Priorities in stress research: a view from the U.S. National Institute of Mental Health

Abstract

The mission of the National Institute of Mental Health is to transform the understanding and treatment of mental illnesses through basic and clinical research, paving the way for prevention, recovery, and cure. In consultation with a broad range of experts, the NIMH has identified a set of priorities for stress biology research aimed squarely at creating the basic and clinical knowledge bases for reducing and alleviating mental health burden across the lifespan. Here, we discuss these priority areas in stress biology research, which include: understanding the heterogeneity of stressors and outcomes; refining and expanding the experimental systems used to study stress and its effects; embracing and exploiting the complexity of the stress response; and prioritizing translational studies that seek to test mechanistic hypotheses in human beings. We emphasize the challenge of establishing mechanistic links across levels of analysis to explain how and when specific and diverse stressors lead to enduring changes in neural systems and produce lasting functional deficits in mental health relevant behaviors. An improved understanding of mechanisms underlying stress responses and the functional consequences of stress can and will speed translation from basic research to predictive markers of risk and to improved, personalized interventions for mental illness.

Keywords:

• Psychiatry
• neurobiology
• translation
• cognition
• emotion
• model systems
• neurodevelopment
• sex differences

Introduction

The mission of the National Institute of Mental Health is to transform the understanding and treatment of mental illnesses through basic and clinical research, paving the way for prevention, recovery, and cure. As part of that mission, NIMH identifies critical gaps in our knowledge base, designates high priority areas for research, and encourages investigators to pursue innovative, rigorous, and high impact projects. In this context, and in consultation with a broad range of experts in the field of stress biology research, we have undertaken a 2 year process of identifying and articulating a set of priorities for stress research aimed squarely at creating the basic and clinical knowledge bases for reducing and alleviating mental health burden across the lifespan (https://grants.nih.gov/grants/guide/notice-files/NOT-MH-18-058.html). Here we discuss these priority areas, which include: understanding the heterogeneity of stress, including stressors and outcomes; refining and expanding the experimental systems used to study stress and its effects; embracing and exploiting the complexity of the stress response; and prioritizing translational studies that seek to test mechanistic hypotheses in human beings. Addressing these priorities will result in a program of basic and translational research that considers the contexts, timepoints and mechanisms whereby adverse experimental manipulations and environmental exposures lead to neurobiological or physiological dysfunction and subsequent psychopathology, setting the stage for novel treatments and preventative interventions.

The heterogeneity of “stress”

The word “stress” is an umbrella term, with several usages in both the scientific literature and conversational English. Because a lack of semantic precision can impede rigorous scientific discussion, it is important to first clarify the scope of biobehavioral stress research we are attempting to address. “Stressors” are heterogeneous, cannot be reduced to single experiences, nor can their impact be explained by uniform processes. Research seeking to facilitate an understanding of the brain mechanisms underlying stress and its potential pathophysiological outcomes must therefore be similarly heterogeneous.

Basic stress neurobiology research investigates the effects of stress-inducing experimental manipulations on brain mechanisms and functional processes contributing to cognition, emotion regulation, motivation, and social behaviors. The properties of adverse experimental manipulations and environmental exposures can vary in type, timing, intensity, duration, chronicity and context. Neurobiological, physiological and behavioral responses are both dynamic and multifaceted. Since paradigm-specific manipulations and exposures may lead to unique responses in different species and populations (Tractenberg et al., 2016; Bonapersona et al., 2019; Shields et al., 2017; Loi et al., 2017; Cavigelli et al., 2018), it is critically important to optimize experimental designs to address the specific experimental questions.

In translational and clinical studies in humans, stress biology research includes investigations of the functional impact of adverse environmental exposures, including subjective perceptions of such exposures across the lifespan. Adverse and stressful exposures include, but are not limited to, neglect, abuse, social isolation, major life events, discrimination, exposure to violence or natural disasters, and can occur across a continuum of intensities, durations, and environments.

