Sensory and Short-Term Memory Flashcards
How do we call vision-specific memory? Why?
Iconic memory - since the mental representation we create is called an “icon”
How was iconic memory tested? What were the findings?
Sperling created display of letters in rows (3 x 4 matrix) and presented them for 50 ms
- Control condition: report as much
=> Reported about 4 or 5
- Experimental condition: p. given indication via a tone (high, medium, low) which row should they report
- Experimenter varied the time at which tone was presented
=> If tone briefly after matrix disappeared p. near perfect performance
=> If longer less
Conclusion: Iconic memory has a very high span BUT decays fast
What is anorthoscopic perception? How can we demonstrate it? What’s the purpose?
= the seeing-more-than-is-there phenomenon
- If we’re passing a picture behind a slit at normal speed = p. report seeing more of it than they actually do
- From evolutionary perspective it’s helpful to use iconic memory to quickly integrate info and figure out e.g. whether there is a predator behind the bush
What is the trans-saccadic memory? What’s the mechanism?
= aspect of iconic memory that enables integration of separate saccades to build a cohesive image
- Mechanism relies on Object Files = representations of objects created by neuronal assemblies - aka it keeps track of basic characteristics of objects around us
- Similar machinery as with mental rotations - explains why it takes more time to rotate an object and also move eyes
How can we study echoic memory? What do we know from such studies?
- People were played 3 lists of recordings each with 3 digits
- Presented to one ear, 2. Second ear, 3. Both ears (middle)
- Control: report all
- Experimental: only indicated list
- Findings:
- Same as Sperling
- BUT echoic lasted for longer than iconic (~ 4s)
- Due to sounds being there usualy once (while for image we can make a saccade), doesn’t have to be quickly exchanged for new info
How do we call the memory of touch? What does it include? How about its span?
= haptic sensory memory
- accounts for pressure, temperature, location (while some body parts more/less sensitive), amount of space covered
- similar to previous = large span but quick decay
What is the supposed span of STM? Is it really? How can we easily impove it?
- The most cited is 7 +/- 2 chunks (units)
- BUT some studies also found 4 +/- 2 chunks
- We can use Chunking = method in which we group smaller information into meaningful chunks (e.g. remembering words intead of letters)
- We can build our “chunking ability” with expertise -> the more we know the easier to find meaningful patterns
How does expertise influence STM?
Example: Researchers tested a person for over a year -> He went from 7 +/- digit span to repeating 80 digits
- Possible only because he was a runner and each 4 digits represented e.g. running time
- His STM didn’t improve for letters
=> expertise allows us to get LGM to help STM in creating meaningful chunks
What is the primary issue regarding forgetting of STM?
Does it occur as a result of decay OR interference?
- If decay then it has to be connected to the passing of time
- If interference it must be that any new info in STM is competing with the older one making it difficult to retrieve
Explain the research on “decay” paradigm.
P. given 3 letters string (TPZ) and then, to ensure they wouldn’t rehearse it, given 3 digits (274) and asked to subtract from them (e.g. -7).
- Experimenters varied time between presenting and reporting of letters
=> after 18 s all forgotten
How about research on “interference” paradigm?
They did the same thing as with “decay” but used only 3 trials - believed that what caused the forgetting was interference of previous letters put into STM
=> indeed the first trials showed no forgetting (this only rose with subsequent trials)
Recall the other test on interference theory.
P. given 16 digits -> at the end presented with a probe (indicated by a tone) -> now they have to recall what digits followed the “probe digit” in the previous sequence
- gave control to how much interference was going on
=> Greater interference with earlier probe reardless of the digits being presented as slow or fats (to control for decay)
Explain the Serial Position curve. What 2 effects contribute to it - mechanism, brain areas?
= an U-shaped function in which memory is better for either start or end of set rather than the middle portion
- Primacy effect = better m. for first items
- items that come earlier have more opportunity to get rehearsed and stored in LTM (there are no other items taking away the consolidating resources)
- The slower the presentation of items (more time) the larger P.e.
- Involves LTM areas e.g. left hippocampus, left temporal lobe
- items that come earlier have more opportunity to get rehearsed and stored in LTM (there are no other items taking away the consolidating resources)
- Recency effect = better m. for the last items
- Due to STM, have not been replaced by other items yet
- Areas: parietal lobe
In what scenerios may serial position curve NOT work?
- With memory of performed actions, odors
- Recency effect = yes
- BUT primacy effect = no, as the person devotes full attention to the new action they cannot rehearse the previous (doesn’t get to LTM)
- Suffix effect = recency effect is diminished when extra info is presented at the end (e.g. experimenter saying “go” after STM task)
- The closer the suffix to the target stimulus -> the larger it gets
What do chaining models mean for serial order memory?
- Chaining models = STM includes a series of associative links which could be retrieved when moving along this chain
- BUT that would mean that forgetting one item would break it = not the case