Cognitive, Models of Memory, WEEK 7 Flashcards
1
Q
Ebbinghaus
A
- first real memory researcher + researched his own memory
- Studied memory using (consonant vowel) nonsense syllables to exclude prior knowledge such as PAB, SER, NID > presented with 20 of these syllables using a metronome timing how long they are presented for (one run through the list > single repetition)
- Then would test himself by giving himself the first syllable of the list then test how many other syllables he remembered
- To keep it scientific he tested himself everyday at the same time of day (standardised) > results showed insight into learning (law of repetition) and forgetting (savings method) > forgetting is the opposite of what we see for the law of repetition (high retention to forgetting curve?)
2
Q
Law of repetition
A
- Found that the more you rehearse, the better your retention > at the start performance was quite high (after first and second repetition) but the more that he repeated it, there was less new information because he had learnt it all
- The levelling off of the curve shows that no new data was learnt + memory retention was high because they learnt the info
3
Q
Phases of Memory
A
- Acquisition/Encoding: how new information is manipulated + placed in memory.
- Storage: how and where it is held in memory ( aka “memory trace”) > probably 2 steps: temporary memory trace in hippocampus, then integration in cortical information networks.
- Retrieval: “remembering”, memory brought back into active use. Once we do this, memory becomes malleable again > when we store it again the new memory will be a bit different to the original one
4
Q
Atkinson & Shiffrin Model of Memory Storage (multi-modal model)
A
- Argue there are three types of memory > sensory memory, short term memory + long term memory > these are discrete processing stages
- Sensory input goes to sensory memory, if we pay attention to it, it transfers to the STM where it can be rehearsed. Rehearsed info then gets stored in the LTM
- Unattended info in sensory will get lost, unrehearsed info in STM will get lost and some info in LTM could decay over time
5
Q
Sensory store
A
- At any given moment, you are being bombarded with sensory information (visual, auditory, smell, taste etc..) > this is info we cannot remember for our whole life
- The sensory store can fill in blanks when there is regular stimulation
- Sensory memory somehow takes a snapshot of what is happening in the world which you use to attend to what you think is important (e.g. if you ask someone what did they say but then you can recall it before they say it again because it was still in your sensory store)
- Or when you have a sparkler and turn very quickly it makes a circle but a picture would not see this you would just see one point of the sparkler > if you stop moving around the trail will disappear very quickly so we know the store is time limited
6
Q
Sensory store function
Evidence on capacity and duration: George Sperling (1960)
A
- keep sensory information in mind so we can attend to it > this is modality-specific
- Classic experiment on visual sensory store
- Sperling presented 12 letters (3 by 4 matrix) for 50ms + ppts name as many letters as possible (full report condition) > people could name around 4 letters > doesn’t mean the capacity of the sensory store is 4 letters because ppts can sense that there were more letters
- Takes time to read out things from the store and information fades out with time so while you are reading out, the information starts to fade > doesn’t mean the info was not there in the sensory memory to begin with
- In the partial report condition, ppts would fixate on a cross then shown the letters, a tone would be chimed either before the display (100ms before) or up to one second after the display ended > the tone would be high, medium or low, ppts were trained so the tone indicated what row to report (high=top row, medium = middle row, low = bottom) > max that can be remembered by individual ppts is 4 because 4 on each row
- To understand capacity we need to do a lot of trials + take average performance for each row so that we have maximum of 12 remembered
- Results showed if the tone was made at the same time as presentation of letters, over 9 items could be remembered, after 100ms this drops to 7 and continuing on > longer the delay between presentation of image and tone, the worse the performance
- Can conclude that a lot can be held in the iconic memory (perhaps anything entering the visual system) > as for duration, performance in the partial report got worse + more similar to full report condition between 250ms-500ms > means info in iconic memory decays very quickly (half a second) but this depend on the kind of stimuli
- E.G. if you give people a natural scene, this could be remembered for a second or using auditory stimuli in experiments show retention for around 3 seconds
- Any info left in sensory store transferred to STM before it was lost
- All letters were held in sensory memory for short time but decayed rapidly. Tone acts as a spatial attention cue. Attention > STM
7
Q
Modalities of sensory store
A
- iconic memory for vision
- echoic memory for hearing
- haptic memory for touch
- olfactory memory for smell
- gustatory memory for taste
8
Q
Short-term memory
A
- Function of STM is to process information consciously > attention is important here because it limits what information will be attended to and processed further
- From the sensory register, the brain encodes information > “Once we receive sensory information from the environment, our brains label or code it. We organize the information with other similar information and connect new concepts to existing concepts” > things can be encoded visually, acoustically or semantically (e.g. reading a text using visual encoding so reading letters or acoustic where you remember the sounds of the words or semantic where you remember the meaning)
