Chapter 5: Learning and Cognition
5.1 Learning Allows Animals To Adapt To Their Environment
1. Learning: a relatively permanent change in behaviour as a result of experience .
LEARNING AS AN ADAPTIVE TRAIT
2. One of the simplest approaches to answer this question is to compare the
behaviour of individuals that differ in age and experience.
1. If learning improves fitness, then older more experienced individual should
outperform those that are less experienced.
3. Sullivan tested this hypothesis in a simple experiment with yellow eyes juncos,
small sparrow s that live in the mountain of southwestern North America.
1. Previous worked allowed Sullivan to age birds into four categories, recently
fledged juveniles (four to seven weeks after leaving the nest, or fledging),
young juveniles (eight to ten weeks after fledging), older juveniles (11 to 14
weeks after fledging), and adults (greater than 1 year old).
2. Juncos manoeuvre food in their bill for several seconds to prepare it for
consumption. Sullivan predicted that as birds age, they gain more
experience handling food items and so become more efficient foragers.
3. She tested this prediction by providing good that required different levels of
manipulations mealworms were cut into small pieces (0.009g dry weight_ that
were easy to handle or large pieces (0.026 g dry weight) that were more
difficult to handle. Birds had access to one food type for several days to be
sure they had some exposure to each food type.
4. Each day over the following two weeks, Sullivan recorded the handling time
of each food item (the time from first contact until the item was consumed) for
each bird. She found that for both prey types, adults had the lowest handling
time while recently fledged birds had the highest.
1. She also calculated energy gain for the foraging birds. The difference in
handling times led to differences in energy intake rate: it was highest for
adults and lowest for recently fledged birds.
1. Sullivan attributed these patterns to differences in experience. Older
birds with more experience had learned how to best handle insect prey
and this became more proficient at the task.
EVOLUTION OF LEARNING
4. Theory indicates that two factors affect the evolution of learning: environmental
stability and the usefulness of past experience.
5. Two possible types of worlds: one is fixed and nothing ever changes (i.e. no new
predators or parasites ever appear), while the other is dynamic and changes
unpredictably (e.g. new predators and parasites may appear in the middle of
the breeding season)
1. Now imagine two habitats (A and B) that differ in quality of reproduction:
breeding in one leads to high fitness, while breeding in the other leads to
reproductive failure. 2. In the fixed world, the highfitness habitat (A) will have always be the best
place to reproduce. In the dynamic world, the highfitness habitat can change.
Sometimes A will be best and at other times B will be best, thanks to the
unpredictable nature of predators and parasites.
1. Let’s assume that in this changing world, each habitat has a 50% chance
each breeding season of being high quality or causing reproductive
3. In both worlds, learning would not evolve. In the fixed world, individuals that
breed in the highfitness habitat would quickly outcompete those that
reproduce in the other habitat.
1. If habitat choice is influenced by genes, the world would soon be full of
individuals that selected only the highfitness habitat, and there would be
no need to learn about the other habitat or the differences between them.
2. On the other hand, the dynamic world, there is nothing to learn, because
learning is useful only if the individual can benefit from their experience.
There must be a predictable relationship between experience and the best
option now available.
1. In this world, the habitats change unpredictably, and so both habitats
always have a 50% chance of being the best option, no matter what
happened in the past. Again there is not benefit of learning.
4. These factors are independent of each other but are not mutually limited. AS
environmental regularity increase, learning will become less favoured
because in a completely regular world, evolution will fix behaviour. As the
reliability of experience increases, learning will be strongly favoured because
individuals that learn will have higher fitness than those that do not.
GREEN FROG HABITUATION TO INTRUDER VOCALIZATIONS
6. Habituation: the reduction and then lack of response to a stimulus over time.
7. An environmental stimulus is anything in the environment (biotic and abiotic) that
an individual can perceive, and any reaction to the stimulus is a response.
8. Deer enemy hypothesis: territory owners will show reduced aggressive
interactions toward territorial neighbours compared to strangers.
1. The mechanism for this reduction is aggression is habituation. Strangers
represent potential threats as new competitors for territorials space, while a
neighbours represents a much lower threat, because it has already established
a nearby territory. A territory owner can save time and effort by essentially
ignoring a neighbour while aggressively interacting with strangers.
9. Owen and Perrill investigated whether the dear enemy hypothesis explains
aggression in territorial green frogs. Males defend territories around ponds,
where they vocalize to attract females. Territorial males respond to strangers with
both aggressive vocalize to attract females. Territorial males respond to strangers
with both aggressive vocalizations and physical attacks. 1. Aggressive vocalizations can be distinguished from advertisement calls used
to attract females by their lower dominant frequency. They predicted that if
males habituate to familiar stimuli, there should be a decrease in response
to new rival’s vocalization after an initial aggressive response.
