lecture Flashcards
what is the definition for Pavlovian classical conditioning?
a behavioural change caused by a predictive relationship between a signal (controlled stimulus) and a biologically importat stimulus (unconditioned stimulus or reinforcer)
what was meant by ‘insight’ when observing the mentality of apes by Kohler?
claimed to observe ‘insight’:
changes in behaviour from trial to trial not explicable in terms of observable trial and error or reinforcement
adoption of a more cognitive approach to psychology over behaviourism
however, his account of monkeys stacking banana crates on top of each other to reach for a banana hanging from the ceiling is firstly anecdotal and secondly, the crates had been there for weeks and the monkeys had stacked them before
so may not be insightful problem solving
describe Skinner and his approach to animal cognition of radical behaviourism?
why initially successful?
weaknesses?
suggested that all we can observe is behaviour so psychology should only be about behaviour
performed box studies with rats: key concepts of operant response and reinforcement and key procedure of shaping
this operant conditioning (reward vs punishment) was applied through ‘behavioural modification’ to clinical etc. settings
initially successful:
- replacing unsolvable theoretical arguments with directly observable ‘control over behaviour’
- synthesising cognitive effects by using operant conditioning procedures
failed:
- denial of an ‘inside story’ incorrect
- attempt to account for language was mocked by Chomsky (showed that humans have a predisposed ability to pick up language from birth, which Skinner argued against)
- experiments on operant conditioning in humans showed differences from animal research
what is meant by the dual process theory in human cognition?
human mental life is understood in terms of a combination of associative ad more cognitive (rule-based) principles
give some examples of human cognition being applied to animals?
(as opposed to usually the other way around)
Sutherland & Mackintosh - ideas about attention from human studies used to explain discrimination learning
Olton - demonstrations of powerful spatial learning seemed to call for animal concepts of memory
describe social co-operative learning in pigeons?
- example of being wary when imputing our more cognitive abilities to other animals and should instead but focusing on the associative (complex over simpler)
one pigeon ‘the sender’ can see two lights
other pigeon ‘the receiver’ can see right hand or left hand response keys
both get reward if coordinate red light with left and green light with right
can solve problem after a lot of training
but not what most people would view as ‘communication’
- communication?
receiver develops a position habit by staying on either right or left hand side due to getting food 50% of time
sender notices that gets food when pecks certain light compared to other, due to receiver pigeon staying in same place and also either can or can’t see pigeon so using spatial information to get grain
more like conditioning than communication
describe imitation in the rat?
- example of being wary when imputing our more cognitive abilities to other animals and should instead but focusing on the associative (complex over simpler)
Heyes et al
rats observed a demonstrator rat pushing joystick to demonstrators right
then transferred to demonstrators chamber to push it in either direction
majority pushed in same direction of demonstrator, despite both directions releasing food (so this not primary motive)
is this imitation? no as pushing joystick with snout will leave one side scent marked, making it the preferred side for the observer
as shown trhough when joystick replaced with unscented one, preference for ‘imitated’ side disappears
describe self-awareness in chimpanzees?
- example of being wary when imputing our more cognitive abilities to other animals and should instead but focusing on the associative (complex over simpler)
- Gallup, 1970
Gallup, 1970
reactions of chimpanzees and macaques to their mirror reflections (mirror self-recognition experiment (MSR)):
over time, chimpanzees showed increase in no. of self-directed behaviours that relied on the use of the mirror e.g to groom parts of body usually visually inaccessible and picking bits of food from teeth with aid of mirror
macaques (monkeys) however reacted to mirror socially, as if treating it as a conspecific (which the chimpanzees did during first couple of days)
the mark test:
bright red marks put on visually inaccessible locations on body
when re-exposed to mirror, chimpanzees touched mark more than other body parts
macaques didn’t touch marks more than other parts
(also found in dolphins using mirror to look at marked parts)
- self-awareness/concept? perhaps not as e.g when shaving, able to use mirror as feedback to guide shaving without explicitly recognising it as yourself, so these 2 events don’t necessarily happen together so chimpanzees don’t necessarily need to know that its them using the mirror to use it to guide their actions
describe self-awareness in chimpanzees?
- example of being wary when imputing our more cognitive abilities to other animals and should instead but focusing on the associative (complex over simpler)
- Chang, 2015
Chang et al, 2015
rhesus monkeys trained to touch an irritant light spot on their head using a mirror
after 2-5 weeks, could touch non-irritant light spot or dye in front of mirror
5/7 showed mirror-induced self-directed behaviours like touching mark and smelling/looking at fingers and using to explore unseen body parts
4 monkeys did same thing but without light spot irritant (visual-somatosensory training) didn’t pass mark test or show any mirror-induced self-directed behaviours
- showing not clear divide between macaques and chimpanzees as after a while they stop treating mirror image as conspecific and start eprforming mirror-induced self-directed behaviours
suggestion about the need for conscious awareness in Pavlovian conditioning?
(McLaren et al, 1994 textbook chapter)
suggested that successful Pavlovian conditioning in humans relies on their conscious awareness of contingencies between the unconditioned stimulus and the conditioned stimulus
and no evidence of conditioning shown when unaware of the contingencies
what are the different interpretations of the meaning explicit and implicit learning?
(McLaren et al, 1994 textbook chapter)
explicit = conscious and implicit = unconscious (so recall could be either depending on whether consciously retreiving something or not)
explicit = declarative knowledge enabling verbal report and implicit = tacit (understood without being stated) knowledge of abstract rules underlying
processing distinction: explicit = hypothesis-testing processing when problem-solving and implicit = automaticity in processing
explicit = involving both cognitive and associative processes (dual process theory to understand human mental life) and implicit = merely learning through association
which approach is predominantly used to describe animal learning theory?
and examples where this may not be true?
