Lecture 1 and 2

Question 1

What did Darwin give to behavioural ecology?

Answer 1

. By asserting continuity of animals and humans he opened the way for disciplines such comparative psychology
. Invoked sexual selection to explain sexual dimorphism. Made the distinction between intra and inter sexual selection
. He opened a way for comparative where we can compare other animals to human beings

Question 2

What year did Darwin release the origin of species?

Answer 2

1859

Question 3

What is the difference in the way classical ethology and behavioural ecology come up with hypotheses?

Answer 3

. In classic ethology they usually set up experiments on animals and observe the animals and come up with hypotheses from this
. In contrast in behaviour ethology usually come up with hypotheses to explain behaviours before actually sitting and observing the animals for a long time and they are often generated by models, so a mathematician will write a (computer) model of a system and often behavioural ecologists test the predictions of that

Question 4

How do classic ethology and behavioural ecology differ?

Answer 4

. The way they come up with hypotheses
. In ethology you have detailed descriptions and qualification of behaviour fairly similar in that respect. But tend to deal with performance or fitness measures in behavioural ecology and there is quite a big use of molecular tools whereas there isn’t in ethology

Question 5

What is adaption?

Answer 5

The idea that all behaviours have evolved to be optimal (not that bothered about the mechanisms that underline them)

Question 6

What do they assume in classic ethology?

Answer 6

Tends to assume in adaption that things do evolve to be optimal but there is some kind of structural constraint as to what can evolve

Question 7

What do they tend to ignore in behavioural ecology?

Answer 7

The structure of organs and things like that, structure of the brain and just assume things will become optimal

Question 8

What does behavioural ecology think about adaption/ selection?

Answer 8

That it is at the level of the individual

Question 9

What does classic ethology have greater emphasis on than behavioural ecology? How does this differ?

Answer 9

Classic ethology has a greater emphasis on the structure of the brain whereas behavioural ecology doesn’t and also behavioural tends to ignore genetics and just assumes there will be enough genetic variation in a population for a behaviour to occur.

Question 10

What do both classic ethology and behavioural ecology not take into account?

Answer 10

Don’t really take genetics into account

Question 11

What does behavioural ecology focus on? How does this differ to ethology?

Answer 11

Behavioural ecology focuses mainly on the adaptive value of a behavioural trait.
But ethology studies the mechanisms so they might study the neural circuits in the brain that cause the behaviour but behavioural ecologists tend not to.

Question 12

What is the period after birth that birds will attach to whatever is there?

Answer 12

3-10 hours after birth (humans were found to have a sensitive period as well but not as much/ not to the same extent

Question 13

What is game theory?

Answer 13

It basically tells you what the best strategy for an individual is if interacting with one or more individuals and the pay off for doing a certain behaviour is dependant on what the other individual is doing. So, it comes up with the optimum strategy depending on what others are doing, it is a mathematical computational technique

Question 14

Introduction of evolutionary biologists to game theory predominantly developed what?

Answer 14

Economics

Question 15

What is Evolutionary stable strategy (ESS)?

Answer 15

Is a strategy that, if adopted by all members of a population, cannot be bettered by an alternative strategy (so the strategy that is most common in your population is the best strategy)

Question 16

How do computer scientists model evolutionary stable strategy?

Answer 16

What computer scientists tend to do is to have a pool of individuals interacting with each other and every now and then they throw in a new strategy and they interact in a certain way and sometimes in early simulation these strategies become dominant in the population because they are fit but as these simulations go on in a population with many generations if you put in a new strategy they tend just to be outcompete so eventually you get to a stage where no new strategy in your simulation can outcompete the others so that is when you get to an evolutionary stable state.

Question 17

What is game theory designed to handle?

