Optimizing Brain Performance: Identifying mechanisms of adaptive neurobiological plasticity

Abstract

Although neuroscience research has debunked the late 19^th century claims suggesting that large portions of the brain are typically unused, recent evidence indicates that an enhanced understanding of neural plasticity may lead to greater insights related to the functional capacity of brains. Continuous and real-time neural modifications in concert with dynamic environmental contexts provide opportunities for targeted interventions for maintaining healthy brain functions throughout the lifespan. Neural design, however, is far from simplistic, requiring close consideration of context-specific and other relevant variables from both species and individual perspectives to determine the functional gains from increased and decreased markers of neuroplasticity. Caution must be taken in the interpretation of any measurable change in neurobiological responses or behavioral outcomes, as definitions of optimal functions are extremely complex. Even so, current behavioral neuroscience approaches offer unique opportunities to evaluate adaptive functions of various neural responses in an attempt to enhance the functional capacity of neural systems.

1. Introduction

In general, optimal performance is viewed as the ability to achieve desired outcomes, especially in the context of goal-directed tasks (Paulus et al., 2009). Returning from a foraging excursion with sufficient food or successfully escaping a threatening predator are examples of adaptive, optimal performance that enable animals to survive and thrive in their natural habitats. As habitats change, flexible response systems, accompanied by modified neuroplasticity, lead to successful adaptations necessary for survival (Kempermann, 2012; Rymer et al., 2013). Multiple neurobiological perspectives offer insights into the mechanisms underlying optimal brain performance. In this review, environmental characteristics associated with adaptive biobehavioral outcomes are explored, as well as the various mechanisms of neuroplasticity underlying these context-specific adaptations. Although adult neurogenesis is often emphasized, it is important to recognize that there are other manifestations of neuroplasticity including but not limited to, altered morphology (e.g., increased dendritic spines, modified dendritic branching), altered neurophysiological functions such as Long-Term Potentiation, and modified neural networks. Alterations in the processing of emotional responses are also explored, especially in the context of complex hippocampal functions, to determine factors associated with enhanced resilience against the onset of allostatic load or enhanced wear and tear on physiological functions that deter optimal neuroplasticity-related responses (McEwen and Gianaros, 2011). Variations in these systems, including sex and individual differences, are considered, as these variables provide important clues about the modulating systems that maintain adaptive neural functions. Finally, the concept of cognitive and neurogenic reserves is considered in the context of adaptive functions such as optimal contingency processing and behavioral flexibility especially, but not limited to, contexts characterized by adverse conditions (Lambert, 2018; Prakash et al., 2011; Vanenzuela & Sachdev, 2009; Kempermann, 2008).

2. Optimal Environmental and Neurobehavioral Landscapes

It is well-established that enriched, complex environments alter neural structure and function (Rosenzweig et al., 1996; Volkmar and Greenough, 1972; Bardi et al., 2016; Lambert et al., 2016), even enhancing rates of adult neurogenesis (Kempermann et al., 1997; Garthe et al., 2016). Although an emphasis has been placed on the complexity and novelty of the environmental landscape in these studies (vag Praag et al., 2000; Kuhn et al., 2017), early thoughts about neural adaptations by 18^th century Italian anatomist Michele Vincenzo Malacarne highlighted the importance of activity in neural adaptations (described in Rosenzweig et al., 1972). This observation has been confirmed in subsequent studies (Park and Bischoff, 2013; Kolb and Muhammad, 2014). The complexity of the relationship between environmental influences and activity levels was suggested in an unpublished study in which pigs housed in a barren, rather than enriched, environment actually had more rather than less dendritic branching in the somatosensory cortex. This surprising result prompted subsequent observations revealing that the barren-housed male and female pigs were more active, engaging in more belly nosing responses (flank nudging directed toward other pigs) than the enriched animals (Grandin, 1989; Grandin and Johnson, 2009). Although these results were inconsistent with previous enriched environment studies (Bennett et al., 1964; Diamond et al., 1964), this study serves as a reminder that the relationship between environments and neuroplasticity responses is far from simplistic—with behavioral interactions, as well as exposure to novelty, playing critical roles in adaptive and maladaptive plasticity. In her book Animals Make us Human, Grandin stated, “What makes dendrites grow was the animals’ behavior and actions in its environment.” (Grandin and Johnson, 2009, p. 12). In support of these observations, subsequent research indicated additive effects between physical activity and enriched environments (Fabel et al., 2009). Thus, it is important to consider the relevance of context, as well as intra- and interspecies characteristics, in the assessment of environmental and behavioral influences on neural functions (Kovesdi et al., 2011).

Related to another form of enriched sensory experiences, the impact of repetitive responses on neural plasticity was also observed in lactating rodents, as reported by Michael Merzenich and his colleagues (Xerri et al., 1994); specifically, rodents receiving ventral stimulation through nursing pups also exhibited enhanced neural plasticity of the somatosensory cortex. Although neuroplastic effects have often been observed in the visual cortex (Volkmar and Greenough, 1972), enriched environment studies have identified neuroplasticity in various cortical areas, such as the auditory and previously mentioned somatosensory cortical areas (Greenough et al., 1973; Markham and Greenough, 2004). Thus, the neuroanatomical effects of enriched environment exposure are quite prevalent across the brain, opening the door for environmental-induced modifications that can potentially enhance adaptive neural functions.

