6 - Configural Learning Flashcards
define configural theory
strength of response to novel stimulus determined by the similarity btwn it and previously experienced stimulus, multiplied by assoc strength of prev experienced stimuli
define discriminate
using experience to know two stimuli are different
define generalisation
recognising one stimulus is similar to one we have previously experienced so expected a similar outcome
define negative patterning
training animals to expect an outcome in the presence of A, of B, but none w A and B together
how is negative patterning differnt to blocking
animals trained all associations at the same time
what do animals learn in negative patterning
struggle to learn A+B=no outcome as have assoc of A and B leading to outcomes
why if the RW model summative
when learning ab A and B’s assoc strengths, they are added together when presented together
what does the RW model predict, causing issues w the model
animals would expect A and B lead to twice the outcome and wouldn’t expect no outcome
the revised RW model accomdated what and how
negative patterning by including X
when A and B presented together, animals see A+B+X, with X leading to no outcome
in RW, what does X act as
abstract and configural cue entering association that A+B presented as configuration
X is defined as what else apart from a configural cue
conditioned inhibitor so we expect nothing in presence of X
Pearce stated that we see A+B as what
not as a compound but as its own stimulus whereas RW/elemental say A+B is a configuration
Pearce states we respond to compounds how
in the way we respond to encountered stimuli during training, so expect AB to lead to double the outcome as similar to A and B
what does the generalisation equation look at
change in excitatory strength instead of assoc strength
the configural model states that all assoc strength will be given to what if we train that AB leads to a shock
to AB as a compound and none to A or B separately as don’t learn anything ab A or B
the configural model states that in trials of only A or only B, the animal will expect what
a shock since A and B are both 50% similar to AB
(configural) if trained that A and AB lead to a shock, what will happen
learn AB is 50% similar to A so generalise and learn ab AB as a compound as AB only 50% similar plus expect a shock w just B as 50% similar
RW states what happens to assoc strength if train that AB leads to a shock
half given to A and half given to B and predict a shock in trials of just A and B
blocking states what about generalisation
if trained that both A and A+B lead to a shock, won’t expect a shock as haven’t learnt anything ab B
what were the two conditions in the peak shift study
control: taught that at 550nm there is a reward, not exposed to any other wavelengths
experimental: taught 550nm = reward and 560nm = no reward
results from the peak shift study
control responded at 500nm as expected reward
exp responded most at 540nm and further from wavelength they were trained at and far from non-reinforced
repeating the peal shift study in humans showed what
humans have pattern of responding based on how near or far the item is from trained stimulus but only w artificial stimuli as they are novel to us
how is peak shift explained
wavelengths are compound stimuli 560nm is made up of inhibitory 550nm made up of inhib and excitatory 540nm has excit so respond further away from 560/550 as 540 and below has no overlap w inhib
having fewer words and experience for some colours means what
can’t discriminate btwn colours which are objectively obvious to others who have more words and experience of the same colours as separate stimuli