10. The researchers studied males at four ponds. They simulated intruders by
synthesizing the calls of two males and played them from a speaker placed 1 to
2m away from a focal calling male. Synthesizing the calls standardized their
length and intensity. One intruder’s dominant frequency was set at 350 Hz, while
the second intruder’s dominant frequency was set at 450 Hz.
1. For each focal calling male, the researchers first recorded advertisement.
Next, they played one of the synthetic intruder calls at 515 second interval
for up to an hour. During playback, the researchers recorded the focal frog’s
movement towards the speaker and it’s aggressive vocalization.
2. When the focal frogs stopped moving and began to produce advertisement
calls again the researchers stopped the playback and started 15minute rest
period, after which they initiated the same intruder call once more.
3. The researchers again recorded the focal male’s movement toward the speaker
and aggressive calls. When movement ceased and advertisement calls were
being produced, the researchers played the second synthetic call to stimulate
a new intruder and recorded the vocalizations of the focal frog.
4. Habituation to the intruder calls was observed in both the focal frog’s
movement and vocalizations. At first, focal frogs made several movements
toward the speaker increased following the rest period and reinitiation of the
synthetic call, but at a lower level. Calling frogs produced advertisement calls
prior to the playback of the synthetic intruder but switched to aggressive calls
when the synthetic call began.
1. Eventually, focal frogs switched back to advertisement calls until a new
synthetic call was played.
5. These results suppose the deer enemy hypothesis and demonstrate habituation
to a stranger: both physical and vocal responses to the simulated intruder
were initially high but over time, both responses declined as the intruder
became more familiar and was perhaps perceived to be less of a threat.
5.2 Learning Is Associated With Neurological Changes
11. Information from experiences and environmental stimuli are perceived via
sensory receptors and relayed to the central nervous system through nerves.
Nerves are composed of neurons, cells that receive and transfer electrical and
12. The junction between two neurons, the synapse, is believed to play an important
role in learning and memory. Two aspects of the nervous system is believed to
play an important role in learning and memory: both changes in neurotransmitter
and the number of synapses between the neurons are associated with learning. 13. A typical approach to understanding such proximate mechanisms of learning is to
characterize synapse characteristics before and after a learned experience to
determine what changes occur.
NEUROTRANSMITTERS AND LEARNING IN CHICKS
14. Imprinting: rapid learning of a phenotype in young animals.
15. Filial Imprinting: the learning of the phenotype and identity of parents by
16. Lorenz made imprinting famous when he showed how graylag geese hatchling
imprint on his boots when these were the first objects the saw. In the absence of
their parents, they simply followed him around. More recently, this behaviour has
been used for the reintroduction of endangered birds their former habitats.
17. Many birds, like domestic chickens, visually imprint on a stimulus when they
hatch. In their brain, the intermediate and medial parts of the hyperstriatum
ventral (IMHV) appear to play an important role in memory for imprinting;
earlier work demonstrated that lesions in this area of the brain prevent
imprinting, as does the blocking of postsynaptic neurotransmitter receptors.
18. Meredith and colleagues studied another imprinting mechanism in chicks by
investigating whether the releases of neurotransmitters from the presynaptic
neuron is also associated with such learning.
1. The researchers divided chicks into two groups. Half were trained by
exposure to a visual imprinting stimulus (a red box or a blue cylinder). The
other birds were used as a control group: they were not trained and so had no
visual stimulus for imprinting.
2. During training, birds were placed on a running wheel; the birds typically
attempted to move toward the stimulus, and the wheel recorded these
movements. Less than ten minutes after training, the researchers measured the
strength of imprinting, or preference score, by sequentially placing two
objects in front of the chick while it was on the running wheel. One was the
imprinted object, while the other was a novel object.
3. The preference score was calculated by dividing the amount of running
toward the imprinted object in a fixed time period by the total amount of
running in response to both objects.
1. A score of 50 showed no preference for the imprinted object, while higher
scores showed greater preference and a higher strength of imprinting. At
the end of the experiment, chicks were sacrificed, the IMHV tissue
dissected, an assays conducted to measure the release of several amino
acid neurotransmitters, including glutamate and gammaaminobutyric
4. The research team found higher glutamate in the IMHV tissue of trained
chicks compared to that of the controls but no difference in GABA. However,
among the trained birds, chicks with the highest preference scorethose that
most strongly imprinted on the test objecthad higher levels of GABA in their
brain. 1. These results suggest that neurotransmitters play a role in imprinting. One
neurotransmitter glutamate, was more strongly released in the brain of
birds that visually imprinted on an object, while another, GABA, was
correlated with the strength of imprinting. Illustrates that
neurotransmitter release in presynaptic neurons played a role in
imprinting over the short timescale of the experiment.