To capture this heterogeneity while facilitating growth of a cumulative knowledge base, we need not only a diversity of approaches, but also precision in experimental design and terminology. Specifically, when designing and publishing stress biology research experiments, it is useful to distinguish independent from dependent variables and to provide a clear rationale for choices of stress perturbations or environmental exposures. The quantitative parametric properties of the adverse manipulations and exposures, as well as the neurobiological, physiological and behavioral responses measured should be clearly described. In this way, data can be aggregated across studies to understand the differential impacts of different kinds of stressors.

Finally, behavioral outcome measures are most useful when they directly relate to the hypothesized neural circuits, molecular signaling pathways, and/or physiological systems investigated. In this way, the statistical hazards of fishing expeditions can be avoided. Even more importantly, formulating and testing these hypotheses through basic research sets the stage for testing their translational relevance in clinical syndromes.

Studying stress in experimental systems

Noteably, basic studies aimed at identifying the mechanisms by which stressors lead to enduring changes in neural systems and produce lasting functional deficits requires basic research in both humans and in non-human model systems. However, multiple inherent limitations exist when attempting to study psychopathology in non-human animals (Bale et al., 2019). NIMH has articulated its perspective on animal neurobehavioral approaches in mental health research (https://grants.nih.gov/grants/guide/notice-files/NOT-MH-19-053.html). This perspective is predicated on the notion that emotions and thought processes accessible in humans only by self-report or through clinical diagnoses (e.g. depressed, anxious) are difficult or impossible to identify in animals. It encourages hypothesis-driven, mechanistic studies of the underlying behavioral components of these processes (e.g. anhedonia, motivation, or defensive avoidance) rather than attempts to model specific psychiatric syndromes through untestable categorizations or subjective descriptions of animal behavior.

These priorities are especially salient for stress research. Conceptualizing pre-clinical stress research as modeling DSM diagnoses or even “depression- or anxiety-like” behaviors is potentially problematic. Focusing on these syndromes or descriptors emphasizes face validity–correspondence to clinical phenomena–over utility–the ability to enhance our mechanistic understanding of stress mechansisms. Moreover, categorizing animals as generally “susceptible” to stress or describing particular behavioral responses as “maladaptive” can lead to over-simplistic interpretations of complex systems. Instead, when considering the multi-faceted impacts of stressors on brain and behavior, behavioral assessments in animals would be better focused on specific responses within mental health-relevant domains of function (see also RDoC). Given that outcomes can vary in type and severity, responses should be measured parametrically whenever possible, to maximize the interpretability and translational relevance of the data.

The translateability of this hypothesis-based, mechanistic approach can be enhanced by cross-species research, which has the potential to determine which of the myriad of stressor effects on brain are most relevant to specific emergent functional deficits in mental illnesses. Several objective measures of adverse consequences of stress are prominent across species and may serve as useful pillars for building translational bridges between basic and clinical research. For example, there is evidence that conserved brain circuits might underly aspects of reward processing that are sensitive to exposure to stressors across development in animals and humans (Novick et al., 2018). Stress-induced deficits in attention, working memory and cognitive flexibility have been directly linked to neuronal and synaptic plasticity in rodent medial prefrontal cortex and more indirectly associated with changes in prefrontal structure and function in humans (Radley et al., 2006; McKlveen et al., 2015; Jett et al., 2017; Girotti et al., 2018). However, much more work is needed to develop a full picture of when and how stress effects align across species.

Accordingly, to determine the extent to which mechanistic processes contributing to the effects of stress exposures in animals are conserved in humans, we need neurophysiological and behavioral measures that can be used across rodent, non-human primate, and human populations. To this end, NIMH has developed funding initiatives to encourage the identification, development, optimization and evaluation of computational models, fine-grained behavioral measures, and in vivo circuit-based assays for use in cross-species research (PAR-19-289; PAR-19-214; https://grants.nih.gov/grants/guide/rfa-files/rfa-mh-19-240.html). Investigators are also encouraged to take advantage of every opportunity to collaborate with their counterparts across the animal-to-human and basic-to-translational research space.

