Psychobiology: Learning & Memory, WEEK 1 Flashcards
1
Q
Why is learning + memory important from a neuropsychological perspective?
A
- It helps us predict the future > when you learn things, you will remember this and apply it to future situations > this is called adaptive behaviour and assists survival
2
Q
Learning definition
A
- Different types of learning including behavioural + non-behavioural
- As psychologists, learning is defined by acquisition of behavioural information > general learning = acquisition of ANY information > DISTINCT PROCESS
3
Q
Classical conditioning
A
- An example of unconscious learning which predicts the future > Pavlovian conditioning
- Showing a stimulus before a biologically significant event (like food as it helps survival) = association between the two > next time the situation occurs the animal behaves like they expect food = anticipatory response (anticipate response occurs)
4
Q
Acquiring behavioural information
A
- To be relevant to psychology, we look at acquisition and retention of info which guides behaviour or leads to a change in behaviour (adaptive behaviour)
- E.G: a flashdrive can hold information and retrieve it, but this is not behavioural, it does not change behaviour.
5
Q
Memory definition
A
- Retention of behavioural info
6
Q
Principles of Learning
A
- Memory improves the more we learn, but this doesn’t mean the amount by which it improves is the same each time > progressively the amount you learn will reduce as you retain info from before. > Law of diminishing returns
- Learning + memory is to be adaptively beneficial, this means we cannot encode and retain everything because it is not an efficient process
7
Q
Law of diminishing returns
in relation to principles of learning
A
- The more you learn or the more effort you put into learning generates less additional value as it stays the same when you have no more left to learn. > supported by R-W model
- Creates a characteristic learning curve where memory strength at a point stays the same > amount learnt gets more shallow each time till theres no more to learn
8
Q
Rescorla-Wagner theory of learning
one theory of many
A
- Focuses on the element of surprise in learning > amount learnt depends on the amount of surprise at the outcome > the more surprising, the more we learn.
- Developed to explain Pavlovian learning
9
Q
Rescorla-Wagner rule of learning is described in what equation?
A
Delta V = Alpha Beta (Lambda - V)
- Delta = describes changes V = learning > Delta V = changes in amount learnt
- Alpha + Beta= learning rate parameters (LRP)(speed+amount) > A = LRP for predictive stimulus B = LRP for outcome (eg: we are more likely to learn about a loud noise than a quiet one (A) and about something with high reward than low (B)
- Lambda=maximum amount which can be learnt + V is high in start of learning if it is surprising. > the gap between these two = prediction error > when V gets close to Lambda, less is being learnt because there is less surprise.
10
Q
Prediction error signal (PES)
A
- The difference between Lambda - V = a PES > the more we learn, the more we know which makes the difference smaller. > on a learning curve this would present as V being close to Lambda > eventually Lambda=Lambda as the maximum is learnt.
- A big PES means A LOT is learnt, small PES means A LITTLE is learnt.
- When lambda-V= small, Delta V also becomes small as there is a change in level of learning + vice versa
- When we know the outcome, V=Lambda and Lambda-V=0 as there is no learning left to do. > means regardless of AB, deltaV=0 so PES=0
11
Q
Role of surprise in learning rate parameters
A
- The amount we learn is proportional to the amount we are surprised at the predictive stimulus (A) + the outcome (B)
- More surprise at the LRP for predictive stimulus = more learning
- More surprise at LRP for outcome = more learning
12
Q
Pavlovian blocking as evidence for R-W rule
Not biological evidence due to lack of resources at the time
A
- Supports validity of R-W model > if an outcome is fully predicted, R-W argue that there should be no learning due to a lack of surprise.
-Describes how if there are 2 stimuli which are presented together, then learning of the learning of second stimuli will be blocked.
-EG: Stimuli X is learnt + predicts a particular outcome = learnt to maximum capacity.
Stimuli X and Y (XY) are presented together + leads to an outcome. > Y could be predictive of the outcome as seeing Y = experience of outcome, BUT, Y was presented by X which was already experienced so there was no surprise to learn from (PES=0)
-X blocks learning of Y > supports R-W
-Blocking also leads to adaptive behaviour > eg: allergy to peanuts leads to avoidance of peanuts (peanuts alone or peanuts in other foods (other stimuli like XY)
13
Q
Schulz et Al, 1998: supports prediction error signal
A
- Taught monkeys pavlovian relationships by showing a series of images where certain ones = reward. > focused on ventral tegmental area of brain (VTA) where neurons release dopamine > used activity of dopaminergic neurons to assess PES > high activity = high PES.
- Results > (1) UCS followed by reward = high activity after receiving the reward as they were surprised > high PES
(2) CS followed by R = high activity after seeing CS, this is because they learnt R would follow + predicted it > low PES
(3) CS followed by no R = high activity seeing CS as they predicted R. No R=pause in activity then rises high again due to surprise at outcome > this is a negative error signal as there was an over expectation which needs to be reduced.
14
Q
Berns et Al 2001: evidence for PES in humans
A
- Used FMRI to look at activity in the nucleus accumbens in human brain.
- Created a predictable and unpredictable situation between delivery of water + juice > one sit=W/J/W/J another sit =J/J/W/W/J/W
- Found activity was higher in the unpredictable sit than the predictable as they were learning = high PES
- Nucleus accumbens was consistently active which is where dopaminergic neurons project from VTA
- Shows PES exist in humans but do they regulate memory?
- VTA projecting to nucleus accumbens (NA is part of striatum) can show us PES
15
Q
Does PES regulate memory?
Takahashi et Al (2009)
A
- Studied rats using Pavlovian over-expectation
- Used lights and sounds as predictive stimulus and food as the outcome
- Procedure: 3 auditory stimuli + 1 light shown, only 2 auditory stim + light followed by reward (learning stage). > over-expectation training: an auditory stimulus is shown with light (both individually= reward earlier) > rat expects x2 of R but get 1.
- When tested > results show lower predictive behaviour to the auditory stimuli shown with the light as they didnt get two rewards + higher predictive beh for second auditory stim which got R.
- When VTA is inhibited (reduces activity of dopaminergic neurons), PES is inhibited + shows reduction in prediction for stim 1 from overexpectation training is not there + predictive behaviour is high again.