(McLaren et al, 1994 textbook chapter)
associationist approach to learning
whereby animals solely learn through associations (implicit learning)
characterised in terms of the establishment of links between representations e.g Pavlovian conditioning
habituation and sensitisation may be non-associative precursors to associative learning
suggestions of underlying cognitive processes unerpinning learning may be found in some animals
explain the Overtraining Reversal Effect (ORE)?
McLaren et al, 1994 textbook chapter
in rats, overtraining of discrimination between black and white cards (through treating the former), they will learn reversal of this discrimination faster than those who were just trained
due to in overtrained, black card a good predictor of reinforcement to follow and a relevant cue (between colours not positions etc.), so higher associability, so entering into new associations more rapidly as hypothesis test this cue first to gain treat before others
whereas in just trained to criterion, not as good a predictor and associability of relevant cue not as high (compared to other irrelevant cues), so reversal of discrimination learning is slower due to testing other cues as well
therefore seems a better example of associative over cognitive learning
describe the cognitive and associative account to explaining the effects of ‘blocking’ on animal learning?
(McLaren et al, 1994 textbook chapter)
blocking is a method employed where initially pair stimulus (e.g light) with reinforcer (e.g food) but then in second trial a second stimulus (e.g a tone) will co-occur with the first stimulus with the same reinforcer
animal will learn little about relationship between the second stimulus (tone) and reinforcer (food) due to being ‘blocked’ by the previous relationship of the 1st stimulus (light) with the reinforcer
- will be disrupted when the affective characteristics of the 2 US’s presented one on each training session are different but e.g food in 1st and water 2nd will enable blocking to occur as both appetitive
cognitive explanation - light becomes a good predictor of food, whereas tone is more ambiguous when compared to the light, so wait for further info about whether tone good predictor (playing it without the light)
associative explanation - blocked stimulus (tone) with decrease in associability compared to the first stimulus, due to being a poor predictor of reward
- better supported through animals being slower to learn that tone predicts food when tone played separately to the light, having been blocked (cognitive account would predict that learning would be fast as ‘wait for info about tone as predictor in isolation’)
describe the double dissociation between implicit and explicit learning in the serial reaction time tasks in humans?
(McLaren et al, 1994 textbook chapter)
lights flash and have to respond to it by pressing correct key
in experimental group, lights flash in a certain pattern, enabling prediction of future flashes
implicit - when faster reaction time due to continguency compared to controls, in absense of explicit processes
explicit - when verbally able to predict future flashes or when able to verbally explain the contingencies between flashes (relying on working memory and episodic memory in the long term)
both potentially explained through associative learning as implicit learning of contingencies and gain a ‘feeling’ regarding which flash will appear next
indeed found that pigeons show positive transfer from sequence production (faster RT) to sequence discrimination (predictions) (both elements of ^^^)
what was Tolman’s view on understanding human cognition?
claimed that decisions made by rats in a maze setting can tell you all you need to know about human cognition
bar social psychology and language learning
what is the key difference between instrumental (operant) and classical conditioning?
in classical conditioning (Pavlov), the animal/human doesn’t have to do anything to learn
in instrumental/operant conditioning the human/animal has to perform actions and learns from the consequences
what is meant by the law of effect and what were the findings prompting this?
Thorndike, 1898
placed cats in a puzzle box and gave them incentive to escape
once they found the correct thing to press to escape, time taken to escape would reduce
leading to the ‘law of effect’ whereby any behaviour followed by positive consequences was likely to be repeated but followd by unpleasant consequences is likely to be stopped, so reinforcement establishes link between the stimulus and response
(led to theory of operant conditioning)
issue: reduced process in animal to mere stimulus-response connections, with no place for expectancy of reward/punishment hence unable to explain findings like increased errors in a maze when reward decreases in quality compared to a control group - as were expecting a better quality reward
what is meant by scala naturae of mental function, proposed by Romanes?
suggested that at one end of the ‘ladder of nature’ is amoeba (least intelligent) and at the other is humans (most intelligent)
and that all animals can be placed on a scale within this regarding level of mental function
e.g rats and pigeons in middle and humans, chimpanzees, parrots and dolphins at the top
however, cannot assume that there is this
what is Morgan’s Canon regarding human and animal behaviour?
that humans tend to see animals doing something amazing and think ‘wow that’s exactly what i would’ve done’ and infer that they have the same mental functions
but instead, should use the simplest principles to explain that behaviour and only explain using complex principles when explicit and undeniable evidence for it
give an example of natural non-associative exposure learning in the male white crown sparrow?
Marler
learn their song in a critical period early in life but if this is missed then an impoverished song is produced
e.g in isolation experiments, these birds will show deficiencies in their own song upon maturation (no access to intact adult song) but does still contain some valid elements
young birds reared in presence of taped songs will learn and present that song, even if from another species (little difficulty for song sparrow to learn swamp sparrow song)
give an example of natural non-associative exposure learning in the aplysia (marine sea snail)?
very simple organism so enables direct manipulation of neurons to study basic learning processes
Kandel
when siphon or mantle or gill (under their body) stimulated (touched) then contract
show habituation - (reduced response over repeated presentations without reinforcement) when usually respond to touching siphon by withdrawing gill and siphon but after many trials this response mainly disappears
the connection strength of siphon sensory neuron synapses on motor neurons for the siphon and gill weakens during habituation due to tiring out the synapse and reduced levels of neurotransmitter release as overuse of the pathway
show sensitisation - after aversive shock to the tail, initially weak response to touch by withdrawing gill and siphon becomes more vigorous
connection of siphon sensory neuron synapses on motor neurons for siphon and gill strengthen due to facilitation by interneuron (transmitting impulse from tail sensory neurons to synapse between siphon sensory and motor neurons) and shown by increased motor neuron activity
what is meant by habituation?
can this effect occur the other way round?
