Animal Studies

Holland and Straub

Supports S-S. Food aversion when paired with noise

Holland

Supports S-S. Aversion to food when paired with tone, as tone was previously paired with illness

Cunningham and Rescorla

Supports S-S. Sensory preconditioning. Aversion to sugar, as acid was paired with illness and sugar and acid were previously paired together

Fudim

Sensory preconditioning. Ate more banana when craved salt as previously paired together

Rescorla and Freberg

S-S associations acquired during sensory preconditioning are subject to extinction

Colwill and Rescorla

Outcome devaluation to discriminate between R-O and S-R. Supports R-O

Verway

Supports S-R. Based on features (line orientation and colour). S-R was found for line orientation but not colour. Due to long term development

Mackintosh

Overshadowing in conditioned suppression. More fear to light if only light was paired with shock than if both light and noise were paired with shock

Kamin

Blocking in conditioned suppression. More fear to noise if both light and noise were paired with shock than if light had first been paired with shock

Rescorla (contingency)

High contingency = high fear

Rescorla and Wagner

Role of surprise. Based on summation

Westbrook and Lovibond

Disagrees with S-S models (due to expectancy differing from memory).

CR may not be the same UR (Preparatory vs protective, drug administration, context shift for fear)

CR varies as a function of the CS-US interval (Hierarchy of defensive distance)

Mixed evidence on failures of US revaluation (disagrees with S-R)

Expectancy = performed better than those with only food memory (challenges SoP A2 state)

The perruchet effect in humans (expectancy and gambler’s fallacy

Problems with HeiDI and other stimulus substitution models as a combination of factors determines which CRs are triggered by US

Ommission effect = suggests reflexive model

Overall = integrated expectancy model is most likely (supports S-S with the specific role of expectancy; allows for differences in CR and UR by suggesting a specific expectancy state)

Zimmerhart and Rescorla

Extinction of conditioned inhibition. Inhibitory properties don’t diminish when presented alone

Rescorla; Pearce and Redhead

Superconditioning. Increased response intensity when stimulus is really surprising

Kremer; Rescorla and Wagner

Overexpectation. Lambda<SumV = associative strength decreases

Reiss and Wagner

Preexposure retards acquisition of both excitatory and inhibitory conditioning

Pearce and Hall (attention)

Controlled and automatic attention. Greater attention if previous trial has large expectancy discrepency. Habituation and latent inhibition reflect decline in attention. Consistent consequences lead to latent inhibition. Partial condition = increased attention. BUT does not account for context specific latent inhibition

Lovibond, Preston and Mackintosh

Contextual specificity of latent inhibition. If rats with preexposure changed context, they no longer showed latent inhibition

Wagner’s SoP model

Inactive = waiting to be activated

A1 = elicts UR, initial reaction

A2 = less effective processing, secondary activation, like habituation and latent inhibition

Inactive → A1 = Presentation of novel stimulus

Inactive → A2 = multiple stimuli (e.g. tone and shock, tone goes to A1, memory of shock stored in A2)

A1 → A2 = rapid decay. Short term habituation

A2 → inactive = decay

If preexposure of a stimulus = will be in the A2 thus less likely to cause UR or be learnt about (habituation and latent inhibition)

Honey, Good and Manser; Honey and Good

Context specificity of habituation. Increased attention in different context but decreases quickly to same level as same condition

Pavlov (stimulus substitution)

CR to the CS often resembles UR of the US

Problem = CR does not always resemble UR and can be influenced by the natures of the CS

Problem = Different measures provide the basis of drawing potentially opposite conclusions about associative strength

HeiDI model

Solves Pavlov’s stimulus substitution problems

Reciprocal associations allow conditioning to be affected by properties of both CS and US

Mitchell and Hall

Applications of discrimination learning to perceptual learning in humans

Strong evidence of Rescorla; Mackintosh; and Hebbian theories of discrimination

Presence of feedback links to human perceptual learning

Preexposure of a stimulus in rats leads to better performance of discrimination (particularly difficult to discriminate)

Likely partly due to latent inhibition (but cannot fully explain it)

Pavlov (stimulus generalisation)

If you create a CS then the CR will happen to a stimulus that's close enough to the CS (but to a lower extent)

Hilgard and Marquis

Discrimination is needed when it leads to different outcomes

Rescorla-Wagner rule

Elemental theory of discrimination

Discrimination occurs when the unique element leads to positive and the other unique element leads to negative (common element remains neutral)

Problems: does not explain negative patterning (except with configural cue); lack of catastrophic interference; does not account for similarity (based on summation)

Wilson and Pearce

Lack of catastrophic interference. If A+ and AB- (will respond less to AB-). Then B+. Does not remove discrimination of AB-

Redhead and Pearce

Similarity in discrimination learning. If A+, BC+, ABC-; then BC+ is harder to discriminate than A+ due to increased similarity to ABC-

Pearce (configural theory)

Compares stimulus to memory of stimuli

Strength of CR will depend on similarity

Smith et al

Suggests categorisation is due to a combination of exemplar and prototype theory with categorisation rules

Largely disagrees with exemplar because when prototype was directly compared to exemplar, prototype was almost always more accurate

However the XOR task forces exemplar and showed 75% accuracy

COVIS model uses neuroscience to suggest category rules (Similar to humans and that requires some verbal system so primates perform better than pigeons)

Neuroscience suggests that exemplar and prototype use distinct brain regions and starts with exemplar then moves to prototype as training continues

Transfer on rule-based task but not information-integration tasks suggesting categorisation rules

Lots of evidence of concrete categorisation

Hernstein et al

Cerella

Watanabe et al

Scarfa et al

Rote learning as an explanation for categorisation

Not a good explanation as categorisation occurs with novel stimuli

Feature learning as an explanation for categorisation

Most likely explanation.