2.1. Neurobehavioral affordances and optimal outcomes

Although enriched environments provide sensory stimulation in the form of varied textures, colors, sounds, smells, and odors, they also provide new opportunities for animals to interact with physical and social stimuli. Interestingly, enriched environments constructed in the laboratory are more similar to an animal’s natural habitat whereas the standard laboratory environments are more representative of an impoverished environment (see Lambert, 2019 in this issue). In a natural, complex habitat, animals interact with various components of the environment, leading to an enhanced behavioral repertoire that informs future adaptive responses. Ecological psychologist James Gibson introduced the term affordances to refer to actions prompted by aspects of the physical environment (Gibson, 1986). For most humans, coffee mugs afford being picked up by the handle, chairs afford being sat in and ladders afford climbing (de wit et al., 2017). Experience builds cognitive capital, or reserves, that inform decisions and subsequent behaviors in adaptive ways (Sterelny, 2012). Thus, varied interactions with the environment expand behavioral repertoires and the potential for optimal performance.

In the Lambert lab, interactions with naturalistic physical stimuli in complex environments appear to alter behavioral and social affordances more dramatically than observed in male rats housed in artificial-enriched environments (Lambert et al., 2016). Compared to rats housed in environments with artificial and manufactured stimuli/toys, rats housed in an environment with natural stimuli such as dirt, rocks, and sticks interacted with the environmental objects during the dark phase approximately 50% more than the traditional enriched animals. In this context, it appears that natural stimuli invite more interactions and affordances than manufactured, artificial stimuli. Further, although the topic is beyond the scope of this review, more natural environments may alter an animal’s microbiome in ways that enhance adaptive responses; thus certain bacteria may serve as “psychobiotics” (Sarkar et al., 2016; Deans, 2017). Regardless of whether objects are natural or artificial, increased interactions with physical stimuli have been observed to alter cognitive responses in adaptive ways, including the acquisition of flexible and adaptive strategies in different spatial tasks (Leggio et al., 2005). Accordingly, environmental contexts that enhance functional interactions between the animal and environment enhance adaptive affordances in future contexts (de wit, 2017). Additionally, considering the large proportion of the brain associated with movement (e.g., cerebellum, motor cortex, and striatum), physically active brains facilitate the networking of adult-born neurons in the hippocampus (Vivar et al., 2016), often leading to activity-dependent neurogenic reserves (Kempermann, 2008). In humans, physical tasks such as playing tennis require considerable physical training to enable the brain to generate highly skilled and competitive action-outcomes, merging physical and cognitive domains in the generation of optimal performance (Wolpert and Flanagan, 2010). Thus, more varied and diverse environments increase behavioral affordances, leading to optimal action-outcomes in the same manner as diverse major histocompatibility complexes increase immunological recognition and health outcomes (Sommer, 2005).

2.2. Emotional regulation and optimal responses

When individuals are exposed to extreme environments characterized by threatening psychological and physical conditions, optimal performance is compromised (Paulus et al., 2009). The stress response associated with these challenging conditions is important for survival; however, in extreme and chronic doses, stress is known to diminish brain functions involved in adaptive decisions, especially related to the hippocampus (Kim et al., 2015). Consequently, strategic tactics, such as effective coping strategies, can potentially mitigate runaway stress responses and preserve adaptive neural and cognitive performance.

Resilience refers to the ability to adaptively respond to adverse conditions and bounce back with minimal damage (APA, 2010). Neurobiological investigations indicate that individual variability in stress and resilience responses is affected by many factors, including genetic, experiential and neurobiological influences (Southwick and Charney, 2012). Research with special-forces military trainees revealed that plasma dehydroepiandrosterone (DHEA) is positively correlated with superior performance in an extreme underwater navigation challenge task (Morgan et al., 2009). DHEA is an adrenal steroid viewed as a buffer against the toxicity of chronic corticosterone exposure (Kimonides et al., 1998; Jin et al., 2016). Compared to standard controls, male rats exposed to both natural and artificial enriched environments (see Figure 1A, B and C) exhibited less anxiety in the presence of a novel object; however, in more extreme conditions of a predator threat, the natural enriched animals exhibit less anxiety than the artificial enriched animals. Further, when exposed to a rodent version of the military water navigation task (see Figure 1D,E), both natural and artificial enriched rats exhibited less basolateral amygdala fos activation when compared to standard housed animals (interpreted as less fear); however, the natural-enriched rats had more fos activation in the nucleus accumbens core than the standard housed animals (interpreted as altered motivation; Lambert et al., 2016). DHEA responsivity has also been associated with animals exposed to naturally-enriched environments. Naturally enriched male rats exhibit higher DHEA/CORT ratios (associated with healthier stress hormone profiles) following exposure to a forced swim task (Bardi et al., 2016). Additionally, increased social interactions and accompanying increases in oxytocin-immunoreactive tissue in the supraoptic and paraventricular nuclei of the hypothalamus have been observed in natural-enriched male rats, effects that are also associated with emotional resilience (Neal et al., 2018). Taken together, these preclinical enriched environment models offer potential for laboratory explorations of stress responsivity and emotionally resilient neurobiological phenotypes and, accordingly, may provide insights into the maintenance of mental health profiles that promote optimal performance. Thus, in addition to building resilience against the onset of mental illness, these resilience phenotypes may also deter age-related neural decline and neurodegenerative diseases (Perneczky et al., 2019)

Figure 1.

Rats housed in natural-enriched (A) environments interact more with the physical and social aspects of the environment during the dark phase than animals housed in artificial-enriched (B) and standard environments (C). When exposed to a stress challenge water escape task (D), both enriched groups exhibit less fos-activation in the amygdala (E), suggesting heightened emotional regulation (Lambert et al., 2016).