Embracing complexity

An important consideration in stress biology research, whether in experimental animals or in human beings, is giving consideration to the wide variety of biological variables that could influence experimental outcomes. Experiments must be quite carefully designed to tease out the main effects of population-level biological variables, such as sex and age, and to understand the complex interactions between individual factors and environmental contexts. Stress exposure history, stress type, intensity, timing and chronicity will have different effects on brain physiology and development as well as behavioral effects that depend on sex, age of exposure, and other genomics factors.

A substantial body of literature documents a sex bias in the incidence of mental illnesses, such as depressive and anxiety disorders, that are particularly sensitive to adverse experience (Gater et al., 1998; Rubinow & Schmidt, 2019). Experimental stressors have repeatedly been shown to lead to sexually dimorphic neurobiological and behavioral responses in both animals and humans (Galea et al., 1997; Cahill, 2003; Bangasser & Valentino, 2014; Manoli & Tollkuhn, 2018; Rincon-Cortes et al., 2019). Despite this well-established phenomenology, many investigators have ignored females in their studies of stress, and substantial gaps remain in our understanding of the mechanisms that contribute to sex differences (see also Clayton & Collins, 2014).

Any full understanding of stress biology and mental health also requires critical consideration of developmental context, sensitive periods, and neural plasticity (see also https://www.nimh.nih.gov/about/advisory-boards-and-groups/namhc/neurodevelopment_workgroup_report_33553.pdf; https://grants.nih.gov/grants/guide/pa-files/par-19-027.html). Because neural circuits, physiological systems, behavioral responses, and coping strategies change across the lifespan, the impact of adverse experiences and exposures depends upon both the neurodevelopmental stage of the exposure and the time at which outcomes are measured. For example, in rodents, the timing (and type) of social isolation determines the specific types of reward processing deficits observed: maternal separation during infancy produces long-term deficits in willingness to work for reward, while adolescent social isolation causes increases in incentive motivation mixed with deficits in reward learning (Novick et al., 2018). The impacts of some experiences or exposures may be measurable immediately, while others may not be manifest until later in the lifespan (Lupien et al., 2009; Bale & Epperson, 2015). For example, exposure to childhood adversity is associated with changes in functional network connectivity and rates of depression at menopause (Epperson et al., 2017; Shanmugan et al., 2017). Moreover, in both rodents and humans, exposure to stress during perinatal, early childhood or adolescent periods impacts developing males and females differently (Sandman et al., 2013; Bale & Epperson, 2015; Oyola & Handa, 2017; Kim et al., 2017; Chan et al., 2018; Goodwill et al., 2018).

Individual factors beyond sex and developmental context can also impact the intensity, phenotypic expression, and persistence of stress responses. Behavioral outcomes can depend upon an individual’s genetic background, prior experiences, and physiological status at the time of an adverse experience. Genomic factors, such as inherited genetic risk, and epigenetic modifiers influence stress neurobiology in model systems and in humans. Immune activation differences resulting from prior exposures can regulate the neural and behavioral impact of subsequent stressors (Kerr et al., 2005; Clark et al., 2014; Davidson et al., 2018; Deak et al., 2017; Deslauriers et al., 2017; Pape et al., 2019). Stress experienced during pregnancy could not only alter maternal hypothalamic pituitary adrenal axis (HPA) axis and immune function, but also shape neurodevelopmental trajectories and stress responsiveness in offspring (Jašarević et al., 2017; Carlson et al., 2018; Davidson et al., 2018; Gustafsson et al., 2018; Rasmussen et al., 2019; Spann et al., 2018; Thompson et al., 2018; Graham et al., 2019).

The environmental and social contexts in which an adverse experience occurs impact the range of responses observed. Deprivation or an impoverished environment can lead to different outcomes than direct threat in an otherwise normative setting (McLaughlin et al., 2014; Dennison et al., 2019). Vivarium conditions such as group housing, cage type, and ambient temperature modify stress responses in animals (Hankenson et al., 2018). Multiple stress-induction paradigms in animals and humans use social stressors, and social supports have been shown to buffer against negative health outcomes (Hawkley & Cacioppo, 2010; Beery & Kaufer, 2015; Allen et al., 2017; Masis-Calvo et al., 2018; Kiyokawa & Hennessy, 2018). Activation of threat and fear circuits during neurodevelopment shapes emotional processes, and the impact of such experiences can be buffered by physical cues from a caregiver (Callaghan et al., 2019).

