Context and trade-offs characterize real-world threat detection systems: A review and comprehensive framework to improve research practice and resolve the translational crisis

Highlights

• In nature, antipredator decision-making is shaped by context and characterized by trade-offs.

• In the laboratory, neurobiological models of fear and anxiety control context and limit trade-offs.

• Translational science is failing because models fail to address real-world conditions.

• We develop a mechanistic model to show why contextual factors should be experimentally manipulated.

Abstract

A better understanding of context in decision-making—that is, the internal and external conditions that modulate decisions—is required to help bridge the gap between natural behaviors that evolved by natural selection and more arbitrary laboratory models of anxiety and fear. Because anxiety and fear are mechanisms evolved to manage threats from predators and other exigencies, the large behavioral, ecological and evolutionary literature on predation risk is useful for re-framing experimental research on human anxiety-related disorders. We review the trade-offs that are commonly made during antipredator decision-making in wild animals along with the context under which the behavior is performed and measured, and highlight their relevance for focused laboratory models of fear and anxiety. We then develop an integrative mechanistic model of decision-making under risk which, when applied to laboratory and field settings, should improve studies of the biological basis of normal and pathological anxiety and may therefore improve translational outcomes.

Introduction

Decision-making under risk is relevant to behavioral researchers from a variety of disciplines, including all those who study topics ranging from predator-prey interactions in the field to anxiety disorders in humans (e.g., Clinchy et al., 2011; Mobbs et al., 2018). Anxiety occurs in response to risk and is generally related to a sense of apprehension. This apprehension results, in part, because there is a conflict between a potential threat and the potential of receiving a benefit such as foraging or potential mate (McNaughton and Corr, 2004) It may be quantified as increased vigilance and can become maladaptive when it interferes with otherwise necessary behaviors that enable survival (Box 1). Risk-related decisions are thought to maximize fitness; hence, individuals that make the right decisions are those whose genes are passed on to future generations. A key insight about these decisions is the involvement of trade-offs where the costs and benefits of risky behaviors are evaluated and the optimal outcome maximizes benefits while minimizing costs. Costs primarily ensue from the risk of predation, competition, disease and parasitism (Gallagher et al., 2017). Humans, especially in modern societies, additionally live with the risks of losing social status and employment (Björkqvist, 2001). A key lesson from behavioral ecology is that in order to optimize outcomes, individuals cannot avoid all risks; by doing so it would be impossible to acquire resources or mates (Blumstein, 2008).

Both, anxiety and fear, are emotional states, associated with physiological and psychological responses. Anxiety and fear are adaptive responses if it is possible to correctly differentiate safe and threatening stimuli. If individuals are unable to distinguish between threatening and safe stimuli, anxiety and fear may become maladaptive. If such a maladaptive state lasts longer, it becomes pathological and in humans we might diagnose an anxiety disorder. Decisions about risks may vary according to a range of internal factors (Kiyokawa et al., 2009) and the external environment (Campbell-Palmer and Rosell, 2011; Orrock and Danielson, 2009) at the time a threat is detected. For example, moon phase is known to influence antipredator decision-making in several prey species. Oldfield mice (Peromyscus polionotus) and woodmice (Apodemus sylvaticus) are more likely to respond to predator cues on full-moon nights when the mice are most visible, and hence vulnerable, to their predators (Orrock and Danielson, 2009; Orrock et al., 2004). By contrast tammar wallabies (Notamacropus eugenii) increase the time spent foraging under moonlight, suggesting that they feel safer under illumination when high visibility improves their ability to detect predators (Biebouw and Blumstein, 2003). Humans are also influenced by moonlight. In Tanzania, where lion attacks are common, the full-moon causes anxiety among people (Packer et al., 2011), ostensibly because the coming nights will be darkest. Among non-vertebrates, European leeches (Hirudo verbana) modify their behavior according to how deep they are within water, and whether they have been fed a blood meal (Palmer et al., 2014; Palmer and Kristan, 2011). Thus, there are a variety of ways that context (light, depth of water, or a meal in these cases) modifies decision-making based on the assessment of these factors.

Consequently, there is widespread recognition that trade-offs, and the context in which they are made, must be incorporated into models of decision-making (Caro, 2005; Lima, 1998). However, not enough laboratory research considers these cost-benefit trade-offs and contextual variables that characterize anxiety and fear in nature (Gray and McNaughton, 2000; McNaughton and Corr, 2004). This is especially important if we are to make progress in solving a crisis in translational biomedical research (Manjili, 2013) where results from preclinical studies cannot approximate pharmacological effects in clinical studies or real-world settings (Kinsella and Monk, 2009; Oppenheim, 2019), or similarly, in conservation and wildlife management where fear-based management tools cannot approximate outcomes from fear cues in the laboratory (Parsons et al., 2018).

For instance, an isolated laboratory rodent might be exposed to a specific stimulus that could generate fear or anxiety, but the type and extent of its behavioral response is shaped by other factors, such as the animal’s current satiety status (internal drive; Lõhmus and Sundström, 2004) or the presence of further potential dangers (external context; Nersesian et al., 2012; Orrock et al., 2004; Parsons and Blumstein, 2010). Despite this potential of contextual variables to profoundly influence behavior (Wolff, 2003), many such variables are sacrificed in laboratory/neurobiological models specifically to improve power and replicability (Klumpers and Kroes, 2019). The resulting animal models allow us to use powerful approaches such as optogenetics and chemogenetics to understand neurobiological mechanisms underlying decision-making under risk. They also allow us to manipulate internal and external conditions we wish to understand, but cannot be directly manipulated in humans. Yet, these models are not intended to account for the influence of context in real-world decision making.

Here, we review and discuss the trade-offs that should be taken into account during antipredator decision-making in wild animals, and integrate them with neurobiological models of fear and anxiety. We emphasize the importance of internal drives and external contexts in decision-making via an integrative mechanistic model, for which we have summarized the neural basis for threat-induced defensive behaviors. While calls for real world models have been increasing over the past decade (Kinsella and Monk, 2009; Oppenheim, 2019), our approach is novel in that we develop a more integrative model that shows how decision-making in response to conflicting internal drives is influenced by fear and may result in anxiety. Lastly, following our review, we propose how decision-making research can be improved, and thereby facilitate new semi-realistic and naturalistic approaches. These outcomes are intended to address the growing bench-to-bedside gap in translational medicine (Manjili, 2013), while also enhancing wildlife conservation and management.