Considering that behavioral-environment interactions modulate emotional regulation and neural resilience, attention has been directed toward models targeting training of adaptive environmental interactions. Effort-based reward training, for example, involves training animals to search for salient foraging cues in digging mounds that are repositioned for each trial. Thus, throughout training, mounds of bedding are viewed as affordances for digging to produce desired food treats. When these contingent-trained rats are examined in additional challenging tasks, adaptive learning strategies and stress hormone ratios have been observed (i.e., higher DHEA/CORT ratios) (Lambert et al., 2014; Bardi et al., 2013). Interestingly, neural data suggest that this version of contingency training mitigates neuroplasticity markers in the form of BDNF-immunoreactivity in the lateral habenula, a brain area associated with freezing behavior in rodents, a response potentially similar to dissociative responses in humans that compromise optimal performance in stressful contexts (Kent et al., 2018; Morgan et al., 2008). Interestingly, the lateral habenula has been implicated, in conjunction with the rostromedial tegmental nucleus, in diminished effort and reduced motivation---two phenotypes that are characteristic of major depression disorder (Proux et al., 2018; Proux et al., 2014). Optimal performance in the effort-based reward in male contingency-trained animals has been observed in a novel problem-solving task (Lambert, 2006; Bardi et al., 2012); additionally, the trained animals exhibit behavioral flexibility in the form of more diverse search strategies in response to a prediction error in a spatial foraging task (Kent et al., 2018). Thus, unlike learned helplessness, a condition in which an animal’s contingency history and natural passive response to fear compromises problem-solving attempts in subsequent challenges (Maier and Seligman, 2016), effort-based reward contingency training increases adaptive persistence in challenging tasks (i.e., learned persistence vs. learned helplessness; Lambert et al., 2014; Lambert, 2006). Persistent exertion of effort in this task, as well as in other environmental challenges, has been associated with adaptations in mesolimbic dopaminergic activity, contributing to the likelihood of animals achieving optimal performance by enabling them to overcome challenges encountered in the pursuit of desired stimuli (Salamone and Correa, 2018).

Optimal performance is also influenced by effective coping strategies. In humans, coping strategies that enhance self-efficacy, a term referring to the perception of one’s capacity to survive a threatening situation, are often related to optimal performance (Southwick and Charney, 2012). Alternatively, in some cases, high levels of self-efficacy have been associated with psychopathy (Schonfeld et al., 2017); thus more research on the complex interaction of self-efficacy and optimal performance is necessary. The rodent effort-based reward model, however, suggests that experiences contributing to mastery over environmental challenges modulate stress responses in ways that enhance optimal performances (Lambert, 2006). Other coping strategies such as flexible emotional responsiveness have been associated with the successful navigation of stressful situations (Waugh et al., 2011). In general, flexible psychological and behavioral approaches lead to greater success in solving problems, increasing the probability of surviving threatening situations (Southwick and Charney, 2012; Bryan et al., 2015)—although the benefits of flexibility are not always evident, especially in stable environments (Tello-Ramos et al., 2018).

Flexible coping strategies have been observed in rodents using a modified back-test that was adapted from a pig model (Koolhaas et al., 1999; Lambert et al., 2006). Male rats profiled as flexible, or variable, copers have higher NPY-immunoreactive tissue (associated with emotional resilience) in the basolateral amygdala and bed nucleus of the stria terminalis (Hawley et al. 2010). Further, compared to less flexible copers (e.g., consistently active or passive copers), flexible copers had higher NPY-ir in the CA1 and CA3 hippocampal subfields following training in a cognitive task (Bardi et al., 2012). More recently, male and female rats profiled as flexible copers exhibited higher DHEA/CORT ratios following chronic unpredictable stress; additionally, they exhibited less fos-immunoreactivity in the basolateral amygdala in response to prediction errors in a probe trial, as well as less hippocampal glucocorticoid receptor immunoreactivity following both chronic stress exposure and cognitive training (Kent et al., 2017). These results suggest that flexible coping responses are associated with adaptive stress responses and optimal performance.

In both preclinical and human investigations, valuable information about adaptive responses can be determined through close examinations of environmental and social interactions. Italian educational researcher and clinician, Maria Montessori, described demonstrable improvement in the performance of children with cognitive developmental delays through the use of repeated and targeted interactions with physical stimuli in their educational environment (Rathunde, 2001). In addition, predisposed factors, such as individual differences in coping strategies (de Boer et al., 2017), contribute to adaptive neural functions and optimal performance. Thus, it is becoming increasingly clear that an animal’s physical environmental landscape alters the accompanying neural landscape in ways that may provide intervention opportunities to achieve and maintain optimal neural functions. In humans, structural equation modelling of fMRI scans of male and female Berlin residents living in environments characterized by various physical aspects such as the presence of forest growth, green urban areas (e.g., parks and zoos), water-adjacent property and wasteland, revealed that the forest growth was predictive of a heightened integrity of the amygdala, suggesting that forest habitats have a salutogenic, rather than pathogenic, effect (Kuhn et al., 2017). The impact of environmental green space was also identified in an assessment of approximately one million Danish residents demonstrating that children growing up with more green space were 55% less likely to develop a psychiatric disorder (Engemann et al., 2019). These environment-induced neural outcomes are not universal, however, requiring the investigation of individual differences in positive adaptive outcomes.

3. Neurobiology of Individuality: Relevance for optimal functions

If experience shapes the brain, not only “functionally” (whatever that might mean) but also structurally, the result must be highly dependent on both the prerequisites and potential outcomes, as well as the actual experiences encountered by the animals. As these will vary massively between individuals, it follows that the space spanned by gene × environment influences must be vast. And indeed, it is: complex phenotypes tend to follow normal distributions in large populations. Thus, genetic and environmental contributions to individual phenotypes vary greatly. In addition, the consideration of environment (“E”) in the equation brings in a component of unpredictability, diluting genetic determinism. Interestingly, “E” also possesses the characteristic of being at least potentially modifiable by individual actions. There are two general forces in the “E” contribution: those that tend to make us similar and those that tend to make us different. The first is commonly referred to as the shared environment, the second as the non-shared environment, referring to that part of environmental influence that is not shared between members of a population with a common experience. In this case, the distinguishing feature is by itself not independent of genetic predispositions and the shared environment (in that different environments might elicit different scopes of behaviors); however, the key emphasis lies on how the individual responds to extrinsic stimuli and, in addition, how the individual acts within the given environment (as discussed in section 2). The non-shared environment consequently comprises the impact of one’s environment-dependent activity patterns on the critical and defining aspects of an individual’s development. This concept was conceived in 1987 in an article asking “Why are children in the same family so different from one another?” (Plomin and Daniels, 2011), which resulted in considerable discussion (Plomin, 2011), but by and large became generally accepted and influential (Turkheimer and Waldron, 2000). Nevertheless, neurobiological research on the phenomenon has been sparse, with the distinction between shared and non-shared environmental influences rarely considered.

