Memory Systems

Question 1

What are the two main double dissociations covered in this unit?

Answer 1

-declarative vs working memory
-declarative memory vs habit learning

Question 2

How is memory divided?

Answer 2

memory —-> working memory and long term (declarative (explicit) or nondeclarative (implicit))

Question 3

How is declarative memory divided and what areas of the brain are related to it?

Answer 3

declarative memory - facts and events
hippocampus, medial temporal lobe, diencephalon

Question 4

How is nondeclarative memory divided and what areas of the brain are related to it?

Answer 4

skills and habits - striatum and motor cortex and cerebellum
priming - neocortex
basic associative learning – emotional responses (amygdala) and skeletal musculature (cerebellum and hippocampus)
nonassociative learning - reflex pathways

Question 5

What is a loose definition of memory?

Answer 5

how your past experiences influence current behavior

Question 6

What regions of the brain did HM have removes and what did he present with afterward?

Answer 6

medial temporal lobes and hippocampi for epilepsy treatment
-anterograde amnesia - cannot form new memories

Question 7

What types of memories could HM not form?

Answer 7

new declarative memories - cannot learn names or recongize the faces of new people
-could not learn new facts

Question 8

What was HM good at?

Answer 8

normal IQ
able to carry on conversation (working memory)
remembered things before surgery
-had intact short term or working memory or long term nondeclarative memory

Question 9

What was HM’s primary issue?

Answer 9

encoding new declarative memories
-storage and retrieval seemed fine

Question 10

What are the two main double dissociations between memory systems did we talk about?

Answer 10

-declarative long term memory vs working memory
-declarative memory vs nondeclarative memory for skills and habits

Question 11

Why are phone number seven digits long?

Answer 11

the digit span task to measure working memory is essentially learning a new phone number
-found that random digits the order is important
-the magical number seven plus or minus two

Question 12

What deficits did patient KF have?

Answer 12

difficulties with working memory

Question 13

What double dissociation is provided by a comparison between hM and KF?

Answer 13

declarative memory and working memory

Question 14

If KF was given a list of unrelated words how many could he repeat back? How many could HM repeat back?

Answer 14

KF - could repeat back 1 word
HM - could repeat back six words

Question 15

If participants were gievn a ten word list they needed to remember long term how many repetitions did it take KF/controls and how many for HM?

Answer 15

KF/controls - 7 repetitions - repetition takes the place of working memory har because KF cannot internally repeat back cause working memory is compromised
HM - 25 repetitions - could only repeat back 5 words

Question 16

What were the short term memory impairments that KF displayed?

Answer 16

recalls of lists longer than one was poor
-could learn 10 word lists after multiple repetitions and could recalls most of the words
one to two months later
How can you possibly form new declarative memories if you cannot hold more than one or two items in working memory at a time?

Question 17

What was done in the wagner et al study?

Answer 17

-imagine the process of encoding declarative memories
-wagner et al presented participants with words and had them make semantic decisions (is this an abstract concept or concrete thing)
-after the imaging session they were presented with the words they had read and asked if they remembered seeing them or not
-the researchers then compared brain activity during the semantic decision task between items that were remembered and forgotten- trying to turn the information into something that you can store and then retrieve

Question 18

What regions of the brain were more active for words later remembered than the words that were later forgotten in the wagner et al study?

Answer 18

frontal and medial temporal regions were more active for words that were later remembered than words that were later forgotten
-control areas in visual and motor cortices did not show this effect
-these areas are similar to what we see in working memory tasks with written words as stimuli as well
-Visual cortex and motor cotex have to respond when you see a word on a screen and push a button and those areas are the same for remembered or forgotten items and the left fusiform gyrus is analogous for the right FFA and the left FFA is visual word form area involved in processing printed words and can observe that teh remembered words are processed or elicit more activity than the fusiform gyrus - inferior frontal gyrus areas like brooks area and they are involved in these and encoding of longterm memory involved coordinated areas - so more intense processing leads to better memory

Question 19

Good performance on working memory tasks in people with amnesia depends in part on what?

Answer 19

the familiarity of the items
-test with unfamiliar unnameable items reveal impairments after as little as ten seconds of delay
-the sparing of working memory may be specific to lists of familiar items
-patients with working memory deficits have issue with forming long term memories to unfamiliar items
-rapid learning of word-word pairs but very slow learning of word-nonword pairs

Question 20

More intense processing of a stimuli aka a stimuli is harder to interpret leads to better what?

Answer 20

better memory of the stimuli

Question 21

How did HM help show the dissociation between declarative and non declarative learning (especially habit learning)?

