CORDIS - EU research results

The Evolution of Cognitive Performance

Final Report Summary - EVOCOGN (The Evolution of Cognitive Performance)

The evolution of cognition is one of the most important, yet poorly understood issues in modern biology. The field of comparative cognition has generated a wealth of research into the cognitive processes underpinning animal behaviour, yet the evolutionary forces that have driven and shaped these processes are poorly understood. Understanding cognitive evolution requires examination of the causes and fitness consequences of variation among individuals in their cognitive performance. This requires a system in which large numbers of individuals can be assayed for performance in a battery of psychometric tasks. It is necessary to quantify and control for factors that affect the heritability of cognitive performance. Crucially, it is necessary to determine how natural selection in the wild acts on cognitive ability, and so determine how cognition evolves. Pheasants Phasianus colchicus offer a unique system in which large numbers of individuals can be reared under controlled conditions and then released into the wild where they face pressures of natural selection.

We have developed and deployed a battery of cognitive tasks that assay individual performance in: colour and spatial discrimination, reversal (of colour and space cues) learning, set learning, spatial memory (at fine and coarse scales), inhibitory control (via detour reach tasks and A not B task variants), time-structured memory, social learning, motor performance, and classical problem solving paradigms. We have demonstrated that some of these tasks do not provide much information about cognitive performances, but rather indicate aspects of motivation and chance. We have demonstrated that an individual’s propensity to participate in such tasks is non-random, but depends on aspects of personality with potential for bias in cognitive sampling. For tests that are robust, we demonstrate that individuals differ in their performance in these tasks and that such individual differences correspond to differences in fitness (survival) on release. Although we have not yet found any support for ‘g’, a single measure of general intelligence, we have found correspondence in individual performance on tasks based on similar cognitive domains, and we can summarise individuals as being generally good or poor at associative learning processes. Surprisingly, for at least some of the performances, individuals who performed poorly were more likely to survive. This provides an explanation for why cognitive performance does not runaway, but is checked by some fitness aspects. For other measures, we detect indications of positive fitness correlations with performance, but these need further investigation and confirmation. We have an example of how early life rearing environment (specifically habitat complexity) may lead to differences in cognitive performance in spatial memory tasks, with individuals reared in complex habitats having better spatial memory, and we found that these individuals were less likely to be found dead post-release.

Our empirical work has led us to question current methods of assaying individual differences in cognitive performance, and to suggest robust alternatives. Although we have developed a paradigm to run such tests with large numbers of captive young birds, we are continuing to develop variants suitable for testing of free-living adult birds. We still need to quantify the relative contributions of genetic, early life and adult factors to the expression of cognitive performance. Having found that gross fitness outcomes may relate to individual differences in cognitive performance, we are now drilling down into some of the mechanisms by which these fitness outcomes are determined.

Our outputs have attracted much interest from other researchers in both psychology and biology and is contributing to the development of the growing field of cognitive evolution. The work has also attracted attention from the UK shooting industry who wish to understand how altering early life rearing conditions may influence the survival and behaviour of the ~40 million pheasants released annually for shooting.