DENDRITIC SPINES AND LEARNING IN MICE
19. Memory: retention of learned experience and is a critical factor in animals’
ability to utilize their prior experiences.
20. Neural plasticity: structural changes in the brain, especially in the number of
synapses and the strength of chemical synapses between neurons.
21. Dendritic spines: small protuberances on a dendrite that typically receive
22. Yang, Pan and Gan examined spine formation associated with learning in mice.
One group of mice was trained to learn a new motor skill: running on a rotating
rod suspended above the cage floor.
1. The mice had to learn novel motor coordination skills and balance in order to
stay on the rod at its speed increased.
2. Another group of mice received no training and served as control. The
researchers used transcranial twophoton microscopy to examine the
fluorescentlabeled dendritic spines of living subjects.
1. This technique involves surgically thinning the cranium to allow for high
resolution imaging of dendritic spines.
3. After one to two days of training, both young and old mice that learned the
new motor skill showed significantly higher levels of dendritic spine
formation than controls.
1. The performance of mice on the rotorod (the revolution per minute they
could achieve) was associated with the number of new spines formed:
mice that had developed more new dendritic spines performed the task
1. Over the course of the following 2 weeks, most of the new spines
disappeared, so that the net result was essentially no difference in the
total number of spines between trained and control mice. In fact, the
researchers estimated that less that 0.1% of the new spines would
persist for life.
23. One possible explanation for how this related to lifelong memory is that is may be
associated with both the formation of new spines during learning and the pruning
if synapses that, in essences, remodel the brain during learning. AVIAN STUDY OF STORED FOOD
24. Cache: food stored in a hidden location for later retrieval.
1. Caching provides for the future but is only useful if that food can be relocated.
25. Episodic memory: memory of a specific object, place and time.
26. Caching is particularly common in two families of birds, Corvidae (crows ad
jays) and Paridae (chickadees and titmice), but catching tendency varies across
species. Some, such as Clark’s nutcrackers and willow tits, cache heavily and rely
on stored food for survival. Others, such as scrub jays and marsh tits cache
occasionally, while species such as jackdaws and blue tits rarely, if ever cache
1. If cache recovery is accomplished through memory, the species that chance
heavily should rely more heavily on a welldeveloped memory.
2. In many vertebrates, spatial memory is associated with the hippocampus,
which is located in the medial temporal lobe. Early work found that across
families, the hippocampus is larger in species of birds that often cache food.
However, species in different families may differ in brain structure for a
variety of reasons.
3. Lucas and colleagues examined how species within the same family vary in
foodhoarding propensity and brain structure. They gathered data from
studies on 13 corvid and ten parid species. The degree of food caching ranged
from noncaching species to those that rely on cached food for survival.
1. They found a strong association across species within both families: those
that relied heavily on food caches indeed had a larger hippocampus size
and function (memory of cache locations) in these species.
5.3 Animals Learn StimulusResponse Associations
27. One common type of learning involves making an association between an
environmental stimulus and a behavioural response a stimulusresponse
1. Many animals make such associations allowing them to learn about their
environment, prepare for any similar future events, and respond accordingly.
28. Innate: (instinct) behaviours that are performed the same way each time, are
fully expressed the first time they are exhibited, and are present even in
individuals raised in isolations.
29. Classical conditioning: (Pavolian conditioning) A type of learning in which a
novel neutral stimulus is paired with an existing stimulus that elicits a particular
PAVOLIAN CONDITIONING FOR MATING OPPORTUNITIES IN
30. Classical conditioning allows individuals to become better prepared for future
events by learning new association. 31. Reagan and MacKillop studied how classical conditioning affects mating
behaviour in Japanese quail.
1. They conditioned adult males and females in the laboratory to two different
mating situations using cages that differed in size, location, and appearance
(wire or Plexiglass construction).
2. Each bird learned that mating always occurred in one cage condition
(conditioned stimulus, or CS+) but not the other (CS). For half the birds, a
mate was added after two minute only when they were placed in the
Plexiglass cage, while for the other half, a mate was added only when they
were in the wire cage.
1. For the focal males, once the female was added, mating ensued. For
females, no mating was allowed during conditioning, because females can
store sperm, which could have biased the results. The male was thus kept
behind a wire mesh screen. Training occurred twice a day for five days.
3. After training, the focal individuals were tested in both cages: researchers
presented these individuals with a mate and then counted the number of
inseminations that occurred. Eggs were collected, incubated, and allowed to
develop for one week before being examined for embryos. Males fertilized
more eggs in the cage, where they had been conditioned to expect a male (the
4. Females also had more fertilized in the CS+ condition. Both sexes thus
achieved greater reproductive success in the condition where they had learned
that mating opportunities occur.