Answer 17

Is designed to handle questions where the pay-offs of an action depend not only on a player’s own choices but also of those of others (so what it is saying is that the pay-off depends on not only the strategy you use but also the strategy of your opponent/ others use)

Question 18

The prisoners dilemma game is not strictly speaking an animal-animal interaction but a ‘social dilemma’ where cooperative or aggressive strategies may be adopted. How is this relevant to behavioural ecology?

Answer 18

It is assumed to be relevant to behavioural ecology and animal-animal interactions

Question 19

How is the prisoners dilemma game useful to behavioural ecology?

Answer 19

Mathematicians and computer scientists need ‘standard’ problems like this that lots of different people can work on, allowing them to reach general conclusions (can all work on a particular problem together looking at different areas of it and takes information from these different fields)

Question 20

What does the prisoners dilemma game allow scientists to determine and how?

Answer 20

By manipulating different parts of the game scientists can determine what factors favour cooperative strategies and which promote aggressive ones. Evolution of cooperation is a major theme in behavioural ecology

Question 21

From the prisons dilemma game what is the best strategy?

Answer 21

In its simplest form the aggressive strategy of ‘always defect’ is the best strategy

Question 22

What does the prison dilemma game assume?

Answer 22

That the players have perfect memory (which obvs they don’t) as the game is usually played across numerous ‘turns’, with scores summed across turns, where players can remember what their opponent did during previous turns which is where memory comes into it

Question 23

Describe the hawk-dove game

Answer 23

Two animals engage in a single contest over a resource that makes a contribution to an individuals fitness. Each animal can adopt one of two actions: ‘hawks’ behave aggressively until injured or until the contest is won, and ‘doves’ display and retreat if the opponent behaves aggressively

Question 24

In the hawk-dove game what happens if two doves meet?

Answer 24

They are both passive so they split the resource in two so they don’t fight over it

Question 25

In the hawk-dove game what happens if a dove and hawk meet?

Answer 25

The dove gets nothing as the hawk is aggressive and the dove moves away and the hawk gets the resource

Question 26

What is the distinct difference between the prisoners dilemma game and the hawk-dove game?

Answer 26

If you add spatial structure to the prisoners dilemma game you get cooperative strategies to be fittest, if you do that in the hawk-dove game the cooperative strategies don’t tend to be fittest

Question 27

What is a criticism of game theory?

Answer 27

It assumes ‘rationality’ (players are assumed to be rational

Question 28

Economically rational preferences in game theory are generally assumed to obey a series of conditions. What are these?

Answer 28

. Transitivity

. Independence from irrelevant alternatives (I.I.A)

Question 29

What is transitivity?

Answer 29

In its simplest form it states that preferences are hierarchical, so if option a is preferred to option b and option b is preferred to option c, then a will be preferred to c

Question 30

What is independence from irrelevant alternatives?

Answer 30

The basic idea is that relative preference (so if you have a resource and another resource that is different value and the relative preference of these doesn’t change if you add a distractor resource as you should still prefer the resource that is worth the most) for one option over another is unaffected by adding or removing options from the choice set

Question 31

Explain the the irrationality test of New York City cab drivers which violates I.I.A

Answer 31

. Cab drivers pay a fixed fee to rent their cab for 24h and then they keep everything they earn
. They have to decide how long to drive for each day
. A maximising strategy is to work long hours on busy days and quit early on slow days
. However, cabbies set target earning levels for each day
. So, they end up working long hours on slow days and shorter hours on good days because they have target levels that they want t earn each day, so this is irrational (So the human economic behaviour can be irrational which is a criticism for game theory because it does not incorporate things like that)

Question 32

What is phenotypic gambit?

Answer 32

Game theory just comes up with the optimal strategy and then we assume that, that will evolve through evolution. So, this assumes that there is enough genetic variation that has the correct genes in a population for that strategy to evolve so this is usually an assumption of this optimum strategy will evolve.
It is called phenotypic gambit because we don’t really know that if in real populations the genetic variation to allow that behaviour to evolve is present

Question 33

What is the perceptron?

