Cognition in a changing world: exploring the evolutionary potential of cognitive abilities in the wild – EVOL-COG

The project was based on the analysis of empirical data collected from two avian study models that were ideally suited to our objectives, thanks to (i) prior deep knowledge on their physiology and behaviour, (ii) the ability to easily monitor both species at a single study site with access to breeding data, physiological samples, individual traits (such as personality) and environmental factors, (iii) whilst allowing for the straightforward implementation of both cognitive tests and experiments on a large number of individuals (> 600 individuals monitored per species per year), and (iv) the long-term monitoring (> 15 years) of the population, which provided essential information (age, growth conditions, pedigree) for the implementation of the project. The cognitive tests used in the project had already been successfully used within this population, and were complemented by new automated devices and new tests to expand the range of cognitive abilities explored (inhibition) established during the project.

The project utilised both data already collected in the years prior to its start (for the long-term objectives – 2010 to 2016) and data collected during the project (for the more specific objectives, in particular the experimental approaches). The project therefore relied on annual field campaigns (between 2020 and 2024) within the study population, coordinated by the LBBE, which involved field teams comprising (i) the project leader and a postdoctoral researcher (overall coordination and field visits to set up the work), (ii) field assistants (3 per year, for 2 to 3 months each) to manage the team’s day-to-day work on site, (iii) a dozen students (undergraduate and postgraduate), to carry out measurements and experiments and to monitor the population and the corresponding data. Five field campaigns were carried out during the project (see below regarding the year of the Covid-19 pandemic).

At the same time, at the IPHC, a research engineer was recruited (for one year) to develop (with the assistance of permanent on-site staff) the new automated system tested during the field campaigns, and the physiological analyses (markers of oxidative stress and telomere length) were carried out in the laboratory under the coordination of local partners and an engineer recruited by the Swiss partner.

Finally, at BioSP, the hierarchical Bayesian approach (SCR-AM), which combines spatially explicit capture-recapture models (spatial capture-recapture, SCR) and quantitative genetic models of the animal model (AM) type, was carried out by a postdoctoral researcher (recruited for two years) under the supervision of the local partners, to implement various models, compare them through simulation and prepare them for application to the data.

Objective 1. Quantifying inter- and intra-individual variations in cognitive abilities under contrasting environmental conditions

We aimed at quantifying inter- and intra-population variations in cognitive performance by using (1) a multi-population approach, measuring performance using the same cognitive test across around twenty populations of great tits in Europe, through a large-scale collaboration, (2) a long-term approach within our main study population of tits and flycatchers in Sweden, to investigate the role of environmental variations across space and time; in parallel, we also explored intra-individual variation in cognitive performance with age, and demonstrated for the first time in the wild senescence in these abilities with age, accompanied by selective disappearance, whose the mechanism remains to be understood.

Objective 2. Elucidating the causal behavioural and physiological mechanisms linking cognitive abilities to individual fitness

We have focused on reproductive success in order to understand how cognitive performance leads to greater success within our population. We first demonstrated in tits that the highest-performing individuals (in terms of innovation) were the most effective at provisionning their young during growth (higher feeding rates and greater biomass provided); through path analysis, we found that it is via this superior provisioning eficiency that males achieve greater success, whilst the situation is less clear for females. We also demonstrated experimentally in flycatchers that cognitive performance (learning) modulated the use of a manipulated information source for breeding site selection, showing that cognitive abilities are involved in the management and use of complex environmental information to optimise decision-making. Finally, we have also demonstrated both correlationally and experimentally that physical condition (growth conditions, level of oxidative stress) affects cognitive performance in the short term (via plasticity) and in the long term (via structural organisation) throughout the life in tits, and that this influence could explain an indirect link (on top of a direct link via provisioning) between cognitive performance and reproductive success.

Objective 3. Obtaining unbiased estimates of the heritability of cognitive abilities

We have developed a hierarchical Bayesian approach, called SCR-AM, which combines spatially explicit capture-recapture (SCR) models – which account for imperfect detection of individuals – with quantitative genetic models. The originality of the SCR-AM approach lies in jointly estimating the trait, the observation process and the genetic structure, rather than first extracting traits from observed positions. Several models have been developed and compared using different measures of the traits in question: continuous versus discrete. The models now need to be applied to the data.