1. These results show that Pavlovian conditioning can affect fitness. The
researchers hypothesized that males may transfer larger or more effective
ejaculates and that females may behave differently to allow more sperm
transfer under the conditioned stimulus.
FISH LEARN NOVEL PREDATORS
32. A prey can innately identify its predators. However, many animals move large
distances when migrating or searching for appropriate habitats and so ma
encounter many different predators.
33. Mitchell and colleagues examined the ability of juvenile lemon damselfish to
learn about novel predators. Lemon damselfish larvae develop in open water for
about a month, after which they settle on the Great Barrier Reed off the coast of
Australia. Settlers are small (1cm in length) and suffer high predation.
34. Chemical alarm substance: chemicals released from damaged epidermal cells in
fish that function as an alarm cure for others.
35. Fish that perceive chemical alarm substances respond with innate antipredator
behaviours: they reduce feeding, increase vigilance behaviour (scanning the
environment for predators), and may spend more time in shelter.
1. Researchers investigated whether fish learn their predators by association
with the presence of the chemical alarm.
36. The experimenters created a “cocktail: of chemical odours using adults of four
fish species caught from the reef: two fish predators and two nonpredators.
Individuals of each species were placed in separated tanks and left for six hours enough time for the tank water to accumulate species=specific odours. The
researchers mixed water for each of their separate tanks into a “cocktail” of
37. Mitchell’s team collect damselfish recruits that had not yet settled using light
traps set to 50 to 100m from the reef. Since predators of these fish are only found
on reefs and not in open waters, the damselfish recruits were assumed to be naïve
in regard to these predators.
1. On day 1, half the damselfish were conditioned to the cocktail odours plus a
conspecific alarm substance. The other half was conditioned to the cocktail
odours to odours from one of the cocktail species of to the combined odours of
two novel species, only one of which was a fish predator. The researchers
recorded the amount of time the test fish spent feeding, their distance from the
shelter, and time spent in shelter.
2. When the odour of each species was added, damselfish conditioned with the
cocktail of species and alarm substance reduced their feeding time
significantly more than did the controls.
3. The test fish had apparently learned a new association between the odour of
each cocktail fish and predation risk, as indicated by the chemical alarm
substance. Individual tested with the odour of the cocktail fish paired with
saltwater did not respond with antipredator behaviours. Individuals also did
not respond to the odours of the two novel fish species.
1. Interestingly, there was no difference in shelter use between the two
groups, perhaps because there were no predators in sight.
4. The results demonstrate that lemon damselfish can rapidly learn associations
between fish odours and the risk from unfamiliar predators. In addition,
because damselfish did not respond to the odour of a novel predator, there
was no evidence that they have innate knowledge of predatory species;
instead, they must learn to identify predators.
38. Operant conditioning:(instrumental conditioning) A learning process in which
an animal learns to associate behaviour with a particular consequence.
39. Any of four types of operant conditioning can occur, according to whether a
stimulus is added or removed:
1. In positive reinforcement, behaviour increases because it is associated with
the addition of a desired stimulus, such as food.
2. In negative reinforcement, behaviour increased because it is associated with
the removal of an aversive stimulus, such as pain.
3. In positive punishment, behaviour declines because it is associated with an
aversive stimulus, such as pain.
4. In negative punishment, behaviour declines because a desired stimulus is
removed, such as food reward. 40. Operant chamber: an enclosure used to study behavioural conditioning.
41. Trial and error learning: learning through repetition that results in rewards or
progress toward a goal.
42. Learning curve: a graphical representation of a change in learning over time.
LEARNING CURVES IN MACAQUES
43. Murray, Kralik, and Wise studied the strength of this preference in rhesus
macaques. They offered six subjects a choice between one and four peanut halves
that were placed in the experimenter’s open hand.
1. If a test subject reached for the hand with one food item, it received four;
when a macaque reached for the hand with four food items, it received one.
Here, one action provided four times more rewards than the other and
therefore represented a more positive reinforcing outcome.
1. Twenty trials were conducted each day with a 20 second delay between
2. As expected, each subject showed a strong initial preference for the hand with
four food items (the incorrect choice). Macaques did learn to select the hand
with one food item in order to receive four, demonstrating trialanderror
learning, but there was tremendous variation in their learning curves.
1. One macaque learned rather quickly, after about 340 trials, but another
macaque took over 2.700 trials to attain a low error rate.
TRIALERRORLEARNING IN BEES
44. Raine and Chittka, using operant condition, examined the learning curves of bees
trained to associate the colour yellow with a food reward in the laboratory
45. The researchers worked with 12 colonies that contained uniquely marked
workers. All bees were first allowed to feed from artificial flowers that contained
sugar water. Each flower was multicolouredboth blue and yellow.