The reluctance to consider the finer elements of environmental influences is surprising given that adaptive neuroplasticity influences both similarities and distinctions among individuals in a given species. The tension between factors contributing to similarities and distinctions is emphasized in the disciplines of psychology, sociology, economy, law, and political sciences; however, these forces must also have a massive influence on neurobiological concepts, as both literally shape the brain. The fine architecture of mammalian brains depends on genetic factors, shared environmental influences and behavioral activity. These contributing factors dictate that each individual achieves a unique collection of neural circuits, which reflects not only inheritance but also past and current experience and action. The reason for the historical emphasis on between-group effects over within-group effects is that a dominating intention of neurobiological research in the past decades has been to unravel the role of individual genes on brain development or function. In this context, variation must appear as a confounder. Efforts to standardize environments have aimed at reducing variability in order to measure the genetic effects. On the other hand, heritability cannot be estimated if there is no variation.

The neurobiology of individuality is still a rather small sub-specialty of the larger field of neurobiology that is interested in environmental effects. The massive literature on the effects of “enriched environments” on the brain has been largely one of mean effects, but, in the Kempermann lab, we argue that the paradigm has a large and underused potential that should be tapped for valuable information related to individual variation on many levels (Kempermann, 2019). The key question is how, under the condition of constraining both the genetic background (through the use of inbred animals) and the nominal environment, individuality might still emerge (See Figure 2).

Figure 2.

Individuality in brain plasticity (as in other complex traits) is dependent on genetic and non-genetic factors. Non-genetic, i.e. environmental factors can be either shared or non-shared. It is the non-shared environment that tends to increase differences between individuals. The non-shared environment reflects, among other aspects individual activity and responses to the extrinsic stimuli. In experimental conditions, both the genetic background and the shared environment can be kept constant (indicated by the clamps) to expose the non-shared environmental effects (Kempermann, 2019).

While variability here becomes the feature of interest, mean effects remain important and often go hand in hand with changes in variance. If the mean rises, there might be an associated increase in variance. But these interdependent effects nevertheless create individuality by emphasizing and augmenting inter-individual differences. At the same time, variability can also change independently of mean effects and both factors can remain independently unaffected. As we have seen in a multivariate study on the effects of environmental enrichments in female mice, the covariance matrix itself appears to be subject to the influence of environment, shared and non-shared (Körholz et al., 2018). In a longitudinal study focusing on individualizing effects, rather stable behavioral trajectories developed, which correlated with adult hippocampal neurogenesis as a highly visible and measurable type of neuroplasticity (Freund et al., 2013). Initial longitudinal behavioral assessments revealed that more active female mice (which in another study had more hippocampal neurogenesis) were more individualistic in the expression of playful responses with respect to their social behavior (Freund et al., 2015).

Indirectly, these longitudinal results likely indicate what had also been suggested in the context of studies exploring brain individualization in humans: i.e., that affective behavior and emotions represent a key factor determining the emergence of individuality. Related to such considerations, a long-term study on avoidance behavior and risk-taking in inbred male mice demonstrated small variation in the innate aspects of fear, but substantial individuality in the acquired response to the adverse stimulus (Hager et al., 2014). Such findings begin to provide a neurobiological foundation of the individuality that is found in observations of optimal performance in the form of human resilience to psychopathology. Ultimately, there is no reason, however, to limit that idea to mental resilience as it is relevant for the assessment of the impact of lifestyle risk, as well as resilience factors relevant for individual susceptibility to neurodegeneration.

As the genetic background in mouse experiments has become increasingly fixed, the behavioral phenotypes that developed have been rather stable over time, suggesting that it is likely that epigenetic mechanisms play an important role in mediating these effects at the molecular level. This observation is in line with the interesting observation that, in monozygotic twins, the epigenetic signature becomes increasingly different with age (Fraga et al., 2005).

Relevant to the discussion of individual differences in optimal and vulnerable responses is the question of the degree to which genetic variations contribute to individuality. It has often been suggested that the 1% genetic diversity in the human genome would represent the genetic blueprint for individuality, but this is unlikely to be true, given the emergence of individuality even under the condition of genetic identity. On the other hand, it is obvious that genetic variation is critical. Even so, genetic diversity in irrelevant traits will still result in homogeneity of others, while genetic identity can nevertheless become the basis of phenotypic divergence. It is the interaction that matters, further influenced by the temporal dynamics of experiential events. And this interaction is to a considerable, if again variable, degree, driven by individual behavioral activity: the non-shared environment. Monozygotic twins exemplify both the dominating role of genetic determinism as well as the subtler power of the non-shared environment. Practically, these differences are of immense importance, confirming the critical aspects of previous behavioral-environment influences in the individuality of optimal responses.

