Cognition 1-4 Flashcards
Paivio (1969; 1971) explained the fact that imageable words are more memorable by suggesting that having numerous retrieval routes improves the odds of successful recall, an idea called the ___________ hypothesis
dual coding
Jenkins and Russell (1952) observed that _____ are more likely to be grouped together in a free recall.
Semantically related words
Neuropsychological evidence for episodic vs semantic memory
Spiers, Maguire, and Burgess (2001) -> 147 cases of amnesia
Dound substantial oe even dramatic loss of episodic memory.
- semantic effects were more variable and generally smaller.
- Damage to the hippocampus (and the MTL) affects episodic memory far more than semantic)
damage to which areas of the brain cause semantic deficits?
semantic = Anterior frontal lobe + Anterior temporal lobe
damage to which areas of the brain cause episodic deficit’s?
Amygdala, Hippocampus, Parahippocampal cortex etc
who studied schemas and how
Barlett, recall of efforful and meaningful stimuli.
what did Barlett find in his schema study “The war of the Ghosts” using Native American folk tales.
when participants recalled the tales, they committed many distortions and errors.
They aimed to make the story more coherent by omitting details.
- Distortions were more consistent with their semantic knowledge.
- Recalled stories were “Westernised”
What did Bransford and Johnson (1972) find about schemas and titles
participants who read a paragraph with slightly ambiguous meaning showed better recall when supplied with a title than not.
suggests that previous schematic knowledge is beneficial for later recall
helps organisation of elements and helps comprehension
Levels of processing theory (Craik & Lockhart, 1972)
lowest to highest facilitator of LTM
Meaning facilitates LTM.
lowest to highest facilitator of LTM
Visual (structure)
Phonological (Acoustic)
Semantic (Meaning)
Evidence for the levels of processing
Words studied and participants asked to make 3 judgments:
Visual processing (e.g. “Is TABLE in upper case?” Y/N)
Phonological (e.g. “Does DOG rhyme with LOG?” Y/N)
Semantic (e.g. “Does FIELD fit in the sentence: ‘The horse lived
in a ___.”Y/N)
found sentance better remembered than rhyme and case. Rhyme better remembered than Case.
Also, YES answers were better remembered, possibly because they are better integrated with existing meaning.
Even when reaction time is matched (some hypothesised that semantic is better remembered as more time is taken to process).
Limitations of levels of processesing
- difficult to define and measure
- processing speed ?
- Levels of processing (features) are not processed in a serial order but simultaneously (e.g. we observe structure and meaning in parallel)
(-) Deeper processing is not always more memorable!
TAP - transfer appropriate processing
cues at encoding should match those at retrieval
TAP evidence
(transfer appropriate processing)
rhyming recognition TESTS better for phonological rhyming encoding
Elaborative vs Maintenance rehearsal. Which is bette
Elaborative. e.g. taking the info in the way that you encoded it, and applying different methods such as connecting to other concepts.
ORGANISATION!
Intention to learn vs organisation
importance in memory
Study
Task
Deck of cards with word on each
4 conditions:
1. Learn the words
2. sort by meaning
3. sort by meaning, you will be tested
4. sort into collums
Found:
2 and 3 showed highest recall.
sorting by meaning with or without knowlege of test produced similar recall.
- worst recall in 4
Conclude: Attention to material and organise is more important.
- intention has minimal effect, while level/type of processing matters more.
however, some argue that if you intend to learn, you will probably organise anyway.
Distributed Practice
Distribute learning trials sparsely across a period
of time
§ Faster improvement rates of learning and less
forgetting
Lag effect (memory)
benefit of repeated study increases as the lag between study
occasions increases
The Testing Effect shows that
having to retrieve the answer, rather than
being presented with, leads to greater retention.
Expanding Retrieval Method
Landauer & Bjork (1978)
Spaced/Distributed Practice + Testing/Retrieval Practice
Ø Motivation to learn may make learning more efficient in
both automatic and strategic ways
How?
Automatic:
Ø External (e.g., reward) or internal (e.g., curiosity) motives
prior to exposure to stimuli improves memory even when time spent studying or strategies used are controlled
Strategic:
Ø People use deeper and more elaborate memorization strategies for high value items
Collins and Quillian (1969) devised which model of semantic memory?
Hierarchical network model
what is the Hierarchical network model
+ cognitive economy
semantic memory organised into a series of hierarchical networks
- Major concepts are represented as nodes
- Properties/features are associated with each concepts.
eg.
top of hierarchy has animal, which has defining features. Node below would be Bird which has features such as has wings can fly. Canary at bottom of hierarchy, with features such as yellow and can sing.
Cognitive economy: the idea is that features are stored at the highest level possible and are not repeated to remove redundancy
Main empirical framework to test hierarcical Network model of semantic memory
Sentence varification task.
shown a sentence, asked if true or false. RT is measured.
greater distances from the node are associated with longer RTs. for both categories and properties.
e.g. A canary can sing has a faster response time than a canary can fly.
problems with hierarchical model
doesn’t acknowledge Typicality (e.g. canary is more representative of a bird than a penguin) or Familiarity (e.g. canary has skin is not familiar)
what model of organisation of semantic memory did Collins and Loftus develop (1975)
Spreading activation model
spreading activation model
semantic memory is organised by semantic relatedness/distance.