- form of exposure learning
response to a stimulus that is repeatedly presented which is not rewarding or aversive will often decline over a number of presentations
e.g responding to noise by turning round and looking at it but if repeated then response will decline or being in a bad smelling environment but not noticing after a while
studied using
- startle repsonse - a rat will move head from side to side and move around when noise presented
- orienting response - rat rears up to inspect a presented light source
also, response to a stimulus can increase over a number of presentations e.g increased reaction to mildly painful stimulus (electric shock) (sensitisation)
suggested that habituation depends on the animals ability to remember stimuli for short periods of time i.e if a distractor then may dishabituate (no decrease in level of response) due to removing memory trace of first stimulus
what is the issue regarding sensitisation and the existence of conditioning?
and solution for this, found by who?
- form of exposure learning
if sensitisation occurs (increase response with repetition) then the increase in responding after pairing a CS and US may be sensitisation and not conditioning at all
Rescorla, 1967 argued that a random control (CS and US occur randomly) would generate the same amount of sensitisation, so conditioning should only be attributed to CS-US pairings if it led to greater responding than this control group (as if same response then sensitisation)
CS-US pairings have specific effect over and above any sensitisation that may occur (much lower suppression ratio meaning higher conditioning when probability of shock when CS presented than when equal in both conditions or high probabilitywhen CS not presented) (Rescorla, 1968)
what are several forms of complex and non-complex exposure learning?
and why are they more complex?
non-complex - habituation & sensitisation
complex - latent inhibition and perceptual learning - important influence on how and what animals and humans learn
who discovered latent inhibition and what is it?
-complex form of exposure learning
Lubow & Moore, 1959
retardation in learning to a stimulus brought about by simple pre-exposure to that stimulus with no noticeable outcome
therefore find it more challenging to learn when stimulus paired with certain outcome
what was found about latent inhibition by McLaren et al, 1994?
-complex form of exposure learning
demonstrated the context-specificity of latent inhibition
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
highly significant context-specificity effect, whereby more latent inhibition (poorer learning) in the same group than in the different context group
effect of context switch doesn’t occur with simple conditioning (if instead of pre-exposure, CS and US were paired before switching contexts and testing)
what was found about perceptual learning by Gibson & Walk, 1958?
PL - faster acquisition of a discrimination due to pre-exposure
after pre-exposure to shapes in the home cage, rats were better able to discriminate between them to get food when trained in the jumping stand, compared to non-preexposed controls
pre-exposed group showed faster learning, as illustrated by reduced error rates
what is perceptual learning?
faster acquisition of a discrimination as a result of pre-exposure to stimuli that bear an appropriate relationship to the target stimuli
e. g 2 pictures being shown before discrimination task enables animals to better tell the difference between them - as opposed to learning discrimination slower in control group not pre-exposed
e. g taste aversion to one solution does not generalise to other solution if pre-exposure to both - as opposed to generalising aversion to both if not pre-exposed to them
so unlike latent inhibition, pre-exposure aids learning, and doesn’t hinder it (due to enabling more rapid discrimination between the unique features of the stimuli, without exposure would pay more attention to features they have in common - as long as common element receives more pre-exposure than unique elements)
in which ways may different species of animals have the same intellectual processes?
face many of the same problems and therefore have to have the same intellectual processes in order to sovle them
e.g finding food, locating young, hiding from prey
ability to learn about recurring sequences of events in order to repeat beneficial behaviour and prevent behaviour with negative consequences
how to use brain size to determine intelligence of a species?
can’t just use brain size as may have bigger brains due to when bigger, need for increased function e.g elephant
find two animals with similar body size then compare the size of the brain (brain mass/body mass - cephalisation index), so brain size isn’t dependent on brain to body ratio, which may account for the size of the brain
assisted by cephalisation index and coefficient which calculates this ratio
should look for deviations from plot of brain mass against body mass - e.g monkeys showing differences in intelligence across species
corresponds to results from Nakajima study on students ranking animal intelligence
the animal with the bigger brain in this situation may then be assumed to have more intelligence
e.g humans and ostrich same body weight humans have bigger brain
for some part of the brain, bigger may = higher intelligence e.g hippocampus and spatial memory
issue: e.g amphibians seem less intelligent in this way but bodies heavier due to water supporting this
describe imprinting as a natural form of exposure learning?
Bateson & Horn
chicks preference for a given stimulus is measured via the speed with which it runs in the wheel runs faster (moving towards) the imprinted stimulus
found changes in synapses in teh area equivalent to the hippocampus after imprinting
what did McLaren 1990 find about latent inhibition in mice?
put in box for 6 days and would hear a tone 8 times (preepxosure condition) whereas control didn’t hear these tones
had 4 days of conditioning where after tone, had temporary access to water from a magazine source
found that preexposure group learnt nothing on day one and were generally much slower learners than the control group
so latent inhibition as retardation of acquisition of learning due to pre-exposure to a stimulus
what did Kaniel + Lubow, 1986 find about latent inhibition as a function of age?
found that at 4-5, when pre-exposed, children took a lot more time to get trained condition to over 90% correct, indicating poorer learning
whereas at 7-10, were a lot faster and no pre-exposure effect, so no latent inhibition exhibited (unlike that found in animals)
suggested this is due to control processes preventing latent inhibition
task was indicating which side one of 2 outside cards was by pressing a key (pre-exposure to middle card of different sized black or white squares)
then in test phase, had to press a key for the side a certain square was that they had been exposed to in the study phase
supported by a recent study by McLaren et al, 2019 finding latent inhibition and lower accruacy and longer response time to pre-exposed over novel stimulus (control group) in 4-5 year olds but no such effect in 7-10 year olds
how was perceptual learning demonstrated from McLaren & Mackintosh, 2000?
aversion conditioned to a novel vinegar (injection) generalised more strongly to the other novel vinegar in the control group than in either a group exposed to multiple previous vinegars or a single, group vinegar (common to all the vinegars)
so multiple group being familiar with 4 different vinegars enhanced their ability to discriminate between the 2 new vinegars, thus not generalise aversion between them (demonstrating perceptual learning aided through pre-exposure)
describe the face inversion effect?
better at discriminating upright faces in tasks than upside down ones
found dog experts showed face inversion effect for dogs due to experience at doing this (preexposure) but novices showed no effect and discrimination ability equal (Diamond & Carey, 1986) - perceptual learning
‘average’ features become less salient and distinctive ones become more salient which helps recognition with upright faces - latent inhibition
why might speed of learning be a poor determinant of intelligence in humans and animals?
