Flashcards
PSYC 318- Lecture #6
Cards in this deck (20)
- 1SKinner's Radical BehaviourismBehaviour is a consequence of environmental histories of reinforcement.
Problem: animals pursue goal's that haven't been reinforced before. - 2Test#1:step 1: put animal in a cage w. two levers, one releases juice, another salt (every time they press the juice lever)
step 2: remove all salt from diet.
step 3: put them in the chamber, make sure neither lever works. The mouse presses the salt water lever, based on how he is feeling now & not on previous histories of reinforcement (proof that he has re-evaluated the value of salt water). - 3episodic vs. semanticepisodic = specific events of the day that is stored in cognitive maps
semantic = we can think about things with general world knowledge - 4Testing Cognitive MapsWhen animals are in a new situation, they assess what would be the best behaviour in this situation.
step 1: put a rat in a chamber with a lot of rooms let home explore.
step 2: lock rat in one room w/ food
step 3: put the rat back in the corridors, he will know where to go to get back to the location of his food (i.e. latent learning)
Rats who take longer to think through each decision, learn faster than rats who rush through the maze. - 5Dichotomies used to describe behaviourstimulus response vs. cognitive map
habitual vs. goal-directed
retrospective vs. prospective
cached values vs. inferred values
model free vs. model based
ego-centric vs. allo-centric - 6goal directed vs. habitual behaviourhabitual: were generally successful in the past (w/out thinking about the goal)
goal: think about a goal that you have thought about being valuable at this very moment - 7Where in the brain, is more than 5% of dopamine needed to learn and maintain pavlovian conditioned responses?Use: CAV (affects axon terminals)
Method: inject virus into axon terminals
it will drive cre expression in all neurones, that project to either the ventral or dorsal striatum
mice born dopamine deficient: dopamine will only be expressed in neurones that express cre
ability to make cre will be in all the neurones that project project from the VTA to the dorsal striatum (red).
A lot of motor cortex and glutamate inputs project here (but effect is insignificant b/c they don't have LoxP sites and can't make cre.
vrDD=DS: Conclusion = 40% dopamine restored to DS and 5% VS (this is because one neurones has widespread axons) but this is not important b/c 5% isn't enough to restore
pavlovian learning. Behaviour = restores learning.
vrDD-VS: Conclusion = 40% dopamine restored to VS and 5% DS. Behaviour = inc. amount of time animals checked the food port, & slightly inc. learning. - 8Test#2: What axons/cell bodies of dopamine neurones are located in the forebrain?Method: Sham Animal + immunohistochemistry
Antibody stains dopamine axons red in the normal animal. This is all in the forebrain basal ganglia region (therefore there are no cell bodies, those are all located in the VTA). - 9Q3: Stimulus-Stimulus Learning occurs via the hippocampus so how is dopamine involved?The hippocampal system, without dopamine, will learn about anything randomly.
Dopamine, directs learning, to pay attention to certain features.
Therefore, restoring dopamine, doesn't change the ability to learn about the world. - 10Paper #6: Decision Makingtransitioned uncertainties: when you chose to go down a certain path, what is the chance you will end up there (i.e. road is closed)
state uncertainties: uncertainty that the reward would have value you thought it would (i.e. the bakery is closed) - 11Paper#6: What are gut feelings based on?Dopamine neurones: encode a temporal difference prediction error.
If dopamine signals say that certain decisions were previously punished, those pathways will become weaker over time, and won't get through the action selection filter.
Yours actions rely on what the dopamine signals have encoded.
The dopamine signal compares predictions from one moment to the next.
The dopamine signal just caries the average value of what happened in that state (not the specifics) this is what is meant by "The dopamine signal collapses, the decision tree and creates cached, values". - 12Experiment#1An animal does a maze 10 times while hungry. If you make the animal thirsty, while doing the maze, they intuitively go left (where the food is), b/c of the dopamine signal and be disappointed (b/c no water). Actually, usually animals find the water, decisions that's haven't been reinforced in the past.
The dopamine signal encodes look up tables of values, if you are hungry turn left (value =4) & if you are thirsty turn right (value =4). Therefore, the brain is constantly storing context dependent values of things. Problem: every moment is unique, so animals can't store an endless look up table. - 13Variable Interval Schedule of Reinforcementstep 1: the lever become active every 30 seconds (the animal presses it every 30 seconds)
step 2: interval change (30, 60, 90) the animal keeps pressing the lever, but not always expecting a reward
Conclusion: animal is expecting both a reward or a punishment, so neither one is surprising.
Purpose: to stop using the cognitive map of the world - 14Drug addictsAct habitually instead of goal directed, even though they are inventive in how they get their drugs.
Habitual = decision to get drugs
Overtime I decided to get drugs the cached value was positive. - 15lesion the dorsomedial striatumland
- 16Intra-limbicasldj
- 17How to demonstrate Habitual Behaviour (rather than thinking)overtraing rats: press a lever for food when hungry
de-value the reinforcer: make a lot of delicious food available
normal rats (who have not been over-trained) will not press the lever
overtrained rats will continue pressing the lever after the reward has been de-valuated (i.e. acting based on habit, rather than value of the reward)
contingency degradation: making the food drop independent of the lever presses: how long does it take the rat to stop pressing the lever? (amount of time it takes from them to stop reflects the degree to which their behaviour has become habitual). - 18How to demonstrate Habitual Behaviour (rather than thinking) Experiment continuedBig Picture: Lesion brain to promote one type of behaviour over another.
Striatum innervated by glutamate from motor cortex, and dopamine from the midbrain.
Lesions in the dorsolateral striatum: prevent habits from forming (i.e. animals are sensitive to the reward despite over-training)
Lesions in the dorsomedial striatum is involved in goal-directed behaviour. (i.e. animals are likely to form habits).
Most of the glutamate inputs from the dorsolateral striatum come from the motor cortex (which relies on habitual behaviours) - 19Lesioning regions of the PFC in ratspre limbic cortex: or OFC in rats = early emergence of habitual behaviours (i.e. they are involved in goal directed behaviour)
Lesions of the infralimbic cortex prevents habitual behaviour from developing (i.e. it's involved in habitual behaviour) - 20infra limbic cortex and reward devaluationstep 1: train the animal- tone 1, says food is available of the right, tone 2 says food is available on the left
step 2: devaluation stage: give the animal a lot of food (correlated with tone 1 i.e. fruit loops), then given him a lithium chloride injection
step 3: next day a normal rat doesn't want any fruit-loops; but over-trained animal still turns right to get fruit loops
step 4: optogenetics used to inhibit neurones in the PFC while the animal is running down the track (habitual neurones) making their decisions goal oriented, therefore stops running towards the fruit loops, and goes towards the path where there is no food available
Information
- Description
- Psychology Flashcards on PSYC 318- Lecture #6 , created by Pascale Bockelmann on 29/01/2017.
- Created by
- Pascale Bockelmann · Jan 2017
- Updated
- Jan 29, 2017
- Views
- 0
- Rating