Lengths of links indicates the degree of semantic relatedness.
- activity at one node causes activation at other nodes via links
- spreading activation decreases as it gets further away from the original point of action
Semantic priming tasks
when presenting one stimulus that is semantically related makes subsequent processing more efficient
semantic links and distance determine
strength and speed of activation spread from one concept to the other
Deese-Roediger-McDermott paradigm.
study items that have one concept in common. e.g. Doctor.
more likely to report the related item as studied even when it was not studied.
pros and cons of the flexibility of the spreading activation model
pros: reflects more empirical findings
cons:
- makes it less specific in predictions.
- more difficult to test
limitations of spreading activation model
- notion that each concept is represented by a node is oversimplified . e.g. what about abstract concepts such as justice
- does each concept have a fixed mental representation?
- situation/context in which we encounter concepts changes the way we process them.
- do different people have similar representations of any given concept?
- no consensus on how to **MEASURE ** semantic distance
situated simulation theory
- concepts are processed in diff settings
- processing is influenced by current context/setting
- Concepts incorporate Perceptual and motor/action related properties (goal related properties)
e.g. concept of bicycle reflects current goals.
e.g. aim to buy bike for daughter, think ab diff bike.
neural evidence for situated simulation theory
similar areas that control foot, arm, face movement are activated when are asked to perceive an action that is possible with the corresponding limb.
words such as kick, lick, pick activate areas in motor cortex.
RT evidence for situated simulation theory
when asked to annwer Q about an action requiring hands with their hands, the RT was faster than the inconsistant limb being used.
limitations of situated simulation theory
new theory so empirical research is still coming through
BUT these questions remain unanswered:
How cariable are concepts across situations?
do concepts consist of a stable core + context dependant elements.
- are these properties secondary - after concept meaning has been accessed, then contextual version
Grandmother cell hypothesis
- semantic memories are represented in the brain as whole objects
- each object/concept has its own node/neuron.
e.g. there’s a special neuron representing your grandmother. - types of nodes are grouped together
- Most evidence suggests that this is not the case
Feature-Based approach.
Diff kind of info ab given object stored in seprerate brain regions.
e.g.
- Visual in one part, auditory in another
becoming increasingly popular
Neuro-cognitive Hub-and-spoke model.
- hybrid model of semantic memory
- Hub = Modality, independent conceptual representations. (the meaning)
- Spoke = Modaliy specific brain areas. sensory and motor processing (sound, action, colour, shape)
Ecidence for Hub and SPoke model.
tDCS (transCranial direct current stimulation) makes certain brain regions more likely to fire.
asked 2 questions. Tool function, (scissors used for cutting)
and tool manipulation question (pliers gripped by handles).
applied tDCS to ATL (hub semantis) showed increased performance in both tool manioulation and function)
applied tDCS to IPL (actions to objects / spoke) showed tool manipulation task was enhanced
Neuropsychological evidence for hub and spoke model
those with sematic dementia (extensive damage to and starting from ATL.
Show general semantic deficits.
Those with Category-specific deficits have greater difficulty naming or identifying whether object is living or non living for example
Limitations of Hub and Spoke model
open questions
- role of ATL (anterior temporal lobe) may be more complex e.g. organising spokes.
- Does familiarity with concepts affect their organisation in the hub
- How many spokes ? can be categorised?
- How is info integrated between spokes and hub ?
spreading activation model in memory retrieval
Activation level =
the internal state of a memory, reflecting its level of excitement.
- determines the accessability of the item
Increases when:
- smth related to the memory is encountered
- persists for some time
SO spreading activation =
the automatic transmission of “energy” from one memory to related items via associations.
- proportional to the strength of connections
what is pattern completion in spreading activation model
regarded as what mechanism
Hippocampal mechanism
process by which reinstatement of a memory, is caused by spreading activation from a set of cues.
e.g. want to remember an event. at time of encoding, the hippocampus connects the name, face, lecture theatre together. If some of these partial cues are present, one is more likely to remember entire event (and pattern completion has occoured).
hippocampal mechanism
Factors determining retrieval success
- Attention to cues
- Relevance of cues
- Cue-target strength
- Number of cues
- Target strength
- Retrieval strategy
- Retrieval Mode
Attention to cues (as a retrieval factor)
Reduced attention to cue impairs its ability to guide retrieval
observed in divided attention tasks.
e.g. during retrieval, it reduces memory performance if secondary task is:
- related to primary task
- Demands a lot of attention
Relevance of cues
- cues most effective when strongly related to target.
encoding specificity principle
more useful if:
- present at encoding
- encoded w the target
- similar to original cue
best cues are ones present at encoding
e.g. using hair colour to remember a name
cue - target associative strenght
determined by length of time and attention spent on encoding the relationship
e.g. encoding the target and the cue seperately is unhelpful
e.g. using lecture theatre cue to remember name
number of cues
access to additional relevant cues facilitates retrieval, providing both cues are attended to
= Dual cuing.