- findings suggest the opposite to intelligence via the cephalisation index (brain:body ratio) bees and carps learnt much faster how to receive a reward constantly than pigeons and human 5 month olds (Angermeier, 1984)
- and speed of learning differs based on measurements used to assess it e.g rats learn much faster when have to press lever for reward than when have to perform this same action to avoid an electric shock (Bolles, 1971) and in taste aversion, when poison paired with salt and light+click paired with shock
- found dolphins had difficulty forming learning sets with visual but were fine with auditory stimuli
- differences between species makes it hard to devise a task which exerts the same demands on them (what the optimal conditions for testing each species are) e.g learning response for food may depend on motor, motivational and perceptual resources - contextual variables (Bitterman, 1965) - links back to first point as may have been that food was more rewarding for certain species or easier to locate lever etc. - so better a battery of tests looking for patterns (IQ type of approach)
what are some of the direct neurological responses to pavlovian conditioning?
- training makes it easier for one neuron to activate the other in the cerbellum
- event excites neurons to a greater extent at the start rather than at the end of training (maximal impact when event unexpected i.e at the start of training)
describe the basic process of the original pavlovian conditioning?
before conditioning: unconditioned stimulus (of food) provokes an unconditioned response (of salivation) neutral stimulus (tuning fork) doesn't provoke a conditioned response (salivating)
during conditioning:
the neutral and unconditioned stimuli of a tuning fork and food, respectively, provoke an unconditioned response (of salivating)
after conditioning:
the tuning fork is now a conditioned stimulus (due to pairing) and salivation is now a conditioned response to this stimulus
describe the different conditions that pavlovian conditioning works in?
in all vertebrate species that have been tested AND some invertebrates
for any conditioned stimulus an animal can detect
shown with spinal reflexes (e.g in brain damaged patients)
in humans e.g blinking
describe some of the common pavlovian procedures?
eye blink to an air puff in humans / galvanic skin response (GSR)
key pecking in pigeons (‘autoshaping’)
taste aversion in rats (falvour + illness)
conditioned suppression of lever pressing in rats
what are the 4 basic effects of pavlovian conditioning?
- stimulus generalisation
- extinction
- overshadowing
- blocking
what is stimulus generalisation?
- as 1/4 basic effects of pavlovian conditioning
- conditioned response will occur to similar stimuli to the conditioned stimulus, the response weakens the further away from the original stimulus it becomes (normally distributed with CS in middle - generalisation gradient) e.g pitch
- found in rabbits that wore trained to blink to a tone (via expectancy of a shock) and that still produced this response when tone sufficiently similar to the orginal and when extent of generalisation incomplete (quite far removed from original) then generalisation decrement
- neural explanation: due to exciting some but not all of sensory neurons initially excited by orginal CS, thus eliciting the CR albeit a weaker one, and explaining that the less similar, the weaker the response as the less sensory neurons excited
what is extinction?
- as 1/4 basic effects of pavlovian conditioning
- start with a well trained conditioned stimulus, then repeatedly present on its own and the CR will weaken as non-reinforced presentations
what is overshadowing?
- as 1/4 basic effects of pavlovian conditioning
Mackintosh, 1976
- train light and noise together (compound) or separately and use different intensities of noise for different groups
- test the light and noise
- light and intense noise paired together has the highest suppression ratio (lack of conditioning) in presence of light as intense noise overshadows light (fine when non intense as only most intense grabs learning and if equal then balanced)
- light and weak noise in training has highest suppression ratio (worst conditioning) in presence of noise as light overshadowing (not the case when intense noise)
what is blocking?
- as 1/4 basic effects of pavlovian conditioning
Kamin, 1969
pair initial stimulus with result (shock) so conditioned response is fear
then in second stage, pair noise AND light (another stimulus) with shock so noise and light lead to fear
in the test, the group who went through both stages showed a lack of suppression to light as blocking prevents relationship between light and shock to be indepedently formed as blocked by pre-existing relationship of noise and shock (0.45)
however, in control condition who didn’t received first stage, had much higher suppression ratio (0.05) to light due to having light and noise paired together, not originally blocked by one of the stimuli (information is novel+surprising, not ‘redundant’ i.e predicted via other stimulus)
what is the relationship between timing and success of conditioning?
when conditioning delayed - conditioned stimulus before unconditioned stimulus then easily established conditioning
when trace conditioning - conditioned stimulus much before the unconditioned stimulus then ease of conditioning depends on length of trace
when simultaneous conditioning - conditioned and unconditioned stimulus presented at the same time, very little conditioning established (shows have mechanisms for understanding that A can be used to predict B, so why it doesn’t work at the same time)
(shown by conditioning being optimal after a 4-8 sec interval and after a couple of minutes, completely drops off)
what is the Rescorla-Wagner Learning Rule?
the combined associative strength of all conditioned stimuli present on the learning episode, predicts the outcome
strength of associations of UCS and CS representations predicts the strength of CR
associative strength doesn’t necessarily translate into responding
associations form synaptic connections so spreading activation whereby representations of CS can activate representation of UCS
what was then, and is now, the significance of pavlovian conditioning?
then:
- form of learning described entirely objectively
- scientific underpinning for concept of association
- contributed to dominance of behaviourism
now:
- reliable phenomenon with determinable laws
- practically useful e.g behavioural therapy
- may account for involuntary forms of human learning e.g phobias
- continued investigation of theories of pavlovian conditioning
- not an explanation of ALL learning but gives insight into fundamental learning processes which may apply to humans
briefly explain conditioned suppression in rats?