Answer 33

A fundamental subunit of the neural network

Question 34

What are artificial neural networks?

Answer 34

Simply lots of perceptrons joined together and passing information between each other

Question 35

What is the confusion effect? Give an example of where it is seen

Answer 35

It is when you have a task and get a bunch of distracting information it makes the task more difficult.
It has been found that with lions chasing wildebeest their ability to target any one prey goes down as the number of wildebeest increase so they get more and more confused, so their ability to capture an individual prey goes down as the number of prey goes up (it is thought to be why wildebeest get into groups

Question 36

What do retina topic maps do? and how this is connected to the optic nerve

Answer 36

Humans have these things called ‘retina topic maps’ so what you get into your eye will be carried into a region of the brain and will be mapped in that region of the brain. So, we have these regions where the retina is essentially reproduced in the brain. With the optic nerve you have this bottleneck as well where the nerves leave the retina, so the number of cells there isn’t that big for the amount of information encoded in these retina

Question 37

How do you model the confusion effect?

Answer 37

Input various different images such as a shoal of fish and input that into the network and train the network to reproduce that exactly. So, when this is input in get it to pass through the network and the network to reproduce that in the map in the brain. Use groups of different sizes and pass each of these through and look to see how well these are produced in the map as the size of the group gets bigger. Would expect the image to be mapped better when there were smaller groups like the confusion effect

Question 38

How did they experiment humans attacking imaginary prey on a computer screen?

Answer 38

They did an experiment where it was a computer game that pointed out one individual and then took the arrow away and they had to strike it with a mouse cursor and repeated that with different numbers of tadpole. They found that the target accuracy in the humans and the neural network goes down as the number of prey goes up (then they found that it went up in both at the end but ignoring that bit it is the confusion effect)

Question 39

How did they experiment the neural network model output?

Answer 39

It has been shown experimentally that if you isolate a target from the side, if you have a herd of wildebeest and one becomes isolated that predators find that very easy to target. So, they input a group with a straggler on the side and found that if you target that individual then your target accuracy is much higher than if you target a central individual

Question 40

How did they experiment that neural networks prefer extreme body ornaments, just like many real animals?

Answer 40

It was a sexual selection sort of study, so they have a 3 layered network, they input the appearance into an animal female brain and then you have a male and that gets input into the retina. They train them to say yes to a male with a slightly longer tail and no to the other one. Then when they played around with these train networks had an innate preference for really long tails, it just spontaneously emerges. So in sexual selection there are various mechanisms that can promote the evolution of these odd traits, could be sensory biases within the nervous system have a contribution to that

Question 41

What are the advantages of using artificial neural networks to model evolution of behaviour in behavioural ecology?

Answer 41

. Evolution is constrained by organic structure of the model. No assumption of ‘rationality’ or ‘optimality’ of behaviour need to be made
. Neural networks are small enough models that a lot of model parameter space (these models might have various different parameters in them that you can vary) may be examined. You usually want to examine as much of that space as possible
. Extremely versatile (any input/ output behaviour in animals. They can be modified to recognise sequences). Most stimulus-behavioural output system can be modelled
. They can be easily modified to display more sophisticated behaviours such as sequence recognition (by adding feedback connections)
. There is a lot of ongoing research on them- a lot is known about them

Question 42

What are the disadvantages of using artificial neural networks to model evolution of behaviour in behavioural ecology?

Answer 42

. It is often difficult to understand why the networks are producing the behaviour they do. Neural networks are considered ‘black boxes’
. They are not especially analytically tractable to mathematicians
. They are an extreme abstraction of cognition and need to be compared to real animal behaviour

Question 43

What did Hamilton look at?

Answer 43

Why we help relations more than we help strangers

Question 44

There have been hypotheses that animals are ‘fitter’ if they are symmetrical. How was this studied and what did they find?

Answer 44

They cut some swallow tails and females didn’t pick them