Aging rhesus monkeys stick to friends and family

Most people have fewer social contacts as they get older. Children move out, people retire, and friends die. But irrespective of these demographic events, there are also notable changes in personal preferences as individuals age. One of the most influential theories on motivational changes with age, social selectivity theory, suggests that the awareness of a finite lifetime leads to an increased focus on fewer but more meaningful social partners (1). However, a limited future-time perspective does not seem to be the only driver for a change in social motivation: Older Barbary macaques and chimpanzees, for instance, also have smaller social networks than younger conspecifics (2, 3), although they are probably unaware of the fact that they are going to die one day. Previous studies on age-related changes in sociality in nonhuman primates were cross-sectional, however, and so, it remained unclear how specifically social networks become smaller. Do nonhuman primates preferentially retain relationships with their “old friends” or are the smaller networks due to other selective forces? In a longitudinal study on female rhesus monkeys, Siracusa and colleagues now provide an answer to this question (4), adding an essential piece to our understanding of the evolutionary roots of social aging (Fig. 1).

Fig. 1.

In their study on social aging in female rhesus monkeys, Siracusa and colleagues contrasted two alternative hypotheses that could account for shrinking social networks as the animals grow older. According to the social selectivity hypothesis, individuals preferentially retain social relationships with valuable partners, such as friends and relatives, here indicated by thicker edges and a similar color as the subject in the center. In contrast, the selective disappearance hypothesis posits that animals who are more social than others are more likely to die and disappear from the network. The authors found support for the social selectivity hypothesis but not the selective disappearance hypothesis.

A profound understanding of social and motivational aging processes is vital in the light of the global demographic development of the human population. Projections indicate that by 2050, persons aged 60 y or above will outnumber adolescents and young adults worldwide (5). Nonhuman primates are valuable model systems in this context. They have a similar life history as humans, with slow development, extended infancy and adolescence, and a relatively long life. In contrast to humans, though, nonhuman primate females do not experience a substantial postreproductive phase, i.e., menopause. Otherwise, the physiological changes across their lifespan are similar to those of humans. With increasing age, nonhuman primates also experience significant changes in brain structure and physiology and marked cognitive decline (6).

Previous studies into primate senescence have mainly used life-history theory as a framework. Life-history theory allows us to formulate predictions about how organisms should allocate resources to maximize their fitness (7). Of particular interest are the trade-offs that organisms face in their development and behavior: Should they invest in current or future reproduction? Should they look out for a new mating partner or devote their attention to their offspring? In contrast, lifespan psychological theories of aging focus on cognitive and motivational processes (e.g., refs. 1 and 8). Several studies on nonhuman primates have used these theories as frameworks for studying age-related changes in social interaction patterns (e.g., refs. 2, 3, 9, and 10).

Siracusa and colleagues used social selectivity theory as their starting point to ask how monkey social networks shrink in size. They studied female rhesus macaques living on the island of Cayo Santiago, off Puerto Rico. The island is a breeding facility where the animals are food-supplemented and receive some medical care. Over many decades, the population has been a source of valuable information on rhesus monkeys’ social behavior and physiology, including a recent study of social network changes after Hurricane Maria hit the island (11). Rhesus monkeys live in “female-philopatric” societies, where females remain in the natal group whiles males disperse. In these societies, female relatives are often the most critical social partners. Rhesus monkeys are also renowned for their despotism, resulting in strict dominance hierarchies.

The authors surveyed changes in the social networks of more than 200 females between 10 and 28 y old. The subjects were tracked, on average, for 2.8 and up to 8 y. One of the challenges of working with such data sets is their heterogeneous structure. For some subjects, information was available for the time when they were 10 to 14 y old; for others, when they were 18 to 24 y old. With such data sets, it is not trivial to distinguish between within-subject and between-subject effects. For instance, one might ask whether a person’s carbon footprint increases with their income—a within-subject effect. Or, we might ask whether more affluent people have a higher carbon footprint than poorer people—a between-subjects effect. If these within- and between-subject effects are aligned, more conventional statistical methods can be used to estimate the relation between income and carbon footprint reliably. However, if the effects are not aligned, one might run into what is known as an “ecological fallacy” (12). To tease apart both components, the authors, therefore, first estimated for each subject how its behavior at a given age compared to its behavior at its average age. For example, consider a female observed from age 10 to 14 y. The authors then assessed her sociality when she was younger than her average age of 12 and when she was older.