! explanation of suppression ratio !
rat learns to press a lever steadily to get food
then pair a tone with a shock
see whether rat freezes when tone played (determined via comparing lever pressing before tone and after tone, and not number of presses as varies between rats)
so suppression ratio closer to 0 indicates better conditioning as bigger difference between lever pressing before and during CS whereas 0.5 if no conditioning as exhibit the same behaviour in both conditions
describe blocking in human participants?
asked to rate likelihood of an allergic reaction when pre-trained with allergic to avocado
then after told allergic to avocado and bacon
bacon given lower rating of likelihood than avocado as blocks formation of relationship between bacon and allergic
demonstration of humans understanding contingency between pressing and a white light?
found that when rated the connection with their pressing and a white light coming on as 0 (meaning tapping doesn’t make a difference) when in a group of white light coming on half the time
found that when light would come on 1/8 when tapping and 7/8 when not, rated much more negatively the effect of tapping on a white light appearing
so shows ppts can track contingencies without working out through calculations and instead just engaging with and thinking about the material
explain the effect of context in conditioning key pressing with a white light in rats?
when white light occurs without the rat pressing a key, context (a constant) will overshadow key pressing so blocking key pressing
so will learn less in this condition as reduced contingency due to interference of context (overshadowing association of key with white light)
what is the suggestion regarding ‘pavlovian conditioning’ demonstrated in humans by Lovibond, 1992?
- cognitive, conscious
- automatic
found that pairing tree with shock and leaves without, humans able to figure out contingency between shock and tree through expectancy ratings and also, skin conductance hgiher for shock pairing than for non-paired picture
BUT for 30% of ppts, not aware of contingency, which manifested in their expectancy ratings and skin conducatnce being similar for both pictures, suggesting that cognitive, conscious awareness necessary so not associative
Hughdahl + Ohman, 1977 found that when ppts instructed that no more shocks would occur, then immediate extinction shown by skin conductance, compared to when non-instructed and merely extinction, which led to slower extinction (cognitive control can switch off response)
lack of cog control - found that when instructed extinction for fear relevant stimuli, not the same immediate extinction as for irrelevant so responses not governed entirely by conscious expectancy and instead an automatic element, regarded as being due to Pavlovian conditioning
when can associative learning be said to have take place in an animal?
when there is a change in an animals behaviour as a result of one event being paired with another
what is an example of autoshaping?
panel illuminated for 5 secs and food delivered
after a while, pigeons will peck it rapidly when illuminated
example of pavlovian conditioning as pairing of panel with food sufficient to create a response of key pecking
findings in rabbits doubting the effect of pavlovian conditioning on a conditioned response?
Sheafor, 1975
2 groups of rabbits: tone + water straight after or tone and water separated by 12 minutes
found that despite second group not having paired tone and water, a conditioned response of jaw movement still occurred (albeit less than in other group) when tone presented (maybe due to constantly expecting water and salient stimulus of tone evokes a response)
suggests need for a control group given truly random control of sometimes pairing tone with water and sometimes separately, to show that the group of consistently paired had stronger response (jw movement) showing better conditioning
difference between excitatory and inhibitory pavlovian conditioning?
excitatory - conditioning allows an animal to learn that one stimulus signals another
inhibitory - conditioning allows an animal to learn that one stimulis indicates the other will not occur
example of inhibitory conditioning in pigeons?
Hearst + Franklin, 1977
food never available when a key was lit
as training continued, pigeons displayed tendency to move away from key that was illuminated
suggesting able to learn that stimuli signal omission of food and that they can withdraw from this
so technique of ensuring US is delivered in absencse but not presence of CS
what is meant by the retardation test to demonstrate inhibitory conditioning?
pairing conditioned inhibitor directly with the US so if inhibitory conditioning results in stimulus being regarded as a signal for absence of US, then difficult to then convert this stimulus into signal for US
Pearce et al, 1982
inhibitory CS paired with shock (slower conditioning than for below group)
control group had lack of prior training so received the stimulus paired with a shock
what is meant by the summation test to demonstrate inhibitory conditioning?
stimulus being a signal for omission of US so this should occur whenever accompanied by a CS that has been paired wih the US in question
Pearce et al, 1982
tone + shock, clicker + shock, clicker-light compound + nothing
strong CR when presented with tone and clicker but not third group
tone trials (presented alone) mixed with light-tone CR much higher for tone than compound so influence of light transferred well from clicker trial (where nothing happened) to tone
so successful summation test as higher supression ratio (not inhibiting response) when light-tone than just tone
what is meant by the term ‘alpha conditioning’?
strengthening of a response through conditioning that was already elicited by the CS e.g withdrawing the gill when siphon or mantle shelf stimulated (Carew et al, 1983)
demonstration of pavlovian, excitatory conditioning in the sea snail?
Carew et al, 1983
when stimulating siphon or mantle of sea snail, gillwould be withdrawn, which was strengthened through the addition of a shock in one condition
in the condition with a shock(CS+), withhdrawal response much stronger and without a shock (CS-) much weaker
example of alpha conditioning (strengthening response already present to CS)
Hawkins et al, 1983
stimulation of siphon followed by tail shock whereas of mantle didn’t
recording electrode placed in motor neuron in contact with the 2 sensory neurons
showed conditioning increase electrical activity considerably when shock compared to when not paired with shock
(ability to excite motor neuron increased when paired with shock due to release of serotonin from interneuron in tail after shock, if sensory neuron fires at same time as this release, causing relatively permanent change enabling release of more neurotransmitters in future from sensory N, so when stimulated, stronger response in motor neuron)
describe the assumptions of the memory model for associative learning?
activation of representation of UCS will active response system thus eliciting unconditioned response
so repeated pairing of UCS with CS will develop connections so when CS presented, will excite the UCS response center and lead to response mimicing the one elicited by UCS
OR
growth of connection between CS and UCS representations so presentations of CS will excite representation of UCS which will then excite a response
what is stimulus-stimulus learning?