Ideally, investigators would embrace this complexity and see it as an opportunity rather than a barrier. Consideration of the mechanisms by which multiple variables contribute to increased stress-associated clinical risk for psychopathology and the development of functional deficits, as well as those that serve as possible protective factors, has the potential to reveal novel pathophysiology and treatment approaches. Most noteably, consideration of these complexities could reveal and/or validate resilience factors, such as social support or cognitive function, that may buffer the impacts of stress exposures.

Speeding Translation

In addition to the behavioral and diagnostic assessments which currently serve to identify dysfunction or psychopathology, psychiatry sorely needs additional non-behavioral biomarkers to improve early detection and prevention of mental illness (IOM, 2008). Unfortunately, for the stress field, as with much of psychiatry, the availability of useful biomarkers remains largely an aspirational goal. Traditionally, investigations of stress-related biomarkers have focused primarily on acute stress responses, including the role of the HPA axis, corticosteroid effects, and coordinated sympathetic arousal (e.g. Joels et al., 2007; Finsterwald & Alberini, 2014). However, despite a significant body of literature from both animal and human studies, cortisol measures and dexamethasone suppression tests have not proven to be clinically useful or reliable (Nierenberg & Feinstein, 1988; Holsboer, 2000).

Therefore, we need to consider additional systems and approaches, keeping in mind the critical importance of mechanisms likely to be conserved across species. For example, the immune system is an integral component of physiological responses across the breadth of stressor types. Exposure to stressful events can impact neural activity by changing the structure and function of astrocytes and microglia (Mayhew et al., 2015). Stressors also impact sleep quality and sleep microstructure across animals and humans. Sleep disruptions may, in turn, contribute to effects of stress on memory, affect, autonomic and immune functions. Several recent studies suggest that REM theta and non-REM sharp wave ripple frequencies may serve as useful biosignatures for stress-mediated alterations in brain function (Girardeau et al., 2017; Kim et al., 2019; Nolett et al., 2019).

Within neuroimmunology, there are opportunites to address key questions regarding how and when immune cells and signaling molecules cause, reflect, and/or predict long-term changes in brain activity and functional deficits following stressors. An increased understanding of the microbiome and mechanisms of communication across the gut-brain axis could help identify critical mechanistic players affecting long term stress effects. Further investigation of the relationships between specific types of stress-induced sleep disruption and clinical outcomes is also warranted. On the other hand, because of the solid existing knowledge base, there is less need for additional studies focused exclusively on HPA axis responses to stressors.

Substantial gaps remain in our understanding of how stressful experiences translate into specific risk factors and interact with protective factors to predict individual mental health outcomes. Although real-world stressors cannot ethically be manipulated in humans, multi-dimensional measures of exposure to stress or adversity can be used to create testable conceptual models and to provide converging evidence of underlying mechanisms and processes. For example, using data from a longitudinal study, Miller et al. (2018) compared several models to demonstrate that unique types of adverse experience act on different mechanisms to shape risk for psychopathology (Miller et al., 2018). Computational approaches that link multi-dimensional measures of adverse experiences, individual factors, and diverse outcomes in large samples to model risk prediction have the potential to further our understanding of the circumstances leading to adverse outcomes.

In addition to objective characteristics of adverse exposures, individual perceptions of experiences have the potential to alter clinical trajectories. Moreover, opportunities exist to identify protective factors that contribute to better-than-expected outcomes, or stress resilience. While there have been limited well replicated results, a few studies aimed at identifying clinical predictors of resilience after exposure to traumatic events have found better prognoses with cognitive reappraisal of emotional stimuli, higher baseline heart rate variability, and structural differences in the brain regions regulating these processes (Rodman et al., 2019; Carnevali et al., 2018; Walker et al., 2017). Taken together, understanding how various exposures to adverse experiences translate into clinical risk can inform both novel, personalized, intervention development as well as broad intervention strategies such as universal prevention programs or improved access to services.