Ultimately, one’s self-image is influenced by what is and is not shared with others, with the non-shared experiences likely contributing more to the establishment of adaptive self-images (e.g., self-efficacy and self-agency) than shared experiences. In a similar manner that personalized medicine acknowledges the individuality of disease and relevant treatment options, consideration of individual differences in the neurobiology of adaptive neuroplasticity offers the potential of novel approaches to achieving optimally performing neural systems that are resistant to the emergence of psychiatric and neurodegenerative phenotypes (Kempermann, 2019). As more is learned about individual variation in neurogenic and adaptive responses, however, it is important to acknowledge the shape of the trajectories that adaptive variations might take across the lifespan, as well as, for example, the influence of sex-related variablesand relevant interactions with experience.

4. Is more always better? Determining optimal neurogenesis: consideration of sex and experience

We often default to the idea that more neurons and more spines are better for improved memory. However, it is not always true that more neurogenesis or spinogenesis will lead to better functioning of the brain. For example, during development, there is an immense increase in both spines and neurons that are pruned during puberty and other time points (sexes not disaggregated: Luo and O’Leary, 2005; Petanjek et al., 2011). If it were simply the case of more synapses and neurons being optimal for function, this peri-puberty neural pruning would lead to compromised functions. Furthermore, developmental disorders, such as autism spectrum disorder and fragile X, are associated with increased neurons and dendritic spines in various cortical and non-cortical areas including the hippocampus (Vanghese et al., 2017), and these disorders are associated with suboptimal functioning of the hippocampus. It is important to recognize that it is not merely the number of spines or new neurons that are produced at a given time, but also the shape of these spines, and connections between those spines and neurons that will crucially affect the function of the synapse and cell, respectively.

The pruning of synapses during development involves the coordination of microglia, known as the resident immune cells in the brain (Paolicelli et al., 2011). Accordingly, new research suggests that microglia are also involved in synaptogenesis and neurogenesis in the adult brain. Indeed, recent research indicates that microglia are actively involved in adult neurogenesis in the hippocampus (sex of subjects not disclosed;Sierra et al., 2014). Intriguingly, how microglia contribute to adult neurogenesis and behavioral output may be distinctly different in female versus male mice. For example, both cx3cr1 −/− male and female mice exhibit reduced levels of hippocampal neurogenesis but this has opposing functional consequences between the sexes. Although cx3cr1 −/− females show improved hippocampus-dependent learning (Maggi et al., 2011), males show impaired learning (Bachstetter et al., 2011). There are other instances of opposing results between the sexes; for example, dendritic branch pruning following chronic stress results in improved hippocampus-dependent performance in female rats but impaired performance in male rats (for review see Luine et al., 2017). These contradictory findings remind us that differences in dendritic morphology may result in different functional outcomes in males versus females.

Additional fodder for the idea that more is not always associated with better function comes from studies examining the volume of the hippocampus. Hippocampal volume is often used as a proxy for brain health, and in certain situations, during aging and in disease, there are numerous reports of cell density or regional or whole hippocampal volume being positively related to better episodic memory (both sexes used but not disaggregated: Gorbach et al., 2017; O’Shea et al. 2016, Sexton et al., 2010; van Petten, 2004). It is not always clear what accounts for differences in hippocampal volume, but if there are any changes in hippocampal volume, these are likely due to changes in neuropil, cell size, and or number of cells, within the hippocampus. including neurons and glia. It is important to be aware that the hippocampus can vary in size dramatically within and across individuals, particularly with hormone fluctuations in both men and women (Lisofsky et al., 2015; Pletzer et al., 2018), pregnancy in women (Hoekzema et al., 2017) and with childhood adversity (Calem et al., 2017). Intriguingly, the effects of childhood adversity to reduce hippocampal volume may be more dramatic in women (Calem et al., 2017). Experiences that alter hippocampal volume may alter the trajectory of aging and relationship between hippocampal volume and memory. For example, testosterone levels in men are associated with larger hippocampi that are moderated by genotype (Panizzon et al., 2010) and cortisol levels (Panizzon et al., 2018). In addition, a meta-analysis indicated that volume of the hippocampus from childhood into early adulthood provides little support in the literature for “bigger is always better,” although sex as a factor was not accounted for in either study or was only conducted in men (van Petten, 2010; Molnár and Kéri, 2014). Indeed, this is likely to be disease- and experience-dependent, as larger hippocampi are associated with poorer memory performance in adults with fragile X syndrome, but conversely, larger hippocampi were associated with better memory performance in adults with a perinatal hypoxic injury in men (Molnár and Kéri, 2014). Thus, it is time that we examine our preconceived notions that greater hippocampal volume is always related to better outcomes.

Similarly, the same contention that more new neurons are related to better functioning is not always supported, and likely depends, in part, on the environment into which the new neurons are born. For example, new neurons are not created equally, as new neurons produced under positive conditions such as running show different electrophysiological properties than new neurons produced under seizure conditions in male rats (Jakubs et al., 2006). In addition, traumatic brain injury and seizures in adulthood increase neurogenesis in rodents (Cho et al., 2015; Wang et al., 2016). While likely a compensatory mechanism, this increased neurogenesis is not associated with better function (Jessberger et al., 2007). Indeed, studies have shown that new neurons directly contribute to worse cognitive performance following seizures (Cho et al., 2015), as these new neurons must integrate appropriately into the exciting circuitry to elicit optimal effects. Furthermore, Sheena Josselyn and Paul Frankland showed in a series of studies that increased neurogenesis is related to infantile amnesia during development (both sexes used but not segregated, Akers et al., 2014) and increased neurogenesis in the hippocampus following learning promotes forgetting of recently acquired memories in adult mice (Akers et al., 2014). Of course, forgetting is an important part of learning, but here again, the simple notion that more neurogenesis is responsible for better memory is challenged. Recent studies indicate that the temporal relationship between increased neurogenesis and training determines whether there is a deleterious or beneficial effect on memory, as increased neurogenesis can promote forgetting to reduce proactive interference in mice (both sexes used but not segregated:Epp et al., 2016; Gao et al., 2018). Consequently, it is important to understand that new neurons must be integrated appropriately into an existing circuitry and this integration will depend on the environmental conditions and, finally, that many new neurons may make connections that are not related to better behavioral functioning. Lastly, when considering the contribution of neurogenesis to memory it is important to understand that there are many aspects of memory processes related to encoding, storage, and retrieval of the information-- and neurogenesis in the hippocampus is likely to only be related to certain aspects of memory.