In addition to continuing to analyse the data successfully collected during the project to address the initial objectives (including the spatial and temporal variation in cognitive performance as a function of environmental factors, a comprehensive exploration of learning performance in the flycatcher, relationships with other cognitive abilities – associative learning, inhibition – and the application of SCR-AM models to cognitive performance data), the data generated by the project open up the possibility of exploring new questions. In particular, whilst the project has focused on reproductive success, the relationships between cognitive performance and other traits and fitness components may be studied; links with survival and dispersal, more specifically, appear to be both promising and crucial for fully understanding the evolutionary potential of cognitive abilities in the wild. The exploration of these relationships will therefore form the basis of new projects and research proposals using existing data (thus requiring no new data collection, or only highly targeted data collection to investigate specific questions in greater depth).

The project also clearly demonstrated (both correlationally and experimentally) that environmental conditions – in particular the early-life conditions to which individuals are exposed during growth – influence the development and maintenance of long-term cognitive performance throughout their lives in a natural environment. These observations appear particularly significant in the context of current environmental changes linked to human activities, which often influence reproductive conditions in the wild and could have an indirect impact on long-term cognitive abilities and, consequently, on individuals’ ability to adapt to these changes. We therefore plan to continue exploring this topic using a biological model that allows us to investigate the impact of environmental variations on a small spatial and temporal scale in the wild (whilst also permitting experimental approaches), something that the main population of tits and flycatchers (nor the multi-population approach for the tit) did not allow us to do, as the environmental variations studied in this project are observed at large scales. We therefore plan to carry out these new studies on another passerine species, using another population that has been monitored over the long term in a study area where there are marked environmental variations within the population, and thus on a very small scale in terms of general environmental conditions, including the presence and type of chemical pollutants, whose influence on the development of cognitive abilities has been demonstrated in the laboratory and is beginning to be explored in the wild.

Cognitive abilities are thought to be at the heart of behavioural flexibility, driving the potential for individuals to quickly respond to environmental changes. Among these cognitive abilities, innovation, learning and inhibition show high variation between species, and comparative studies suggest that they drive species’ ability to optimally cope with environmental spatio-temporal variation. However, we currently know very little about whether these cognitive abilities show adaptive within-species variation and can evolve in response to selective pressures in changing environments. This information is nonetheless essential to understand how the evolutionary potential of cognitive abilities in the wild shapes the rapid adaptation of populations to past, but also current, human-induced, environmental changes. Like any trait, assessing the evolutionary potential of cognitive abilities in the wild requires understanding (i) the variation of these abilities among- and within-individuals, (ii) their links with individual fitness components and the causal mechanisms underlying these links and (iii) their heritability. Current knowledge gaps on all three points prevent a critical understanding of the importance of cognitive abilities in shaping adaptation in a changing world Using two complementary passerine model systems in the field, we will address these three knowledge gaps to provide a comprehensive understanding of sources and adaptive consequences of variation in innovation, learning and inhibition. We will implement powerful approaches, in particular experiments, state-ofthe-art mixed capture-recapture models and large-scale within- and among-population comparisons to significantly advance our understanding of the evolutionary potential of these main cognitive abilities in a changing world. The project is structured in three research objectives: Objective 1. Quantifying inter- and intra-individual variation in cognitive abilities in contrasted environmental conditions to provide insights into the adaptive value of this variation - We will first investigate temporal (within-population) and spatial (among-populations) fluctuations of cognitive abilities, and their link with environmental conditions, and second explore age-related individual variation in cognitive abilities (cognitive senescence), using repeated measures for the same individuals within our main study population. Objective 2. Unravelling the causal behavioural and physiological mechanisms linking cognitive abilities to fitness - We will first test whether cognitive abilities shape individuals’ efficiency in using information, exploiting habitat and provisioning their nestlings using a series of experiments, and second explore whether cognitive abilities and fitness may also be indirectly linked via a third latent variable, such as ageing and oxidative stress. Objective 3. Obtaining unbiased estimates of the heritability of innovation and learning - We will develop and implement models combining multi-state capture-recapture and quantitative genetics approaches to estimate unbiased heritability levels in innovation and learning by accounting for imperfect individual detection and its potential heterogeneity among individuals in the wild. To achieve these aims, we will combine inputs from the fields of cognition, behavioural and evolutionary ecology, ecophysiology and biostatistics, with the complementary expertise of the five partners involved. By the end of this project we will know in unprecedented details to which extent cognitive abilities vary among and within populations and within individuals, whether cognitive abilities can respond to selection, and what mechanisms link cognition to fitness. By moving us considerably beyond typical studies of cognition in the wild, this project will therefore strongly contribute to address the currently open question of the evolutionary potential of cognitive abilities in response to environmental variation.