Signal Value of Stress Behavior:

2022, Jamie Whitehouse et al.;

Introduction:

"The experience of stress is often paired with visual cues across a wide range of species (Troisi, 1999). These include self-directed behaviours such as scratching, face touching and lip-biting (Troisi, 2002) as well as some stress-specific facial movements (Giannakakis et al., 2017; Mayo & Heilig, 2019). There is a body of behavioural, physiological and pharmacological evidence demonstrating that the experience of stress is reliably associated with the production of these behaviours (Maestripieri, Schino, Aureli, & Troisi, 1992; Mohiyeddini, Bauer, & Semple, 2015), and their study has gained significant attention in the clinical and health sciences (Mayo & Heilig, 2019; Troisi, 1999) due to their important applied benefits in psychiatry. However, science has somewhat neglected questions concerning why these behaviours evolved and what adaptive benefits they could provide to a stressed individual and those around them (Tinbergen, 1952). Thus, why stress behaviours actually exist remains a mystery. The emergence of these behaviours is not immediately obvious, as displaying overt visual signs of weakness is difficult to understand from an evolutionary perspective. Producing behaviours so strongly associated with stress could provide opportunities for an individual to be taken advantage of by others and an adaptive strategy should be to conceal stress and other weakened states. But at least for stress, this does not seem to be the case. This could reflect the cooperative nature of humans (Tomasello, 2010) and that this risk of competition from others simply does not exist. Or, it could be that the benefits gained from displaying stress to others outweigh the risks of competition, such as providing key opportunities to elicit empathy and help from others (Dezecache, Jacob, & Grèzes, 2015); social benefits which could act as a strong selection pressure. As a comparison, crying (a signal of negative affect) is known to elicit or enhance shared emotional experience (Preston & de Waal, 2002). However, all this this assumes that stress behaviours are salient to others. It may be that these behaviours are not actually perceived as associated with stress in real-world social interactions, for example, because of their significant overlap with other behaviours. The scratching and self-grooming behaviours often associated with stress also have the more basic function of removing irritations from the skin (Maestripieri et al., 1992). Such alternate and non-affective functions could make any information transfer about stress noisy and unreliable if they are also produced readily in situations where stress is not present. All this currently remains unknown, however, as the accuracy in which humans can perceive the stress of others (and what factors determine this accuracy) is yet to be quantified in detail.

Stress behaviours are frequently referred to as displacement behaviours (Mohiyeddini & Semple, 2013), a term coined by early ethologists who thought their emergence was as a consequence of displaced energy (Troisi, 2002). It was proposed that during a motivational/internal conflict (for example, when there is both motivation for conflict and affiliation simultaneously, McFarland, 1966), a third behaviour often irrelevant to the context is often produced, causing a distraction from the negative stimuli or acting as a ‘sensory cut-off’ (Chance, 1962). In humans, these behaviours appear to mainly manifest through self-directed ‘comfort’ behaviours such as self-grooming face-touching, head scratching and through behaviours through the iterative manipulation of objects such as fumbling with jewellery, and chewing on pens (Troisi, 1999). These behaviours may have proximate function to regulate the experience of stress (Mohiyeddini, Bauer, & Semple, 2013), as individuals who produce more stress-associated behaviour seem to recover from a stressful event quicker; measured through lower self-reported stress (Mohiyeddini & Semple, 2013) and through lower heart rate post stressful event (Pico-Alfonso et al., 2007). In non-human primates, there is also strong pharmacological evidence linking these behaviours to stress, and increased rates of self-scratching in monkeys are positively associated with the administration of anxiety-inducing drugs (and negatively associated with anxiety-relieving drugs) (Troisi, 2002).

The observable link between stress and behaviour suggests a communicative function of these behaviours, either as a signal that has been specifically selected and evolved, or as a phenomena occurring as a by-product of other functions (i.e., a cue: Scott-Phillips, Blythe, Gardner, & West, 2012). Although stress associated behaviours have been proposed to be communicative in the past (Bradshaw, 1993; Maestripieri et al., 1992), these inferences have been made largely from quantifying the relationship between individuals' experienced stress and their behaviour. However, to understand if the emergence of these behaviours was driven (or at least, in some part driven) by a signal function, a shift in focus from the producer of these behaviours to the psychology and behaviour of the observer (or receiver; Guilford & Dawkins, 1991) is necessary. This perspective has already been attempted with species of non-human primate (macaques: Whitehouse, Micheletta, & Waller, 2017), who were found to interact differently (i.e. more affiliatively) around individuals displaying stress behaviours, appearing to respond to these behaviours in a meaningful and adaptive way. As we share many similarities in the way stress is manifested in behaviour compared with non-human primates (many species of which have also been reported to produce self-directed behaviour during periods of stress, Maestripieri et al., 1992; Troisi, 2002), it is possible that we also share function, and that producing these behaviours affords us comparable social benefits.

People can vary in both the extent to which they produce signals (Kanai & Rees, 2011; Mayo & Heilig, 2019), and their ability to accurately read and process them (Duesenberg et al., 2016). Whether these individual differences represent noise, or if there is an adaptive reason for people to behave differently, is an unknown but interesting issue. Communicative complexity is known to be linked to social complexity (Freeberg, Dunbar, & Ord, 2012) and it is thought that those species with more complex communication are consequently able to navigate a more complex social environment. We could also hypothesise a similar relationship at an individual level and expect that someone's communicative skills are then positively associated with their social environment and their ability to maintain bigger social networks. Testing such a hypothesis with stress behaviours could be very useful. If producing or processing stress behaviours is linked to the size of an individual's social network, this would provide further evidence that stress behaviours function within a social environment and that they have an evolved signal function. Additionally, looking at the relationship between an individual's ability to display and detect stress and their social environment may even help us explain why we can observe individual variability within expressivity in general.

In order to assess for a potential signal function of stress, or more simply, if people are able to recognise the stress in others by observing their behaviour, we designed an experiment. A group of participants (termed ‘actors’ here on) were exposed to a stressful task whilst video-taped, and several indices of stress, including self-report, behavioural, and physiological measures, were taken. We presented these videos (without an explanation about the context in which they were obtained) to a larger pool of participants (termed ‘raters’ here on) who were required to give these videos ratings (on the actors' experience of stress and their likeability) and provide some details regarding their own social network. Our aim was to explore the relationship between stress, behaviour and its perception. Although a largely exploratory approach was taken, we had the following predictions: First, we hypothesised that the ratings given by the raters would align with the measures of stress taken from the actors, and that the most reliable indicator of stress in others would be displacement behaviour (i.e. stress behaviours). Second, we hypothesised that those who appear more stressed would also be judged as more likeable, a finding which would further suggest an adaptive benefit to the communication of stress (i.e., relationship building). Finally, we assessed how individual variation in stress-reading skills reflected social network size, with the idea these skills are directly linked to an individual's ability to navigate and develop their social environment. Here we predicted that better stress-reading skills would afford a larger social network."