As eluded to above, there are numerous reports of sex differences in hippocampus structure and function (for review see Yagi and Galea, 2019). It is intriguing to contemplate whether optimal levels of neurogenesis for proper hippocampus functioning may be different between males and females. Work in the Galea laboratory has shown that in tasks in which males generally outperform females (Morris water maze, similar pattern separation) males also show increased survival of new neurons after task performance in rats (Chow et al., 2013; Yagi et al., 2016). However, while males achieve asymptotic performance in Morris Water Maze training faster than females, there were no sex differences in memory for the spatial location of the platform. In line with this, activation levels (measured using the immediate early gene (IEG) zif268) of new neurons was not significantly different between males and females following a probe trial (Chow et al., 2013). Intriguingly, however, better performance during training in females was linked to greater activation of new neurons in the dorsal dentate gyrus but this same relationship was not seen in males (Chow et al., 2013; see Figure 3). These findings indicate that in females the use of these new neurons in spatial tasks is coupled with performance more so than in males. Thus, it may not be the sheer number of new neurons, or how activated they are in response to memory, but whether or not they are contributing to the learning/memory itself in a positive way.

Figure 3.

A and B – the percentage of new neurons that were activated after a probe trial (following standard Morris water training), indicating that in the dorsal granule cell layer more new neurons were associated with positive learning (less swim distance swum to reach the hidden platform) in females but no such correlation was seen in males. These correlations were significantly different from each other. Reprinted with permission from Chow et al., 2013. C and D. The correlations between zif268 activation in the dorsal dentate gyrus (DG) and performance on correct similar patterns in a pattern separation task (Yagi et al., 2016). Data were not included in the original paper, however there was only a significant correlation in proestrous females (r(4)=0.89, p=0.043) but not in males (r(18)= −.08, p=0.73, significantly different from proestrous females p=0.045) or in disestrous females (r(10)=0.26, p=0.43). E. Figure indicating that a significant correlation between density of new neurons in the ventral dentate gyrus was seen in females, but not in males on adjacent or similar trials in a pattern separation task. These correlations were not significantly different. Reprinted from Yagi et al., 2016 with Permission. F. The correlation between asymptotic performance in the trace eyeblink conditioning task with density of BrdU-labeled cells in the ventral dentate gyrus. As cited in the original paper, the correlation was higher in females (r = 0.63, P = 0.09) than in males (r = 0.39, P = 0.2). Reprinted from Dalla and Shors with permission. GCL- granule cell layer, BrdU- bromodeoxyuridine, DG- dentate gyrus, NeuN- neuronal nuclei, m-metres

Sex differences in regional activation between different types of IEGs in response to both pattern separation and cue competition in the Morris Water Maze have also been observed (Yagi et al., 2016; 2017). In both tasks, Yagi and colleagues found that the dorsal CA3 region was more active in females with the IEG zif268, but males showed greater cFos activation. Zif268 is required for LTP and for learning-induced neurogenesis in the adult hippocampus, but not cFos, indicating that zif268 may be more important for learning outcomes than cFos (Jones et al., 2001; Bozon et al., 2003; Veyrac et al., 2013). These findings collectively suggest that the dorsal CA3 region (the main terminal region of the new neurons) is most active in both males and females in response to different forms of spatial memory. However, because different IEGs are active in the dorsal CA3 between the sexes, these effects may lead to different sex-specific downstream events.

Christina Dalla and Tracy Shors (2009) found that following a trace eyeblink conditioning task, in which females typically outperform males, greater ventral hippocampal neurogenesis was observed in trained females versus untrained female rats but no significant differences between training groups in males were noted. There was also a significant positive correlation between performance and density of new neurons in the ventral hippocampus in females but not in males. Intriguingly, dorsal and ventral neurogenesis is differentially correlated with performance in males versus females. Both Dalla and Shors (2009) and Yagi et al. (2016) found significant positive correlations in female performance with ventral neurogenesis in two different tasks that failed to reach significance in males (see Figure 3), while activation of new neurons in the dorsal region was positively related to training performance in females but not in males (Chow et al., 2013). These findings suggest that regional differences may play a role in sex differences or that sex differences favor one region over the other, in the contribution of new neurons to learning and memory.

Lastly, it is important to recognize that beyond sex differences, steroid hormones play a role in both neurogenesis (for review see Mahmoud et al., 2016) and hippocampus-dependent learning (for review see Duarte-Guterman et al., 2015). Only a few studies have examined the link between hormones, neurogenesis, and activation of new neurons in response to learning and memory (Barha and Galea, 2013; McClure et al.,2013; Workman et al., 2015). Chronic corticosterone reduced activation of new immature neurons in response to a probe trial in male rats, even though these corticosterone-treated males showed improved acquisition of the Morris water maze (Workman et al., 2015). In female rats, estradiol treatment increased activation in response to spatial memory in female rats (McClure et al., 2013), with no significant effect on acquisition of the Morris water maze. In both of these studies, activation differences were not necessarily a result of differences between groups in learning, suggesting differential reliance on the use of new neurons during spatial training, depending on treatment. Preliminary evidence suggests that spatial strategy users in a pattern separation task had a greater number of new neurons compared to idiothetic strategy users in the low estradiol, but not in the high estradiol, group (Yagi and Galea, in preparation). Thus, although it is tempting to think that more new neurons may be related to optimal learning and memory, the conditions (hormonal, age, genotype, sex) will impact whether or not those new neurons are integrated appropriately (Wood et al., 2011) and may depend on the type of cognition tested. Furthermore, individual differences, such as those seen between the sexes or hormone groups, may be expressed via the contribution of new neurons and the mechanisms to integrate these new neurons for optimal behavior. As conveyed in the subsequent sections of this review, the successful integration of new neurons into neural systems associated with adaptive functions builds valuable cognitive and neural reserves that enhance an animal’s probability of recruiting these circuits for optimal functions in the future. However, a word of caution about the determination of optimal performance is necessary as the multifaceted, integrative nature of neural networks is considered.