- Elaborative encoding maximises the number of retrieval routes.
- strength of target memory
- weakly encoded targets are more difficult to achieve
- targets start at a lower activation level.
- require a greater boost in activation to achieve
Explains word frequency effect on recall.
more frequent target words start with higher activation level and are more easily retrieved.
e.g. the name you are trying to remember is a long one.
Retrieval strategy
success increased by:
- The organisation of materials at encoding
- Adopting efficient strategies of memory search
also, Adopting a new perspective/strategy can facilitate recall of different objects previously forgotten.
strategy used in recall should be consistant with strategy used in encoding
Retrieval mode
Frame of mind allows interpreting environmental stimuli as episodic memory cues to guide subsequent retrieval.
e.g.
having multiple episodic tasks in a row graduallly imroves performance
- takes time to fully adopt retrieval mode
- BUT there is also involuntary episodic retrieval (smth springs to mind)
harry potter example recognition task
asked to identify harry potter characters names from a list of fictional characters names.
explicit memory tests
ask:
require which cues
amnesics
which brain region
- ask pp to recall particular experiences
- Require a contextual cue
- reveal impaired performance in amnesics
- in many cases rely on hippocampus
implicit memory tests
- measures unconscious influence of experience without asking the subject to recall the past.
- Priming: recent experience with the stimulus improves performance
- Reveal normal performance in amnesics
types of explicit memory tasks
- free recall
- cued recall
- Yes/No recognition (was the word MONK studied)
- Forced-Choice recognition (which word did you study? Apple or Monk?)
Implicit memory retrieval tasks
- Stem completion. “fill in missing letters MO_ _ “
- Fragment completion “Fill in A _ P _ E”
- Conceptual fluency “name as many birds”
context dependant memory
- memory is context dependant
- context reinstates original encoding environment and facilitates retrieval.
Principle of encoding specificity: we encode info along with its context
diff types of contextual cues
External
- Spatio-Temporal / environmental
Internal
- Mood
- Physiological (tired, drunk)
- Cognitive: collection of concepts one has though ab the event
Mood-Congruent memory
about the person/item match.
- easier to recall events that have an emotional tone that matches the current mood of the person
- depressed indivs are more likely to recall unpleasant events
Mood dependant memory
about person/person match
- recall dependant on match between mood state at encoding and retreival
experiment of Cognitive dependant memory
Marian & Neisser (2000) Bilingual study.
Task.
- Memory cues provided in Russian or English.
asked to recall life events from any period.
Found
- when pp cued in Russian, more likely to generate Russian related memories from prev events. vise versa .
Recognition memory. how do you gage the level of guessing or descision making bias for accepting items as old
- add in new distractor, foils or lures.
they will give an indication of error and .: guessing
SDT (signal detection theory) principles in recognition memory
+ outcome matrix
auditory task:
- detect a tone (signal) presented against a background noise.
it is either hard or easy to detect (depending on similarity to background noise).
4 outcomes in a matrix. Stimulus present vs absent. Participants response as YES vs NO (recognition).
YES + Present = HIT
NO + Present = MISS
YES + Absent = false alarm
NO + Absent = correct rejection
Recognition memory applies same matrix but with Old and New rather than Yes and No.
2 accurate, 2 inaccurate responses.
Familiarity graph concept in recognition memory
key concept i acc need to remember: appreciate each outcome from recognition memory test and their relationships define how well performance is.
and can get metrics from this to gage memory performance.
surplus info:
two normally distributed graphs of equal size on an axis of familiarity (against probability ofc).
- old item curve will be higher up as they have been studied and are more faimiliar.
- New items will be lower down.
BUT
there will be some overlap.
possibly because some new words are similar to old words and seem more familiar.
assume pp will set a criteria for how familiar an item is. any over this line will be rated as old (studied).
so can do mathamatical calculations
limitations of SDT and recognition memory
SDT cannot account for all memory phenomena
e.g.
- word frequency effect. low frequent words start with little familiarity (and are worse recalled). However, in recognition tasks, low frequency words do much better.
- SDT incorrectly predicts that low frequency words should be less familiar
Dual process theory
recognition memory is based on two processes.
Familiarity and Recollection.
Familiarity:
- a sense of memory without being able to remember contextual info
- described by signal detection
- Faster and more automatic
Recollection:
- Retrieving contractual info about a stimulus- like cued recall.
- Slower and more attention demanding
Recognition Memory: Measuring
Remember/Know Procedure (Tulving, 1985)
prompt participants.
p.p decide
remember item being presented previously then
- Recollect contextual details
- Measure of recollection
Know it was presented previously:
- seems familiar
- measure of familiarity
