Rescorla Wagner Flashcards
what is rescorla wagner theory of classical conditioning
a formula used to mathematically predict when learning will occur
uses latent parameters (influences on behaviour that we observe but dont have access to) to measure influence on behaviour
determine strength of association
how an animal learns the associative value between cs and us
what is the formula?
delta x v = alpha beta (lamda - v)
what does v stand for
the associative strength depicting the amount of learning
v = v(old value) + delta v
what does alpha stand for
relatve salience of cs between 0 and 1
0 = no attention
1 = max attention
what does beta stand for
relative salience of us between 0 and 1
0 = greater rate of extinction
1 = greater rate of conditioning
what does lamda stand for
max associative value that can be conditioned
reflects strength of ur elicited by us and time lag between cs and us
set to 1 when us present
set to 0 when us absent
what does delta v stand for
‘us suprisingness’
represents the changing associative strength between cs and us
how much learning occurs - change in associative value due to variability
how much you learn is dependent by how suprised you are
what does alpha and beta stand for
the natural relationship between cs and us
ie more likely to associate food with smell than light
what does (lamda - v) stand for
difference between max associative value and actual value
distance from asymptote
what does the formula mean
the change in associative strength of the cs during a trial (delta v) will depend directly (=) on the satience of the cs (a) and the strength of the us (b), and the difference between the max associative value and the true value (lamda - v)
how does the formula explain blocking
one element of cs already fully conditioned
combine with unconditioned
cs = 1 (fully conditioned, no suprise)
ie tone (cs) and shock (us) = fear
tone(cs) = fear
tone(cs) and light(ns) and shock (us) = fear
light and shock = no fear
cs and ns do not change predicitive value as completely predict us
what are the main points that the model proposes
Learning will occur if what happens on the
trial does not match the expectation of the organism
expectation on any given trial is based
on the predictive value of all of the stimuli
present.
how does the formula explain unblocking
when the novel stimulus is introduced
and the US is larger there is surprise,
because the CS predicts a smaller us than what occurs
how does the formula explain extinction
assume cs already conditioned - initially at 100% (1)
max value for lamda to approach will be 0
models problem with extinction
model cant predict spontaneous recovery (where the association reforms after it had previously diminished)
