Exposure learning Flashcards Preview

PSY2304 Biological Basis of Behaviour > Exposure learning > Flashcards

Flashcards in Exposure learning Deck (17)
Loading flashcards...


birds learn song during critical period - if missed, impoverished song produced

May not be a critical period in humans – but have some predisposed language
acquisition device



Bateson and Horn have done a lot of work on the neural mechanisms underlying this phenomenon – changes at the synapse in the hippocampus

The chicks preference for a given stimulus is measured via the speed with which it runs in the wheel. They prefer the imprinted stimulus.

prefer a moving stimulus



Has large neurones

The fact that a response to a stimulus that is repeatedly presented will often decline over a number of presentations is well known. This is what we mean by habituation, and it can easily be observed in humans.

For example - your response to a sudden noise will be to look around for the source of the noise - but if the noise is repeated this response will decline.

Habituation is often studied using what is called the startle response in the rat to a noise (it moves its head from side to side and moves around) or the orienting response to a light (it rears up to inspect the light source).

What is less well known is that the response to a stimulus can increase over a number of presentations.

Again this is easily observed in humans, the most common example being the increased reaction to a mildly painful stimulus (e.g. shock).

This has important implications for any demonstration of conditioning as we will see.

Context can prime you for what you’re going to experience so you habituate – e.g. drug using – tolerance

Sensitisation – response goes up with repeated presentations of a stimulus

Habituate to electric shocks rather than sensitise to them


Aplysia prep

The preparation involving the marine sea snail - Aplysia - allows us to directly manipulate neurons and so study basic learning processes. The research reported here was performed in Eric Kandel’s laboratory.

When siphon or mantle is stimulated (e.g by touch) then they contract, as does the gill.



An initial response to weak stimulation that is not in itself rewarding or aversive gradually declines over repeated presentations.

Aplysia initially responds to a gentle touch on the siphon by withdrawing both gill and siphon, but after many trials this response all but disappears.

Siphon sensory neuron synapses on motor neurons for siphon and gill. The connection strength weakens during course of habituation.

see notes



An initial response to weak stimulation is this time strengthened, but this is not conditioning per se.

Aplysia initially responds weakly to a gentle touch on the siphon by withdrawing both gill and siphon, but after an aversive shock to the tail (that’s not paired with the sensory stimulation) this response becomes more vigorous.

Siphon sensory neuron synapses on motor neurons for siphon and gill. The connection strength strengthens during course of sensitisation because of facilitation by the interneuron.

see notes

Not been found in humans or other mammalian brains


habituation and sensitisation

important phenomena in their own right.

Our lives would be quite different without the ability to adapt to an environment that’s noisy or smelly that habituation gives us.

And sensitisation also has adaptive value, as it increases our defensive responses in aversive situations.

The fact that sensitisation exists also poses problems for the unequivocal demonstration of conditioning itself


the problem

Rescorla (1967)

If sensitisation occurs - then the increase in responding after pairing a CS and US may not be due to conditioning at all - but simply another example of sensitisation.

Rescorla argued that a random control (CS and US occur randomly) would generate the same amount of sensitisation - and so conditioning should only be attributed to CS-US pairings if it led to greater responding than in this control group.

Fortunately this turns out to be the case! CS-US pairings have a specific effect over and above any sensitisation that may occur.


suppression ratio

(responses during CS)/(responses during CS + responses during pre-CS)

40% of time – CS = get a shock

Lower scores better

0.5 = equal number of presses

0 = baseline – strong conditioning

How many times press lever before and after CS

Freeze after if conditioning has worked

0.4 – CS is making no differences – therefore random control – same as control

Shows that conditioning does work – gives effect over and above sensitisation


see notes


latent inhibition

McLaren (1990)

Pre-exposure – box for 6 days – here tone 8 times

Control – no tone

Eat something dry before – thirsty when in the box - can tell when drinking

Tone = can get to the water – how much in magazine with water before and after CS

Control – learn quickly

LI take longer and always behind

LI = retardation in acquisition with pre-exposure to a stimulus

see notes


McLaren et al. (1994)

Rats were pre-exposed to a CS in one context, then appetitively trained (conditioned) with that CS in either the same or a different context.

The conditioning measure shown is CS - Pre-CS scores for magazine entry - higher measures indicate more conditioning.

There is a highly significant context-specificity effect, i.e. more latent inhibition (poorer learning) in the same group than in the different group.

This effect does not occur with simple conditioning, i.e. if the experiment is repeated, but instead of pre-exposure we pair CS and US before switching contexts and testing, then there is no effect of the context switch

see notes


LI as a function of age

Kaniel and Lubow (1986)

Unable to find in humans

How long to get 90% right over 10 trials

Takes longer when pre-exposed

No effect as you get older – control processes come in later on

see notes


perceptual learning

Gibson and Walk (1958).

After preexposure to the shapes in the home cage, the rats were better able to discriminate between them when trained in the jumping stand, compared to non-preexposed controls

Food on other side of correct one

Other side locked and mice just fell down – changed to be more ethical

see notes

Hall (1979)

The blue line are the pre-exposed group - They learn faster than the non-pre-exposed controls, as shown by the more rapid reduction in errors across blocks

see notes


McLaren et al. (1989)

A theory that postulates an associative basis for representation development. Associations lead to salience reduction, which means less learning.

Similarities – better at learning differences between stimuli

see notes


differential LI of common elements as a mechanism for perceptual learning

If the shaded area (bold outline) for a set of stimuli that possess this similarity structure becomes less salient - then enhanced discriminability of the stimuli from one another follows.

There are at least two ways of bringing this about: Exposure to the stimulus set (coloured circles) will tend to differentially reduce the salience of the shaded area because those features are encountered more often and more reliably predict one another.

Alternatively, the prototypical stimulus corresponding to the shaded area can be pre-exposed.

More aware of the difference characteristics

see notes


McLaren and Mackintosh (2000)

The measure of conditioning shown is consumption of one of two novel vinegars (e.g. red wine) after conditioning to the other (e.g. balsamic) following the pre-exposure indicated.

Low scores = high aversion = strong generalisation from conditioning of other vinegar.

Less generalisation is indicative of perceptual learning.

Common elements fade out

Can detect the difference

Connoisseurs drank more

see slides


perceptual learning and the face inversion effect

The inverted stimuli are much harder to tell apart.

This makes recognition for upright faces superior to that for inverted ones.

e.g. Civile et al. (2014)

Application of the latent inhibition of common elements to face recognition.

It’s essentially the same analysis as for the vinegars experiment.

The idea is that the “average” or prototypical features of a face become less salient, the “distinctive” or unique features that depart more from the norm become relatively more salient.

This helps recognition of upright faces.

The inversion effect in an old / new recognition experiment can be obtained with categories other than faces as long as they are extremely familiar to the subjects: e.g. experts (dog show judges!) with gundogs.
- Diamond and Carey (1986) - see notes