- types of it?
associations can develop between 2 stimuli, even when neither of them has any unconditioned properties
- serial conditioning - sequence of stimuli precedes UCS e.g tone-light-food and tone may lead to CR
- sensory preconditioning - initial pairing of 2 stimuli then second pairing of one of the stimuli with something else, causing response to first stimuli not present in the second stage due to memory activation
- second-order conditioning - normal pairing of NS (CS1) with UCS until CR regularly occurs, then new stimulus paired (CS2) with neutral (CS1) and UCS omitted, and CR may be observed with second stimulus despite never being paired, due to CS2 activating representation of CS1 which does the same for US and elicits response
- shown by extinction of food pairing with CS1 leads to reduced responsivity to CS2 as representation has changed (Rashotte et al, 1977)
what are the 2 different characteristics that unconditioned stimuli possess, according to Konorski, 1967?
- specific - those that make the UCS unique, where delivered, intensity etc.
- affective - UCS has in common with other stimuli and reflect its motivational quality e.g shock has aversive characteristic that the animal knows to avoid
what are the different types of conditioned response according to Konorski, 1967?
- consummatory CR - when CS retrieves info about specific properties of UCS (distinguishing), then this CR would be performed e,g salivation to CS when paired with food
- preparatory CR - when CS retrieved info about affective properties of UCS, would elicit a CR (difference to ^^ being not intimately tied to response elicited by US)
- compensatory CR - CR will oppose/compensate for the unconditioned response e.g drug tolerance whereby CS of imminent injection leads to CR which counteracts the effect of the UCS (drug)
what are the 4 instrumental conditioning procedures?
- positive reinforcement - appetitive - more R
- punishment - aversive - less R
- negative reinforcement - no aversive - more R
- omission training - no appetitive - less R
how did Hull expand Thorndike’s law and effect explanation?
repeat actions leading to ‘satisfying state of affairs’
Hull: reinforcement due to drive reduction, so animal will work for food if its hungry etc.
what is meant by a schedule of reinforcement?
what are examples of these?
a rule for deciding which responses we reinforce
different schedules therefore lead to different, predictable patterns of response, recognisable on a cumulative record
- continuous - CRF - reinforce every response
- fixed ratio - FR - every nth response - response = pause after each reinforcement followed by fast responding
- variable ratio - VR - every nth response on average e.g 8 and 12 if 10 - continuous fast responding
- fixed interval - FI - 1st response after time t has elapsed since last reinforcer - pause after each reinforcement followed by gradual increase in response rate before interval
- variable interval - VI - variable time period -continuous moderate response rate as unsure when interval is
e. g VI60 means mean time from first response to reinforcer given has an average of 60 secs
briefly, what is the difference between ratio and interval response schedules?
ratio - reinforcement depends on number of responses
interval - reinforcement depends on time interval
what are the 2 suggested forms of instrumental learning?
- where animal has knowledge of the consequences of its actions (so be able to represent the outcome and its relationship to the action performed) - make inferences in combination with other knowledge e.g knows outcome valuable under certain drive state and knows outcome prdocued by certain action (never before performed), then can combine this to give appropriate response
- stimulus to response (S-R) reflex supports habitual responding (overtraining leads to this)
what is the Rescorla-Wagner equation for conditioning with a single CS and what does each symbol represent?
∆V=a(λ - V)
triangle V = change in associative strength between UCS and CS on a particular trial
a = doesn’t vary during conditioning and has value between 0-1 (will be given this)
λ = magnitude of UCS, reflecting maximum association that the pairing can achieve
V = strength of the CS-US association
for example:
when a=0.2 and λ=100, on the first occasion CS is paired with UCS, V will be 0, so…
∆V = 0.20 x (100-0) = 20 so increase in associative strength
then second trial…
∆V = 0.20 x (100-20) = 16
third trial…
∆V = 0.20 x (100-36) = 12.8 (as do 20+16 for the V value as association increases)
thus predicting that association strength and therefore CR strength will increase in rapid increments at the start then cease with continued training (when ∆V = 0 and V=100)
how to apply the Rescorla-Wagner equation for conditioning with a single CS to extinction?
∆V=ab(λ - V)
∆V = change in V (associative strength which increases the more conditioned together) ab = given in questions and usually .5 when multiplied λ = outcome so 1 when outcome e.g food given and 0 if not - magnitude of US SV = expectation on trial and predictive value (surprise value on first trial as completely unexpected) - strength of CS-US association (so 0 on first trial)
presenting CS without the UCS means value of λ would be 0 (as no magnitude of UCS)
so if the a = 0.20 and V = 100 (received sufficient training to maximise strength of association) then…
∆V = 0.20 x (0-100) = -20 so first extinction trial will result in CS losing 20 units of associative strength
thus able to explain the timings of extinction as initially rapid decline in association followed by smaller loses in associative strength
complete when ∆V=-100 and V=0
how does the Rescolar-Wagner equation for conditioning with a single CS account for better conditioning when UCS is surprising?
