The story of how cognition and fitness relate may not be simple, but simple stories and complex systems rarely go together

Cognition and reproductive success in cowbirds. David J. White, J. Arthur, H. B. Davies & M. F. Guigueno. Learning & Behavior, Dec 16 2021. https://link.springer.com/article/10.3758/s13420-021-00506-0

Abstract: Understanding the relationships between cognitive abilities and fitness is integral to an evolutionary study of brain and behavior. However, these relationships are often difficult to measure and detect. Here we draw upon an opportunistic sample of brown-headed cowbird (Molothrus ater) subjects that had two separate research experiences: First, they engaged in a large series of cognitive tests in David Sherry’s Lab in the Advanced Facility for Avian Research (AFAR) at Western University, then subsequently moved to the Field Avian Research Megalab (FARM) at Wilfrid Laurier University where they lived in large breeding flocks in aviaries with other wild-caught cowbirds. Thus, we had extensive measures of cognitive abilities, breeding behavior, and reproductive success for these birds. We report here, for the fist time, the surprisingly strong connections we found among these different measures. Female cowbirds’ spatial cognitive abilities correlated positively with how intensely they were courted by males, and with their overall egg production. Males’ spatial cognition correlated positively with their ability to engage in singing contests (“countersinging”) with other males. In addition, a separate non-spatial cognitive ability correlated positively with the attractiveness of the songs they sung. In sum, these results suggest the cognitive skills assessed in the lab were strongly connected to breeding behavior and reproductive success. Moreover, since certain cognitive abilities related to different aspects of breeding success, it suggests that cognitive modules may have specialized adaptive value, but also that these specialized skills may interact and influence fitness in surprising ways.

Discussion

Despite the small sample size, the different cognitive scores correlated with several aspects of breeding behavior and reproductive success in both males and females.

Females

Those females who reliably scored highest on spatial tests in the lab received the most courtship song from males. The number of songs sung to females is an important variable associated with pairbond strength and copulation success, and thus is integral to reproduction (White et al., 2010a, b, c). It is unclear what drives this correlation. It is possible that there is something about these females that makes them more attractive to males. Results of past work, however, would suggest that something about the females’ behavior is important in stimulating the males to sing to them more often (Maguire et al., 2013). What females do to get males to sing to them more often is unknown – though one possibility is the use of chatter in response to males’ song (Maguire et al., 2013).

The other variable significantly associated with cognition for females was egg production. Females are often highly variable in egg production between and across groups and past work has been only marginally successful in explaining this variation. Most of those explanations have revolved around the idea that females invest more in egg production in circumstances where they have the most valuable information about the quality of males present (White et al., 2010a, b, c). The cognition score used here is by far the strongest explanatory variable we have ever found for egg production. Perhaps females who have better cognitive abilities can best engage in the behaviors associated with selecting the highest quality, or most compatible mate, building the most successful pairbond, and therefore most likely to invest in egg production. An interesting aspect of this relationship is that laying more eggs leads to a higher spatial memory demand because it requires finding more nests. No matter what mechanism explains this relationship, the connection between spatial cognition and reproductive success suggests that sexual selection can be a driving force on spatial cognition in females.

Males

We had the opportunity to examine how two measures of cognitive performance related to males’ fitness. First, we found that song attractiveness, as measured in playback tests related to the males’ performance on cognitive tasks that used color stimuli in delayed match-to-sample tests. That song attractiveness related to cognitive performance supports the theory that those males best able to learn are the ones who can produce the most attractive signal – a theory of the functional value of song that has been posited for songbirds in general (Nowicki et al., 1998, 2002) and cowbirds specifically (West & King, 1988). These results connecting cognitive performance and song differ from work in song sparrows (Soha et al., 2019), where no connections between cognition and song could be detected (see also Templeton et al., 2014). Why song attractiveness should relate to cognition for color per se, is unclear. Perhaps song attractiveness and performance on the color tasks are linked by another unmeasured variable relating to male quality (health, “good genes”, or stress responsiveness). This would appear unlikely since past work has shown that song attractiveness is highly dependent on developmental (West & King, 1988) and immediate (Gersick & White, 2018) social experiences. Thus, the most likely route leading to variation in song attractiveness involves interacting and learning from the visual responses of females and other males to singing overtures. The color tasks were designed as a control for spatial memory performance and not designed specifically to examine an aspect of cognition hypothesized to be important to male breeding behavior. Thus, the color tasks may be measuring some more general aspect of visual acuity, attention, or learning. More work is needed to determine exactly what cognitive mechanism is driving color discrimination and song development. It is clear, however, that the cognitive ability measured using the color task was distinct from spatial memory skill because performance on spatial tasks did not relate to song attractiveness.