5. Enhanced neuroplasticity through fine-tuned emotional neural networks

Another important consideration of “optimized” brain performance is what particular brain functions are measured. For example, a common perspective on brain and particularly hippocampal function is that learning and memory improvements are reflective of improved brain performance. However, this statement overlooks the fact that the hippocampus subserves a range of brain functions (Andersen et al., 2007; Scharfman, 2011), which includes memory but also mood regulation. The array of hippocampal-linked functions has sub-hippocampal regional specificity. For example, the septal region of the rodent hippocampus (or temporal/posterior in the human hippocampus) is linked to memory, while the temporal region of the rodent hippocampus (or septal/anterior in the human hippocampus) is linked to mood regulation, with some overlap (Fanselow and Dong, 2010; Tannenholz et al., 2014; Jinno, 2015). Such subregional functional differences have led to a relative segregation of the research on hippocampal function. In fact, it is rare to find a study that addresses how a given manipulation influences both memory and mood. This division where most literature studies either memory or mood, but infrequently both, has been referred to as “the memory-mood divide” (Yun et al., 2016).

One reason for the field of neuroscience to challenge the memory-mood divide is that many neuropsychiatric disorders are marked by both abnormal hippocampal structure and function, phenotypes that normalize with successful treatment. However, if “function” of a brain region or circuit is too narrowly defined, opportunities for long-lasting treatment - or even for developing new therapies - may be missed. For example, in Major Depressive Disorder (MDD) and animal models for MDD, remission or recovery is often measured by improved depressive-like symptoms without consideration of whether other hippocampal functions (such as memory) are improved. To cross this memory-mood divide, or better yet to redistribute attention to all meaningful phenotypic dimensions, integrative approaches such as NIMH’s research domain criteria (RDoC) matrix can be employed (Cuthbert and Insel, 2013; Woody and Gibb, 2015).

Within the current framework of MDD as a multidimensional “circuitopathy” (Lozano and Lipsman, 2013), it is notable that preclinical work from the Eisch Lab has shown that brain circuit-specific stimulation improves indices of both memory and mood (Yun et al., 2018). Prior work in humans and mice showed stimulation of the entorhinal cortex, a brain region upstream from the hippocampus, led to improved memory. In Yun et al., 2018, they stimulated the entorhinal cortex of male and female mice using both a molecular (knockdown of a psychosocial stress-induced protein) and a chemogenetic approach and found improved memory as well as antidepressive-like effects. While excessive entorhinal cortex activity can drive seizures and lead to pathology, Yun and colleagues tuned their stimulation to well below those levels, a fact confirmed by network activity measurements. These findings emphasize the power and potential of entorhinal cortex afferent stimulation - previously well known for the ability to influence learning and memory in both rodents and humans - for MDD treatment.

One other consideration of “optimized” brain performance is whether “improvement” in a given brain function always leads to the desired outcome, or whether a more precise, agnostic phrase can be employed. For example, humans who have a diagnosis of post-traumatic stress disorder have what has been termed “excessive generalization” of spatial contexts and cues relative to control subjects (Kheirbek et al., 2012), and this generalization may lead to panic attacks and other PTSD symptoms. This is a case where “improved” generalization is not optimal. Similarly, humans who have a diagnosis of autism have better discrimination of inanimate objects (e.g., discriminating slightly different automobiles) relative to control subjects (Pallett et al., 2014). Notably, in this study, Pallett and colleagues also repeated an earlier finding that the discrimination of faces is worse in humans who have a diagnosis of autism vs. control subjects. This is a case where multiple dimensions of function are needed to assess the determination of “optimal” in the context of adaptive functions. In addition to human research, basic research also provides examples of results in which “improved” performance is not necessarily “optimized”. For example, rats with hippocampal lesions are better than control rats at discriminating a novel tone (Quinn et al., 2009), but this “improved discrimination” is not necessarily “optimized”. Thus, in our efforts as neuroscientists to understand the neurobiological mechanisms of various functions, it is important to frequently revisit the concept of “optimization,” and proceed with appropriate caution when interpretations about optimal and suboptimal outcomes are identified. Remaining vigilant of the multifaceted functions of neural networks, as observed in hippocampal functions, will facilitate the necessary caution to avoid generating sweeping claims and evaluative judgments about altered behavioral responses and relevant underlying mechanisms.

6. Building cognitive reserves to sustain optimal neural functions across the lifespan

Like a fine wine, it is important for brains to mature in an appropriate environment; however, unlike fine wines, efficient brain functions don’t generally improve with age, unless there are targeted interventions. As brains mature from infancy, repeated sensory and motor experiences build effective neural networks that enable human infants to mature from individuals dependent on others for their survival to fully operational beings, complete with fine-tuned movements, effective language capabilities, and impressive problem-solving skills (Merzenich, 2013). As sensory and motor abilities decline with advanced age, it is important to stoke these systems in ways that recalibrate neural networks and build cognitive reserves that provide resilience against the onset of age-related cognitive decline.