Despite a multitude of potential targets, what is most missing from the translational effort in stress biology are causal tests of mechanism in human beings. An experimental therapeutics framework can also be applied to the study of stress biology and mechanisms of risk and resilience in humans (https://www.nimh.nih.gov/funding/opportunities-announcements/clinical-trials-foas/index.shtml). Such frameworks are designed to iteratively test whether an intervention results in changes in a putative target mechanism, and how those mechanistic changes relate to clinical outcomes. The expectation is that clinical trials will generate information about the mechanisms underlying a disorder or an intervention response as well as the safety, clinical efficacy, and/or effectiveness of the intervention. Studies of stress biology have the potential to point to just such mechanisms and targets.

Supporting mechanistic clinical trials that build on basic science findings in stress neurobiology research are crucial to move the field forward (https://grants.nih.gov/grants/guide/notice-files/NOT-MH-19-006.html). Ideally, these would be designed to establish mechanistic links between stress, psychopathology, and its underlying neurophysiology. For example, treatments with known efficacy at reducing stress could be used to explicitly test mechanisms of interest for different stressors, in different at-risk populations, or across different periods of longer-term follow-up. This approach has been applied to understand how alleviating maternal depression symptoms during pregnancy impacts offspring neurodevelopment and physiological stress regulation (Davis et al., 2018).

Conclusion

Ultimately, the field of stress research must seek to follow this translational pathway, identifying mediators and moderators of negative stress responses and developing them into actionable targets for intervention. A central challenge for stress biology research is to establish mechanistic links across levels of analysis to explain how and when stressors lead to enduring changes in neural systems to produce lasting functional deficits in mental health relevant behaviors. We need to bridge the gap between experimental “stress” manipulation and real-world experience of stressors, understand the mechanisms of how multiple factors interact to increase or decrease risk, and move beyond broad, proxy measures of stress to identify specific proximal factors which shape the functional impact of adverse environmental exposures and clinical trajectories. This includes supporting research which investigates the timing of sensitive periods and the degree of neural plasticity during development as well as research which will identify individual factors that predict how life events alter transitions between clinical states or lead to different mental health outcomes. Effects of stressors on reward processes, learning and neural plasticity, sleep microstructure, neuroimmune signaling, and evolutionarily conserved outcome measures represent significant research opportunities for enriching mechanistic understanding of how and when stressors produce significant adverse effects as well as those systems that may contribute to resilience. An improved understanding of mechanisms underlying stress responses and the functional consequences of stress can and will speed translation from basic research to predictive markers of risk and to improved, personalized interventions for mental illness.

Disclosure statement

No potential conflict  of interest was reported by the author(s).

Correction Statement

This article has been republished with minor changes. These changes do not impact the academic content of the article.

Additional information

Notes on contributors

Janine M. Simmons

Janine M. Simmons, M.D., Ph.D., previous Director of the Social and Affective Neuroscience Program within the Division of Neuroscience and Basic Behavioral Science within the National Institute of Mental Health (NIMH). She is currently Chief of the Individual Behavioral Processes Branch within the Division of Behavioral and Social Research at the National Institute on Aging (NIA).

Lois Winsky

Lois Winsky, Ph.D., is Director of the Integrative Systems Pharmacology Program and Chief of the Molecular, Cellular, and Genomic Neuroscience Research Branch within the Division of Neuroscience and Basic Behavioral Science within the National Institute of Mental Health (NIMH).

Julia L. Zehr

Julia L. Zehr, Ph.D., is Director of the Integrative Studies of Neurobehavioral Trajectories Program and Chief of the Developmental Mechanisms and Trajectories of Psychopathology Research Branch within the Division of Translational Research at the National Institute of Mental Health (NIMH).

Joshua A. Gordon

Joshua A. Gordon, M.D., Ph.D., is the Director of the National Institute of Mental Health (NIMH).