- Even when you read silently, this is acoustically encoded > research shows if you show a string of letters or words and then recall them, people tend to make acoustic errors (e.g. confusing can with man as opposed to can with tire or mixing B with E)
- Information from our sensory store needs to get encoded to be stored in STM
9
Q
Capacity of STM
A
- Miller, 1956: Magic number 7 +/- 2 argues that we have a very limited capacity in STM > we typically remember 7+2 or 7- 2 items when shown letters > seen as an information bottle neck + Cowan (2001) argues it is probably closer to 4
- This doesn’t mean we can only remember 7 items, we can overcome this by using chunking where we group similar stimuli together
- Chunking depends on the knowledge we activate from our LTM + expertise you have in your ability of finding patterns > e.g. in an experiment with chess players, if they were shown a picture of a chess board with chess pieces from a game then expert chess players could remember where the pieces were better than novice chess players. If the chess pieces were randomly places, experts were no better than novice players
- CHUNK: Meaningful unit of information
- Chunking example here shows 3 chunks of information even though there are 12 letters but it is easy to remember because it relates to our prior semantic knowledge
10
Q
Duration of STM
A
- STM is limited in duration > if we do not rehearse information we will lose this info within 15-20 seconds (depends on kind of information)
- The longer the delay between presentation and recall, the more that is forgotten > decay
- Decay is when you lose information due to time passing. Displacement is when you get a new information and replace old information with this (not much evidence for displacement)
- Rehearsal = the process of repetitively verbalizing or thinking about the information
11
Q
Criticisms of STM
A
- Decay is not the only reason for forgetting. Interference is a more likely contributor to forgetting than decay
- Proactive interference: where something you learnt earlier interferes with something you learnt later
- Retroactive interference: where something you learnt later interferes with something you learnt earlier
- E.G. for PI, James dated Hannah first and this earlier knowledge may interfere with remembering his current partners name so calling his wife (Hannah) Jade.
- E.G. for RI, calling his old girlfriend Hannah his current wife’s name because his new info is interfering with the old info
12
Q
Release from Proactive Interference
A
- This happens when for example you do trials where you have to look at some letters, do a distractor task and then recall the letters (e.g. shown letters then have to count back 2 numbers from a given number then recall)
- The first time you do this performance will be good, but the more you do it the worse you get > this is because there is proactive interference (earlier info inhibits performance for later info)
- But if you change the kind of stimuli being used (e.g. swap from remembering letters to numbers) then there is a “release from PI” where you become accurate again
13
Q
Other criticism of STM component of multi-store model
A
- Info in the STM does not need to consciously be processed > e.g. implicit learning (learning things without directly trying to)
-Simple rehearsal does not ensure LTM storage > Rehearsed info does not necessarily transfer to LTM > Information in LTM has not necessarily been rehearsed during learning - Learning in STM affected by LTM (e.g. info from LTM can be helpful for chunking)
- Double dissociation found in patients supports STM – LTM distinction > comes from patient data, some patients with severely impaired STM but intact LTM vice versa
But some patients with severely impaired STM can still acquire LTM memories > contradicts MSM because it suggests things do not need to be rehearsed in STM to get to the LTM
Solution: WMM > New model developed to address some criticisms of multi-store model
14
Q
Long-term memory
A
- Once information passes from sensory to short-term memory, it can be encoded into long-term memory according to MSM
- Function of LTM is to organise and store information using different system like explicit/declarative and implicit/non-declarative
Capacity is assumed to be unlimited (don’t know if anyone has run out of LTM storage) and duration is expected to be permanent (although it cannot always be retrieved)
15
Q
Working Memory
A
- Proposed by Baddeley and Hitch (1974) > WMM provides depth of STM which the MSM lacks
- They argue that we use STM not just for retention of info but used to perform more complex tasks (e.g. storing information about outcome of earlier processes) > STM is essential for the performance of numerous tasks (which aren’t explicit to memory) including understanding language, mental arithmetic, problem solving, reasoning etc..
- Essentially WMM says STM is not a passive store + is responsible for different things
- Argues STM is multi-component: short-term memory store comprises of 4 working memory components: Central executive (supervisory store), Phonological loop & visuo-spatial sketchpad (slave systems) and later added episodic buffer
- Visuo-spatial sketchpad: if you are asked how many windows are there in your house, you would likely create an image in your mind of how your house and windows look > this is part of your visuo-spatial sketchpad. When you count each window, this involves your phonological loop
You would not randomly start counting your windows, you are likely to use a strategy such as from the front of your house to the back of your house > this requires your central executive