(λ - V) when there is a large discrepancy between these at the start (high magnitude of UCS and low associative strength in that trial) shows that CS will be a poor predictor of the UCS and so occurance of UCS will be surprising
when both values are close together, the CS will be a good predictor of the UCS and the occurence of the UCS will therefore not be surprising
and has beeen shown that conditioning is most rapid when (λ - V) has a large value i.e a large deficit and therefore more surprising leads to faster conditioning
examples of instrumental/operant learning being unable to be explained by pavlovian conditioning?
suggestion that food not reinforcer for pressing lever but instead that contingency built between food and lever so keep pressing
food-tone but not when go to magazine
group of unpaired (food when tone not) and omission (magazine entry during tone cancels food)
found that the omission group gradually learned to stop going in the magazine, showing instrumental conditioning over pavlovian
with guinea pig, left head turn = food after buzzer then learn that right head turn now = food so able to perform opposite action
pavlovian conditining gone against as buzzer maintains same relationship with reward, so ‘natural’ left head turn response not changing due to pavlovian conditioning
differences in extinction responses in trained and overtrained rats in instrumental learning?
press lever for sucrose pellets
devalued group given Lithium chloride on the same day so association betwene pellets and illness
non-devalued group given LiCl next day so not associated
found devalued group pressed lever less as don’t want food anymore
also, when overtrained to press lever for pellets over 500 not 100 trials found that devalued group pressing lever lots as opposed to before due to overtraining making lever pressing a reflex/habit
also, if pellets actually presented then would still press lever lots but wouldn’t eat them (showing habit not want of food)
so suggestion that 2 kinds of IL: animal has knwoledge of consequneces of its actions on the outcome (trained) and S->R reflex supports habitual responding (overtrained)
findings suggesting that representation of the outcome determines performance in rats in instrumental learning?
as one of 2 kinds of instrumental learning
instrumental training: that in light, lever = sucrose water and chain = pellet and vice verse when tone
sucrose water then paired with LiCl
when light on in testing, will mainly pull chain and when tone will mainly press lever as aware these responses will result in desired outcome of pellets not the water
what was found about instrumental learning and drive state?
demonstrating 1/2 types of instrumental learning
trained rats with chain = pellet and lever = sucrose
when made thirsty, should press lever to get water if understand actions for expected outcome but didn’t find difference initially
then, found that they did prefer the sucrose water response (lever) when animal taught that sucorse water is preferrable when in a thirsty state before the test (one reinforcer more valuable in certain state)
so again, demonstration of understanding of response needed to achieve desired outcome not reflex actions (different types of instrumental learning)
going beyond simple association as combining knowledge of what outcome valuable in certain drive state and which action to produce to give outcome
what are the 3 different procedures used when studying discrimination and what are some appartus types used?
- successive discrimination - present one of the stimuli and see how the animal responds - so present one stimuli at a time
- simultaneous discrimination - present 2 stimuli and see which the animal approaches, easier than successive as can compare stimuli when they occur together
e. g S+ is blue circle for tone but S+ is yellow circle for light - conditional discrimination - reinforce different responses in the presence of different stimuli
apparatus types:
discrimination boxes, Lashley’s jumping stands, use of computer displays etc.
what is meant by ‘peak shift’ in relation to simple discrimination?
and an explanation behind this?
Hanson: peck greeny yellow light to get grain (S+) but similar, more yellow light gives time out so less food (S-). found control group with no S- peaked at the S+ colour however, when S- introduced, peak shifted over near S+ (past it), but further away from S-, and also higher and sharper peak
Spence’s explanation of peak-shift: S+ produces excitatory and S- produces inhibitory generalisation gradients. the net responding of the animal is interaction between the 2 (one minus the other). so peak shift occurs when e.g inhibitory gradient low but excitatory high, so difference great and therefore a peak shift in this part of the new curve, this difference is greater than at S+ as it is closer in relation to S- than the new stimulus
prediction that peak shift works best with similar S+ and S- and that shift is greatest in this case, is founded
modern variant using Rescorla Wagner has proven very successful at modelling peak shift
what is meant by ‘transposition’ in relation to simple discrimination?
how explained?
if a discrimination between S+ (one stimulus) and S- (another stimulus) is trained, and then S’ is tested vs the S+, S’ may be chosen
so refers to choosing a different stimulus over the one trained
can be explained via gap between excitatory and inhibitory bigger, so preference for new stimulus over S+ (to get away from S-) so transposition to new stimulus
2 curves on a graph, an inhibitory one with S- at the centre and an excitatory one with S+ at the centre, these overlap and the gap between them is greater for S’ than S+ (similar explanation to that of peak shift)
what is meant by ‘transfer along a continuum (TAC)’ in relation to simple discrimination?
how explained?
training an easy discrimination on a dimension can help the animal acquire a difficult one, more than simply practicing that difficult discrimination
even when total training times are equated
explained via: interacting excitatory and inhibitory gradients so harder when gradients closer and easier when further away. when at H+ (the hard condition), gap between gradients much larger when previously done an easy discrimination whereas gap between gradients when learning hard discrimination only is much smaller
what is the theoretical debate regarding absolute vs relative simple discrimination?
whether effective stimulus enabling learning to respond to one stimulus rather than another (discrimination) is absolute or relative e.g does rat learn to respond to black>white or darker>lighter?
so transposition of discrimination to different values on the stimulus dimension should uncover which one
what is the continuity vs non-continuity theory for simple discrimination?
whether learning is a gradual process (continuity - Spence) or all-or-none (non-continuity - Krechevsky & Lashley)
made harder when realising a continuity account can be made to look like non-continuity (e.g stages of progression) and vice versa
what is involved in the Hull-Spence continuity theory?
- simple discrimination
Spence, 1936
discrimination of absolute stimuli
theory: learning occurs gradually
smooth generalisation gradient around the stimuli to which training has occurred and excitatory generalisation around S+ and inhibitory around S- (internal response tendencies)
observed response tendency predicted from monotonic transformation of the algebraic sum of excitatory and inhibitory generalisaed response tendencies (E-I)
this theory can predict transposition, peak-shift and trasnfer along a continuum
what is involved in Krechevsky and Lashley’s non-continuity theories?
- simple discrimination
discrimination of relative stimuli
theory: learning occurs suddenly
Krechevsky, 1932 - rats form hypotheses about what is to be discriminated and when they get the right hypothesis, the problem is solved instantly e.g left or right then changes to colour difference, which turns out to be correct so instantly know everything
predicts position habits, no impact of pre-solution reversal and transfer along a continuum and fits into modern cognitive ideas about selective attention
biggest support from predicting lack of impact or pre-solution reversal: swapping the S+ from one stimulus to another mid way through learning, and a control group sticking with the S+, found both groups arrive at ‘final solution’ of 100% correct at the same time (so pre-solution reversal making no difference)
continuity theorists wouldn’t expect/predict this
what is involved in the compromise theories combining continuity and non-continuity?