Spatial task performance did, however, relate to one important aspect of singing in males: countersinging. Countersinging is a skill that males must learn in order to attain and maintain dominance among males and to stimulate the reproductive output of females (White et al., 2010a, b, c). Past work has shown that countersinging is learned by juveniles over their first year of life as they approach and sing with adult males (White, King, & West, 2002a). This ability to get close to other males, sing with them in duetting bouts, and temper aggression leads to a cascade of learning other breeding skills and is highly variable among males (White et al., 2007a, b; White, King, & West, 2002a).

No other variables for males or females reached the large effect size necessary for statistical significance (other than chatter patterns in females). There is, however, a distinction that should be made between the correlation strength needed for statistical confidence and for biological relevance. Evolution can act on very small effects. Tables 1 and 2 show some of the effects that did not reach significance but will be the subject of future work, as many of them may inform us of the potential directions of effect occurring with other variables. For example, female chatter is highly stimulating and motivating to males (Burnell & Rothstein, 1994; Freed-Brown & White, 2009; Hauber et al., 2001; Lynch et al., 2017; Snyder-Mackler & White, 2011). Perhaps the production and use of chatter is a behavioral mechanism that females use to regulate males’ behavior, stimulate courtship effort, and strengthen the pairbond (Maguire et al., 2013), leading eventually to more egg output. Other interesting positive relationships with females’ spatial cognition include the song attractiveness of their pairmate, and the amount of courtship song they receive only from their pairmate (a measure of pairbond strength we have found in the past to be important for breeding success; Maguire et al., 2013).

The disclaimers here are most likely obvious: the frustratingly low sample size highlights the challenges for neuroecology and studies of animal cognition in general where the depth of understanding of individuals’ cognitive abilities trades off against testing large numbers of subjects and therefore against generalizability and statistical power. The low number of subjects precluded more detailed statistical analyses, and we could only rely on a small number of a priori comparisons requiring very strong relationships to reject a null hypothesis. Also, the birds in this study, while wild caught, experienced years of life in abnormal contexts, raising questions about generalizability to the wild (although in the aviaries they bred in patterns very similar to the resident birds). Finally, it was not the primary goal of the cognitive experiments to subsequently study fitness. Had it been, we would have ensured that we collected measures that were more directly comparable across subjects. As is, it is not clear what cognitive modules we are examining here. The color-discrimination tasks might be measuring visual acuity, attention, learning speed, etc. Spatial cognition here also includes tasks that were focused on numerical discrimination. Thus, this work should be considered an exploratory first step that, even with the limitation inherent in these data, was surprisingly successful in demonstrating relationships between different aspects of cognition and reproductive success. This discovery will drive experiments both in the lab and in aviaries for years to come.

What do these findings mean for the adaptive specialization hypothesis about cowbird spatial memory? We still have not been able to test directly whether spatial cognition abilities allow females to successfully find and select viable nests in the wild – the critical relationship posited by the adaptive specialization hypothesis that started the work with cowbirds. With modern advances in automated tracking technology and advances in neural manipulations, this relationship may be testable in the near future. The findings reported here – that different measures of cognition related to different aspects of effective breeding – support the idea that there are functionally distinct cognitive systems as proposed by Sherry and Schacter (1987). There do seem to be different cognitive domains at work here, similar to food-caching species that show different patterns of performance depending on whether a task is spatially based or color based (Olson et al., 1995). Females’ superiority in behavioral tasks and the hippocampal size evidence suggest that the potential exists for selection to act on spatial cognition through nest-finding abilities. The interconnections between these cognitive systems and diverse aspects of breeding revealed here, however, suggests some cooption, or exaptation of the cognitive system, which significantly complicates determining how selection has acted and may act (Gould & Vrba, 1982; Sherry & Schacter, 1987). Selection may be operating on spatial memory skills for both a specialized demand on the species (finding nests), and also a non-specialized demand (selecting a mate and reproducing). This suggests there are non-additive interactions among cognitive modules and fitness.

The story of how cognition and fitness relate may not be simple, but simple stories and complex systems rarely go together. The complexity of living systems presents many different routes and strategies leading to reproductive success and thus identifying how distinct memory systems relate to fitness can be challenging. Studying the wealth of links between memory systems, however – how they can work independently and together, how they react to different environments, to past experiences and to conspecifics – and ultimately lead to organizing adaptive behavior holds the promise to fully understand the evolution of the brain and intelligence.