Answer 21

HM learned to trace figures in a mirror (harder than it looks) in the same amount of time as healthy controls
-on subsequent days he did not remember having done the task at all but still showed normal learning

Question 22

Since skill learning engages very different circuitry from declarative learning how did the SRRT (serial reaction time task) elucidate this?

Answer 22

-SRTT is used to study simple motor skill learning
-after initial familiarization with random sequences participants learn to tap in a fixed sequence over blocks of trials
-once the pattern is learned a random block is introduced and a slow down in reaction time is observed
-when pattern is reinstated performance immediately reflects the full benefit of skill learning

FOUND activity in the:
-basal ganglia (center of habit learning)
-motor cortex (motor task)
-cerebellum
-but not the hippocampus or medial temporal areas which are the areas involved in declarative learning

Question 23

PD involves the deterioration of what where?

Answer 23

dopaminergic neurons in the basal ganglia

Question 24

What is the specific form of procedural learning the BG is involved in?

Answer 24

reinforcement learning

Question 25

Since the BG is needed for initiation of movement what did we need task wise to test procedural learning in PD patients?

Answer 25

-need a non motor task
-PD patients have tremors and bradykinesia

Question 26

What is the nondeclarative non motor learning task used in PD patients?

Answer 26

weather prediction task - on every trial you try to guess if it will rain or not
-writing the rules for this is way beyond most people capacity
-more than a dozen trial types
-no trial type predicts rain more than 89% of the time
-and yet people do gradually do increase their accuracy over time
-learn to do something or a binary guess and is rain or shine and have patterns which tell you have probability whether it is sign to be rainy or sunny
-some include combinations of the cards like red and green combined is different than just red
-task is designed to frustrate attempts to explicitly remember what is happening so in the same way you are learning a skill of following a gut feeling of whether it is going to be rain or sunshine and can show you will gradually improve over time and cannot explain why

-amnesic patients gradually learn this task just as they gradually learn other skills
-note control patients outperform amnesics early on
-parkinsons pateints do not learn this task well

Question 27

What was the double dissociation between working memory and declarative memory?

Answer 27

working memory - HM and other amnesics generally have normal abilities on things like digit span

KF and rare other patients perform well in long term declarative memory tasks despite extremely poor working memory
although more commonly working memory deficits impact declarative memory

Question 28

What was the double dissociation between habit learning and declarative memory?

Answer 28

habit learning - hm and other amnesics can learn to do things like mirror drawing, tapping in
sequence or the weather prediction task
parkinsons patients impaired at all of these things but have normal declarative memory abilities

the weather prediction task is a fair test of basal ganglia prediction in pd patients because they have primary motor deficits so a lot go habit learning tasks involve complex motor behaviors which they cannot perform - so test what developed to defeat explicit memory approaches - amnesias are good at learning this task but ill come to appreciate the statistical regularities of particular sets of cards or outcomes

Question 29

How did Poldrack et al show how memory systems interact in healthy brains?

Answer 29

-compared performance in the weather prediction task to performance in a paired associate learning task by presenting the answer at the same time as the cards
-set up like in clinical investigation and push a button to make a prediction and get feedback and seeing the outcome of your prediction is important in pattern learning by the basal ganglia -then have a declarative analog of weather prediction see same stimuli told what it corresponds to instead of having to predict what is coming up they just tell you so you can do paired associate learning which amnesiacs and hm are bad at - if you see different parts of the brain active during these tasks can see different memory systems being engaged for these tasks

Question 30

In the paired associate learning task what part of the brain what parts of the memory system were found to be active?

Answer 30

the medial temporal lobe structures associated with declarative memory were active

Question 31

In the weather prediction task what part of the brain and the memory systems were found to be active?

Answer 31

the basal ganglia associated with non declarative learning were active

Question 32

Do the basal ganglia and the MTL compete with each other?

Answer 32

stronger activity in MTL was associated with weaker activity in the basal ganglia

Question 33

What is the trade off between memory systems over the course of the weather prediction task?

Answer 33

-early on the MTL is active and the basal ganglia is deactivated whereas later on the situation is reversed
-this might reflect early attempts to encode predictions as declarative memories
-so the declarative and habit learning systems seem to compete with each other

Question 34

What experiment was done to see if the MTL and basal ganglia always compete?

