged in schizophrenia. This means that stress makes you worse at planning complex sequences of movements.

The stress cascade acts to reduce the amount of information that individuals take with them through time. The PFC generally learns to pick out key stimuli from different scenarios  The more stressed an animal has been the fewer elements of its immediate past can be recalled attentively and applied to the current situation. Animal species that have been endowed with large working memories by evolution can expect that it will be helpful to have many brain areas, corresponding to the last several seconds of thought (or experience), primed at once so that they may bring elements of their immediate past to bear on the present. Less intelligent animals can take only the most salient aspect with them through time, and old, hardwired or subcortical areas are what programs most of their actions in the present.

Many animals, such as a large proportion of invertebrates, that inhabit simple ecological niches have very simple nervous systems and come into the world with nearly all of the behavior needed preprogrammed into their bodies by their genes. These animals can only learn very rudimentary things that only slightly modify their innate systems. As brain to body ratio and encephalization have increased in the animal kingdom, over the last 700 million years, many species have been granted increasing autonomy over their own behavior. More intelligent animals, like those in our phylum, chordates, have the ability to be programmed by the environment so that the behavior of any two individuals of the same species may be very different especially if they developed in differing environments.

The mammalian cortex, which is highly developed in humans, allows very high degrees of behavioral plasticity for planned actions (Moscovich, 2007). Such a high degree of self-government and freedom from instinct would potentially be very dangerous for animals such as reptiles, amphibeans and fish because they would probably often be hard pressed to develop adaptive behaviors on their own. A vastly increased ability for new learning probably developed in mammals because of the presence of maternal care and instruction. Developing mammals were unique because they had doting models from whom to learn memes. But what if a young mammal’s learning does not lead to adaptive behavior and the mother’s own stress detracts from her ability to share knowledge? Might it then be beneficial to return to instinctive behavior? If natural selection could have accomplished it, it probably would have given mammals the connections needed at birth to live, eat and procreate as it does in very simple invertebrates. Mammalian ecological niches are so variable though, that genes could never prepare them properly by preprogramming all of their behavior. For this reason higher cognitive areas are a necessity although perhaps a calculated risk.

People cite many reasons for what pressures caused the human brain to evolve: complex foraging techniques, hunting, social abilities, Machiavellian intelligence. I think that a major contributor to the evolution of human intelligence was the stress free period that resulted from the manufacture of weapons. Intelligence evolved because hominids in our direct line of descent (that we descended from lower forms is ironic) evolved to hunt in groups and fashion weapons, this increased order and decreased predation (intraspecies aggression probably stayed the same after weapons). Now that the world was safer, people had time to see order. The abilities to think things all the way through and to plan ahead were selected. The ability for self discipline and to exercise control over wants was selected for. The ability to do the harder thing and to hold out or to plan, to set things up in advance and “wait for it to be a sure thing.” Sometimes newly appreciated phenomena give us a new vantage point to see other things. I think that the stress cascade gives us valuable insight into the evolution of human intelligence and tells us that safety and order, allowed us time to deliberate which allowed selection to act on the other cognitive theories like hunting, social abilities and such. The way that you think about the things in your environment determines what you extract from them. It seems likely that, like with most other traits, there might be crucial genetic variations that ended up causing an aneurotypical way to think about or perceive the environment. ADHD, depression, PTSD and anxiety may all represent different but similarly adaptive ways to extract specific information from a presumed adverse environment. The predictive adaptive response of the stress cascade made the dangerous expansion of the human mind, not so dangerous.

Perhaps only a minority of people are neurologically well suited for occupations that demand cognitive rigor, planning, foresight and intense concentration. Since there was no electrical engineering, computer programming or rocket science in the ancestral past, sometimes it is unclear how these skills were selected for during human evolution. Perhaps neurological dispositions like these represent an optimistic outlook that expects a systemizable and orderly environment. Delayed gratification, marshmallow studies, waiting longer means more intelligence and school performance. Tie between adversity and delay. Only in a good environment can you delay. The brain evolved to learn about the things that we are motivated about. Today we learn about things that have nothing to do with survival. Hominids probably would have found it hard to get motivated about school work and occupational expertise but would have stayed interested in activities that lead directly to food, shelter and social interaction. And this would have been good for them.

The discipline of behavioral ecology is based on the premise that cognitive traits and the neural substrates responsible for them are shaped by natural selection and hence are fine tuned to respond to the traditional environment of the species in question. A subdivision of this discipline, called neuroecology, attempts to draw causal links from variations in brain structure between species to the lifestyle of these species. For instance, a species that forages over wide terrain or hides (caches) food more than a similar species would be expected to have a larger hippocampus because the hippocampus, an area responsible for spatial cognition, plays a bigger role in its ecological behavior. Neuroecology has received some criticism because the predictions that it tests are largely adaptationist (i.e. if it exists it must have been naturally selected) but since has been largely redefined.

Reality testing of past memories is difficult for one main reason, because the semantic aspects of a memory can be completely isolated from the contextual. The way that we test a memory, to see if it was imagined or not, is to try to recollect the context. I have a strong memory that my friend misheard a song lyric in a popular song. The friend heard “she is a Swedish drum” and the lyric is “she is the sweetest drug.” I asked my friend about this over a year later and my friend denied it completely. I knew that it took place and I scoured my memory… and got no where. All I could remember was that the event took place, but I had no contextual cues. I can’t remember where we were, what else we were saying, what time of day it was, or if I was alone daydreaming the scenario to life. The only other thing I could remember, besides the literal lyrics, was that it was a real memory, is this really trustworthy though? Could the “this was real” memory actually just be a mutated “wouldn’t it be funny if this had been real” memory? Maybe if it all comes back really FAST, then this indicates that it is hippocampal and thus veridical in origin. If it comes back more slowly, it could be confabulated or reconstructed.