Edited photos and removed features (pigeons categorised based on subtle colour feature)

Works for novel stimuli

Exemplar theory for categorisation

Similar to rote learning

Issue: scrambling is not catastrophically disruptive (suggesting feature or prototype theory)

Support: Scrambling can be disruptive even if features are preserved but breaks configuration (could be a potential support for prototype theory)

Abstract categorisation

Matching to sample = ‘sameness’ but could be rote learning

Evidence that only some species can do this (mainly primates)

Issues with matching to sample = could be based on familiarity

Inside/outside = relationship not familiarity, took a while until pigeons were able to learn so maybe not abstract

Bar height = took so long to learn it suggests rote

Anecdotal support

Second-order = forced relationship but not familiarity. Initially only those with language ability could do it but then non-language chimps were able to complete it, if it was adjusted

Harrison et al

Cultural evolutionary theory = adaptive value of information to naive individuals

Some evidence of normative conformity (similar to humans)

Homophily has been found

Imitation and mimicry improve social relations

Webster

Non-grouping animals also show social learning

Although method of doing so is more indirect (e.g. scent markers)

If direct observation they take longer to learn

Galef

Socially acquired food preferences

Stimulus enhancement (does not create preference but increases interaction with cue)

Mate selection (quails)

Social facilitation and stimulus enhancement not imitation (bottle caps)

Preferences can occur prenatally

Hayes and Hayes

Do as I do test (imitation)

Animals imitation as it caused them to be fed (methodological issue?)

Akins et al (imitation)

Move screen either left or right (only saw it moved one way)

Akins et al (two action control)

Showed imitation

Either jump or pecked treadle

Mineka and Cook; Cook and Mineka

Fear of predators

Found learning was necessary and the monkeys did not have an innate fear of snakes

Had experienced monkey react fearfully to snake and naive monket watched the experienced monkey’s reaction

However when replicated with neutral stimulus suggesting it’s specialised

Suggesting ‘preparedness’ or ‘selective’ learning

Selective Conditioned taste aversion

Does not matter what causes the illness, rejection of food still happens (e.g. even when the real source of illness is known people still dislike paired flavours)

Ease of association = it’s easier to associate food with illness than light/etc

Palatability = looked at pairing sucrose with either LiCl or shock. Both led to rejection of the sucrose but the orofacial reactions differed (disgust for LiCl and positive for shock). Suggesting dissociation between taste reactivity and consumption reinforced idea that CTA is about rejecting foods paired with illness.

Similar findings have been found in humans

However, limited research and contrary evidence exists (such as context cues and learning about context blocks learning about tastes, suggesting there is something unique but not selective learning)

Foote et al

Rats showed metacognition (refused to do tasks that they deemed too difficult)

But failed to create a model to exclusively explain metacognitive behaviour

Krupenye

Suggests studying ToM has human applications as we are unsure if ToM is due to inhibitory control

Evidence of chimps showint ToM by hiding food

Blue jays food preference

When accounted for cognitive demands chimps were able to complete false belief tasks

Gallup et al (issue with mirror test)

Found issues with many self-recognition tasks (mainly mirror tests) i.e. misinterpreting results, ambiguous results

Acknowledges there is some strong evidence of chimps possessing self recognition

Hopkins

Cortical thickness = improves performance on mirror tests

Gallup (mirror test)

Apes, elephants and dolphins were able to complete mirror test (although elephants and dolphins required preexposure to mirror)

Some evidence of fish being able to pass the test

Studies for Associative structures in simple conditioning

Holland and Straub

Holland

Cunningham and Rescorla

Fudim

Rescorla and Freberg

Colwill and Rescorla

Verway

Studies for conditions of learning

Mackintosh

Kamin

Rescorla (contingency)

Rescorla and Wagner

Westbrook and Lovibond

Studies for introduction to theories of learning

Westbrook and Lovibond

Zimmerhart and Rescorla

Rescorla; Pearce and Redead

Kremer; Rescorla and Wagner

Studies for attention

Westbrook and Lovibond

Reiss and Wagner

Pearce and Hall (attention)

Lovibond, Preston and Mackintosh

Wagner (SoP)

Honey, Good and Manser; Honey and Good

Studies for the expression of learning in behaviour

Westbrook and Lovibond

Pavlov (stimulus substitution)

HeiDI

Studies for discrimination learning

Mitchell and Hall

Pavlov (stimulus generalisation)

Hilgard and Marquis

Rescorla-Wagner rule (elemental theory)

Wilson and Pearce

Redhead and Pearce (similarity)

Pearce (configural)

Studies for categorisation

Smith et al

Lots for concrete

Mixed for abstract

Studies for Social and selective learning

Harrison et al

Webster

Galef

Hayes and Hayes

Akins et al (imitation - screen door)

Akins et al (imitation - two action control)

Mineka and Cook; Cook and Mineka

Selective conditioned taste aversion

Studies for animal minds

Foote et al

Krupenye

Gallup et al (issues)

Gallup (mirror)

Hopkins

Ostojic et al

Ostojic et al

Scrub jays gave mate food that they showed a preference for