Answer 34

-wimmer and shohamy explored conditions under which there could be cooperation between the hippocampus and striatum
-their experiment had three phases
-learn associations between photographs s1 and fractal images s2
-learn probabilistic associations with reward (s2 only)
-decide between two images either s1 r s2

-associated stimuli had the same reward contingencies in the last phase so to maximize reward participants should generalize reward contingencies across stimuli
-looks at cooperation between habit learning and declarative memory systems

1. association phase - and see photographs and associate photographs with marble images or unnameable patterns and classic paired associate learning and goes all the way back -
2. there is a reward phase and because you are not being asked to memorize the rewarding ones and non rewarding ones and this is associated with classical conditioning - procedural learning is operant conditioning
3. how did these things mesh together and look at decision bias and red which is reward and blue is not and choose between them and gonna pick the red one cause it is a rewarding fractal - more interesting to look at s1 and - never seen that sunset and been given a dollar associate the sunset with a red pattern and the red pattern is later rewarded - must have generalized the red marble rewards to the sunset and chosen the sunset
   How do we associated the sunset with the reward in this experiment? - paired associated learning with s1 and s2 is declarative learning -n need to activate memory of seeing sunset and reward generalized to that

Question 35

Learning was stronger and more consistent for the S2 stimuli than for the S1 stimuli why?

Answer 35

the s1 stimuli were never directly prepared with the reward
-participants did develop some decision bias for some s1 stimuli however so that it was possible to bin these into high medium and low bias after the fact
-Learning is stronger for s2 and classical conditioning works so we have preference for rewarded s2 - average for s1 people are 50/50 for s2 associated with s2 that has been reinforced no strong preference overall - and there is cluster where one prefer one over other and some are in middle and some prefer the opposite - classified into three bins of high medium and low decision bias and looked at high decision bias - the other analogy is that stimulus is not interesting - when you were in the scanner did you exhibit a preference for that image
-interestingly hippocampus activation during the reinforcement learning for s2 strongly predicted later decision bias for s1
-further correlations between the hippocampus and the basal ganglia predicted decision bias
-the association between s1 and s2 was activated during reinforcement learning participants could generalize
-that is remembering having seen s1 and s2 together earlier in the experiment was important for being able to transfer the value of s2 to s1
-these twp memory systems nteract to generlize a reward to a novel stimulus

Question 36

Does cooperation between memory systems mean they are distinct?

Answer 36

no it does not mean they are not distinct
-hippocampal actiavtion is primarily important for the learned association between s1 and s2
-striatum is important for reward learning
-cooperation between the two is important for generalization of decision bias

Question 37

What are some things that are important to note about memory and decision making?

Answer 37

cooperation between declarative and habit learning systems only seems strange since we are used to treating them as disociable systems
-most decions we make consciously involve some combintation of reward circuitry an declarative memory

Question 38

What is prospective memory?

Answer 38

the ability to imagine events in the future is strongly related to declarative memory

Question 39

Where is priming related to in the brain?

Answer 39

the neocortex or anywhere in the brain
-where you see priming effects in the cortex is going to be evidence that show stimuli that something is going to be happening in that part of the brain

Question 40

What are the key features of priming?

Answer 40

a chnage in efficacy of stimulus processinf arising from prior encounter
-same stimulus and target - direct priming
-related stimulus but not same as target - indirect priming

Question 41

What are the forms of direct priming?

Answer 41

perceptual and conceptual priming

Question 42

What are the forms of indirect priming?

Answer 42

semantic priming

Question 43

Does priming depend on conscious awareness?

Answer 43

no in fact when participants are aware of the prime the data is contaminated by declarative memory

Question 44

How are priming and repetition suppression related?

Answer 44

the sharpening hypothesis - on first exposure to a familiar stimulus widespread activation is evoked
-on subsequent exposures the activation is sharpened such that only the most critcal cells fire
-the reduced number of cells firing results in a decrease in the bold response

Question 45

What is repetition enhancement?

Answer 45

when novel stimuli are presented have ot create new representations so the first representaion is little activity but subsequent will actuvate these new representations more robustly
-this increased activity is due to episodic retrieval - see unfamiliar face - be like I just saw that face five faces ao when see it again

Question 46

What is a semantic network?

Answer 46

words are stored in neural network with related words connected to each other so when you say a word some activation spreads to related words

Question 47

Semantic priming is associated with response suppression in what part of the brain?

Answer 47

anterior temporal cortex - this part of the brain that modulates relationships between words

Question 48

What is semantic demantia?

Answer 48

and have early on symptoms of word finding difficulties and accessing most vocabulary or tell u meaning of things or doing non verbal tasks
-deterioration of nervous tissues in anterior temporal cortex

Question 49

Why is priming thought of as a memory process rather than a memory system?

Answer 49

happens through the cerebral cortex
-visual priming in areas associated with visual processing and semantic priming in areas associated with semantic processing
-we can use repetition suppression to study how things are represented in different parts of the brain