- simple discrimination
discrimination involves both learning what stimulus dimension to attend to and what stimulus values on that dimension are correct
Sutherland and Mackintosh, 1971 - attentional learning is slower to reach aymptote than response learning
- attention to multiple stimuli but increased attention to one dimension means less to another (limited)
predicts overtraining reversal effect and impact of overtraining on the relative ease of intradimensional shift and extradimensional shift
advocates make a good case that multiple factors involved in discrimination learning
BUT weaker than either of the simple theories (continuity or non-continuity) as by putting everything into a theory, makes it harder to falsify it
what are the 2 different types of abstraction? and relation to animals ability to discriminate?
- complex discrimination
- perceptual categories e.g all cats - abstraction = prototype
- logical categories e.g all groups in 4s - abstraction = concept
when perceptual e.g cats, animals able to discriminate between this and an absense/something else whereas when logical e.g discriminating between an artifical group of 4s and 3s, only animal that came close is a parrot
what are the differences of bird visual systems from typical mammalian systems?
birds have:
- cone-rich retinas
- desnse receptor matrix over a wide retinal area
- multiple foveas
- classes of cone differ by oil-droplets filtering light, not visual pigment
- more than 3 types of cone
- spectral brightness response and discrimination
- high flicker fusion frequency
- ectostriatum rather than visual cortex
richer visual representation of the world
name some special features of the pigeon visual system?
- two foveas in each eye, one forward (binocular) and one lateral
- the two visual systems have different functions and psychophysical responses
- very wide range of view
- U/V light detected
- plane of polarisation of light discriminated
so when performing visual experiments on pigeons, they don’t necessarily see things how we do
describe the study involving perceptual categories in pigeons?
- complex discrimination
learned to peck in the presence of a pic of a person and withhold when pic not a person
stimuli (holiday slides) varied greatly in no. of people, posture, clothing etc.
after successful learning, transfer trials show correct response to new photos i.e above average pecking when person in photo (so not merely rote learning of the individual pictures and their pos.neg association with food)
‘higher order concept formation in the pigeon’
and also been shown for ecologically valid concepts for that species e.g conspecifics, prey etc.
what are the 3 theories of category discrimination?
- complex discrimination
- rote learning or absolute discrimination
- claimed to be ruled out by successful transfer to new instances but what if there was also stimulus generalisation - multiple linear feature model
- predicts a super releaser (prototype) effect that doesn’t always ocur
- often hard to demonstrate control by multiple features
- seems that animals may use this when lots to remember so find common features to aid discrimination (those that are unique to the rewarded exemplars gain associative strength) - configural (exemplar) models - seems that animals may use this approach to discrimination when only a few slides shown so remember whole configuration
what is simple discrimination? and how is this different from complex discrimination?
simple - discriminate between 2 things (S+ and S-, which animals shouldn’t respond to as bad result) which are similar and on the same dimension e.g different tones
complex- discriminting between categories, not merely a singular S+ and a singular S-
what is the change in generalisation gradient when discrimination rather than conditioning?
the generalisation gradient is more peaked when 2 stimuli to discriminate between as opposed to being broader when stimulus on its own
further support for Spence’s explanation of peak shift through study by Wills + Mackintosh, 1998?
had artificial stimuli where S+ and S- overlapped by around 50% whereas near N+ and N- were further away from their respective opposites and far don’t overlap at all with opposites
found good peak shift whereby N+ (near S+ but further away from S-) was where the peak occured whereas negative peak shift at N- as further away from S+
what did Wills + Mackintosh, 1999 find about transposition when simultaneous and successive discrimination?
when simultaneous: found shift to near discrimination of S+ (darker shade) over S+ itself, so choosing a darker shade and ratio was .75 and in far discrimination, still choosing, more than 50% of the time, to opt for the even darker one over the previously selected ‘near’ one
when successive: still going for near over S+ (so transposed to new discrimination) but then when far, drops down to 50:50 so not choosing one over the other
what is an example of applying transfer along a continuum to learning in real life?
e.g japanese people find ‘L’ and ‘R’ hard to discriminate between when learning the english language
so play them exaggerated versions of these letters (easy discrimination task) then become better at discriminating in naturalistic task (harder task) due to the theory of transfer along a continuum
early experiment examples of short-term memory in animals?
delay of reinforcement - learning impaired if gap between stimuli/response and reinforcer
discrimination - total impairment when 10 sec delay between stimuli
delayed reaction - delay between stimulus and possibility to respond in discrimination
taste avoidance learning - longer delays than above studies ^^^ are tolerated in these conditions
what are the 3 ways that target memories could be lost?
- proactive interference - from info acquired prior to target (info given prior to target disrupts its attention)
- retroactive intereference - from info acquired after target (new with old) (forgetting of info due to being followed by something distracting)
- decay - due to passage of time
describe the different uses of delayed matching to sample (DMTS)?
- doesn’t necessarily require recognition of identity of sample and comparison stimuli *
- oddity from sample - choose comparison which doesn’t match
- symbolic matching to sample - comparison stimuli are not the same as the sample, and subject must learn ‘code’ connecting them
- multiple samples - test of recognition memory or a list recognition task
evidence for proactivate intereference in delayed matching to sample (DMTS) tasks in animals?
some trials were ordinary DMTS and others where 2 samples are presented either 10 or 0 secs apart
animal must respond on basis of last sample
result that performance in 0 sec condition, when one sample immediately followed by other, is worse
so good evidence for proactive intereference in DMTS as memory of old stimuli proactively interfers with the memory of the new stimuli