Research suggests that the parameters of perceived critical periods of neuroplasticity appear to be more flexible than originally defined. When female adult rats, for example, were exposed to moderate levels of auditory noise, it reversed the developmental auditory cortical plasticity trajectory by altering key factors involved in cortical plasticity—i.e., GABA inhibition in addition to BDNF and NMDA activity (Zhou et al., 2011). With renewed cortical plasticity, functional changes pursue, especially when strategic training is employed. In rodent models, age-related cognitive deficits have been reversed by training male rats to differentiate sensory inputs so that neural noise is diminished (de Villers-Sidani et al., 2010). This targeted training opens the door to the identification of effective training strategies to tune up sensory information processing that contributes to sustained optimal functioning across the lifespan. Thus, even in older animals, the potential for sensory-based neuroplasticity remains throughout development when it is strategically activated through targeted training (Mishra et al., 2014).

In humans of any age, distractibility, defined as a diminished capacity to maintain requisite focus on relevant information for goal-acquisition, threatens optimal performance (Strayer and Drews, 2004; Bock, 2008). The classic Stroop Task, an assessment of cognitive interference that requires participants to ignore irrelevant stimuli such as discounting the font color to read the color word (e.g., respond to the printed word green when it is printed in red text), has been shown to be an effective training tool for adaptive cognitive inhibition. In one study focused on older male and female adults, research participants assessed both one and two years following the initial training continued to exhibit improved performance, confirming the long-term effects of an original six-session training session (Wilkinson and Yang, 2016). Even in young adults, the ability to ignore irrelevant stimuli is related to achieved mastery in bilingualism and musical abilities (Schroeder et al., 2016).

In addition to the importance of fine-tuned inhibition of distracting stimuli, the neurochemical noradrenaline, a modulator of attention and learning, plays an essential mechanistic role in sustained effective neural responses (e.g., increased excitability and level of attention) necessary for goal-directed responses (Moran et al., 2013). Noradrenaline is also an adaptive neurochemical marker of surprise associated with unexpected events, or prediction errors, that are a necessary component of adaptive cognitive functions (Aston-Jones, 1999; Merzenich et al., 2014). With the established associations between noradrenaline and alertness, one study exposed healthy older male and female participants to Tonic and Phasic Alertness Training (TAPAT) in the form of computerized training. The results indicated that the targeted alertness training improved performance in varied tasks assessing executive functions (van Vleet et al., 2016), confirming the importance of healthy attentional and arousal systems for the maintenance of optimal cognitive outcomes across the lifespan.

Thus, the ability to maintain optimal neural functions and performance depends on the sustained ability of the brain to operate at adaptive and optimal levels without experiencing the compromising effects of disease, injury, or age-related deficits (Prakash et al., 2011; Merzenich, 2013). As observed in both animal and human models, strategic experiential training modulates sensory processing and additional neural networks that contribute to sustained optimal performance; however, more information is necessary to determine the most efficient ways to stock cognitive and neural reserves that will ward off neurodegenerative diseases and age-related cognitive decline.

Conclusions

Although behavioral neuroscience has amassed an abundance of knowledge over the past century, effective strategies for diminishing psychiatric symptoms and optimizing neural functions have been slower to emerge than observed in other medical conditions such as cardiovascular disease (Insel & Scolnick, 2006). A potential contributing factor to this delay is the challenge of generating psychiatric symptoms in various preclinical animal models (Pariante and Nair, 2013). In this review, an emphasis on adaptive, rather than maladaptive, responses offers new opportunities for understanding biobehavioral mechanisms and optimal neural functions.

Prior research has identified salient variables such as diverse environmental interactions, effective coping responses, and adaptive neuroplasticity that are critical for the production of optimal neural networks and behavioral outcomes. Rather than a one-size-fits-all characterization, close examination of unique environments and neurobiological responses emphasizes the importance of embracing individualized and context-specific perspectives for optimal responses. Further, the mechanisms contributing to mastery over various environmental challenges are influenced by variables such as gender and age. Carefully designed research is necessary to determine contexts associated with optimal increases and decreases in neuroplasticity measures, taking caution to avoid sweeping generalizations about directional outcomes as they are reported in specific research models. Because of its noted capacity for structural plasticity, a thorough understanding of hippocampal processing of not only memory but also emotional, expression and other relevant functional outcomes, is necessary for a comprehensive understanding of hippocampal-directed neuroplasticity that leads to adaptive outcomes. Considering research noting transgenerational effects of various environmental contexts in young adult male mice (Dias and Ressler, 2014), it would be interesting to consider the transmission of context-specific adaptive functions across multiple generations in future considerations of optimally performing brains.

The dynamic nature of neural responses across the lifespan provides opportunities for modifications through targeted environmental exposure or therapeutic activation/training. For example, related to the regulation of the stress response, George Bonanno and his colleagues have introduced the term regulatory flexibility as an adaptive coping response that has a learning component. His lab has identified key aspects of effective coping strategies, including an ability to read an environmental context, the possession of a diverse repertoire of behaviors, and the ability to recalibrate responses based on relevant feedback (Bonanno and Burton, 2013; Southwick et al., 2014). Thus, the elements of regulatory flexibility enhance the potential of perceived affordances in an animal’s habitat (as discussed in this review), leading to a more diverse arsenal of optimal responses. Although the late 19^th century claim attributed to American psychologist William James suggesting that humans only use a small portion of their brains has been appropriately deemed a myth (Beverstein, 1999), current neuroscience research suggests that a comprehensive understanding of the complexity of neuroplasticity in the context of bidirectional signaling between the environment and brain may lead to a higher functional capacity of all brains. If interpreted with appropriate caution, an enhanced longevity of healthy, high-functioning brains may be achieved---providing conditional support for the long-lived hopeful myth that neural circuits can be calibrated for a higher functioning capacity across the lifespan.