"Siracusa et al. used social selectivity theory as their starting point to ask how monkey social networks shrink in size."

The authors considered this within-subject individual variation with age and the overall effect of age in their model. In this way, they could distinguish between two factors that could explain the reduction of the animals’ social networks: social selectivity or “selective disappearance,” according to which more social individuals would be more likely to die sooner. The authors found that female rhesus monkeys had fewer grooming partners as they aged, while their overall time spent grooming remained the same. Thus, as females aged, they focused their affiliative behavior on a smaller number of partners. Crucially, these were more likely to be relatives. Females were also more likely to interact with partners with whom they previously had interacted frequently. This latter effect, however, was independent of age. In conclusion, the authors found evidence for increased social selectivity with age in female rhesus monkey, similar to what is known for aging humans. The key advance of the present paper (4) is to add a critical longitudinal perspective that hones in on the mechanisms leading to increased selectivity.

Still, if we look at the results of the statistical models, another factor seems even more decisive than age—namely, a female’s rank. As noted above, rhesus monkeys have a steep rank hierarchy, and whether a female is high or low ranking explains more of the variation in social interaction patterns than age. These results raise some interesting questions about species differences in social aging studies. Macaques are an excellent model to address this issue, as different macaque species vary considerably in their social style, with some being more socially tolerant and others more despotic (13). For instance, our analyses of social and motivational aging patterns in Barbary macaques, a species classified as rather tolerant compared to rhesus monkeys, revealed partly similar but also some divergent patterns. While the number of social partners and time spent socializing decreased with age, the effects of rank were less strong (14). Thus, similar patterns may be modulated by the specifics of the social system. It is also important to note that motivational changes with age may take different trajectories depending on whether the focus is on social behavior, exploration, or perseverance (2, 15) and what the living conditions of the animals are (16).

The focus on reliable social partners may have adaptive value, as Siracusa and colleagues point out (4). Older individuals—particularly if they are low-ranking—are more likely to be injured. Thus, withdrawal from others might be a strategy to reduce the risk of injury and the associated mortality risk (17). Alternatively, a withdrawal from the social shenanigans in one’s group might be a by-product of having less energy. The combined effect of age and rank reported by Siracusa and colleagues may be due to older low-ranking females being in poorer shape. For Barbary macaques, we found that older animals were less physically active but that could be adaptive (to avoid injury from falling, for instance) or a by-product of lower energy levels (14). The jury is thus still out on this question.

So far, studies on age-related changes in nonhuman primate sociality and motivation have shown that a limited future-time perspective is not the only explanation for increased social selectivity. These results should be highly informative for theories on social and motivational aging in humans. Yet, we also need to be aware of key differences in methodology and in the conceptual frameworks we use. For example, many studies on humans rely on self-report data, while primatologists trudge behind their animals and note the animals’ interactions. However, data for monkeys—even if they are as extensive as the ones analyzed by Siracusa and colleagues—pale in contrast to the data available for humans. These data sets allow us to take much more differentiated looks at how interaction patterns depend on an individual’s role and life stage.

For instance, in the United States, people between the age of 20 and 50 y spend considerable time with coworkers and family. Once the children leave home, people spend more time with their partners. There is also a steady increase in time spent alone from the age of 40 y onward (18). These data not only remind us how important good relations with coworkers are for overall happiness and well-being but they also imply that changes in social interaction patterns strongly depend on a society’s economic system. Therefore, we must extend our knowledge about human social aging patterns beyond Western, Educated, Industrialized, Rich, and Democratic (WEIRD)societies to assess the effect of economic sustenance on social interaction patterns.

In summary, after a trickle of studies on old-aged primates in the last decades, including a pioneering study on the rhesus monkeys of Cayo Santiago (19), the topic of social aging is gaining traction, even beyond the primate order (20). An integrative research program that combines physiological, behavioral, and experimental studies to tap into variation in social behavior, communication, and cognition will shed light on the mechanisms underpinning social and motivation aging. The study by Siracusa and colleagues adds a pivotal piece to this puzzle.