The evolution of cognition

Fish from the Poecilidae family provide an ideal platform to undertake such a research project. The cognitive abilities of fish are complex and diverse, comparing favourably to other vertebrates. In particular, the Trinidadian guppy Poecilia reticulata, is now an established model species in behaviour and cognition research and offers unique opportunities to experimentally test fundamental questions in cognitive evolution.

To test cognition of these fish, our lab has developed a range of cognitive assays. These assays include tests of simple associative learning, reversal learning, numerical abilities, social learning, and many more. Additionally, by measuring performance across a variety of cognitive domains and by constructing ‘cognitive landscapes’, the link between particular cognitive traits can be uncovered.

Ongoing projects

Predation

Predation is a major evolutionary force shaping traits in animals. Individuals generally face a trade-off between predation avoidance and activities such as foraging or courtship. To optimize this balance, animals need to implement adequate behavioural anti-predator responses. These responses can only be implemented if an animal can acquire information about the predation risk and can act on this information correctly. Both information gathering and decision making require cognitive processing. Therefore, increased cognitive abilities should decrease the risk of predation, and likewise a reduced predation pressure, due for instance habitat declines in apex predators, can have fundamental effects on subsequent cognitive performance of the prey species, limiting their ability to respond to environmental changes. An example for the positive link between predation and cognition can be found in the evolution of human cognition. Apart from outsmarting predators to avoid being eaten, living in groups may have originally been an adaptation against predation threat, and the cognitive challenges of complex social interactions are thought to be a major driver of primate brain size and intelligence. Numerous of such indirect lines of evidence suggest that predation selects for cognitive abilities. However, a direct test of this hypothesis is still lacking.

To experimentally identify how increased predation impacts cognitive abilities, and the underlying mechanisms, we are currently establishing six artificial selection lines. Three of these selection lines have been selected for survival under predation by exposing them to a brief period of adult predation. The selective agent for this artificial selection was naturalistic predation. Specifically, a common natural guppy predator, the pike cichlid (Crenicichla alta). The other three lines are control lines. These control animals had a similar experience of high-predation (predator odour, alarm pheromones, bystanders to predatory events) without actually being at risk of predation. This is vital, as indirect effects of predator exposure mediated by stress or fear can impact behavioural, anatomical, and gene expression patterns across generations. Predation selection was repeated over three generations. The offspring of the third generation will then be tested for differences in cognition.

Fish from both lines will be tested in a variety of cognitive tests, including tests of association, numerosity, social learning, and many more. Once all the cognitive assays have been performed, a computational model will be used to create multidimensional cognitive landscapes. This will allow us to visualize how cognition changes under the influence of predation.

Hybridisation

Hybridisation, when distinct species mate and produce viable offspring, is traditionally regarded as a predominantly destructive force, leading to unfit individuals that struggle to survive and reproduce. Yet, as hybridisation is common in nature (10% of animal and 25% of plant species), biologists began to wonder if there may be benefits to hybridisation. Indeed, armed with new theory, methodology, and data, scientists increasingly find examples where hybridisation is advantageous and important in generating biological diversity. For example, hybridisation may lead to a stable new species with higher genetic diversity and capacity to colonize a new habitat. Hybridisation may also produce hybrids with a variety of novel characteristics unlike the parental traits (e.g. unique shapes or colours). Some of these novel traits can become beneficial and promote the rapid formation of new species, as in the radiations of Darwin’s finches or African cichlids. Even our own species likely benefited from a hybrid genomic ancestry. Despite the growing recognition of hybridisation as an important evolutionary event, studies on the outcomes of hybridisation have mostly focused on genetic or morphological characteristics. The impact of hybridisation on more complex traits, such as behaviour and cognition, remains unexplored.

The concept that cognitive traits can be formed by hybridisation processes is an exciting idea only recently proposed. Cognitive traits, e.g., learning and memory, are important for fitness and survival, for instance when finding food or escaping predators. Cognitive traits are also highly heritable. Similar to morphological traits, if hybrid cognitive performance is poor then hybrids may not survive and reproduce, meaning cognition can play a role in strengthening the reproductive barrier between the two species. However, hybridisation may generate novel combinations of cognitive traits. Such establishment of ‘new’ cognitive abilities would offer a window into potential mechanisms of cognitive evolution. To date, studies examining if and how hybridisation impacts learning and memory are scarce and have yielded contrasting results. For example, mules performed better than horses and donkeys in a visual discrimination and a spatial task, but wild chickadee hybrids had poorer performance in a food storing and a problem-solving task compared to the parental species, black-capped and Carolina chickadees. Hybrid performance may even be context dependent, as hybrid cichlids were less efficient than parentals when feeding on common foods but outperformed them when feeding on ‘novel’ food types. Together, these examples illustrate the potential for hybridisation to both positively and negatively affect cognition. However, the scarcity of research in this topic leaves many questions unanswered. Does hybridisation generally impact cognition in different animals? For a given hybrid cross, does hybridisation have similar or variable effects on different cognitive traits? And importantly, can hybridisation act as an engine of evolution by producing novel cognitive phenotypes outside the parental range?

To answer these questions, this project aims to examine the effects of hybridisation on cognitive abilities of Poecilid fish. We are especially interested in testing if and how hybridisation impacts a range of cognitive skills; and if hybridisation can promote variation in cognitive phenotypes.

To answer these questions, I have experimentally produced hybrid crossings and will assay hybrids and parental line individuals in a complete cognitive assay, including tests of simple associative learning, reversal learning, behavioural inhibition, and numerical abilities, among others.