Cognitive Psychology Lecture 7-10 Flashcards
How do we understand the world?
Through automatic and effortless evocation of categories.
we understand categories rather than individual stimuli.
How do we test categorisation?
by slowing down categorization by making very difficult.
What 2 theories do we need to understand theories of categorization?
we need to understand that features may vary in salience (attract attention) and validity (how important they are for categorization)
What type of features are there?
features may be:
Perceptual (Not everyone see the same features)
Functional (features need not be physical)
Relational, Emotion, Causal
Intergral or seperable (Colour-hue, saturation &brightness and Faces)
What are the benefits of categorization?
- Reduce complexity of environment
- enable us to relate classes of objects and events (3 levels-superordinate, basic, subordinate)
- Provides means for identification
- Allows for generalization
- Provides basis for deciding what is an appropriate action
- In real life- recommender systems that predict what we might like to buy
When are generalisation of categories greater?
Similarity: the more similar we think it is, the more likely we are able to generalize.
Typicality: the more typical the thing is to the category, then it allows for greater generalization (Osherson et al., 1990).
Specificity: the smaller the category is from the thing, the better we are able to generalize (Osherson et al., 1990).
Variablity: Generalization is greater when examples are more variable
Name the 3 levels of categorisation and give an example of each.
Hierachical category levels Superordinate: highest - fruit. listed slightly more basic level categories. often functional terms Basic: normal - apple. listed most often. Children speal this first- most used type of terms @ 2 years old during 2 hr recording of MLU based on morpheme count (1-2 morphemes) Often nouns and adjectives Subordinate: specific - granny smith
Category levels are not definite and may change depending on context.
How do similar events/things get categorized?
similar things get clustered together on a map/grid thing.
but the further the things get in the category, may become a whole new category. in this case, there will be no more generalizability
Are concept subtle/obvious? Are concepts simple/complex?
concepts are subtle and complex
Do categories have a large/small degree of variation?
categories have a large degree of variation
What is the classical view of concepts according to Plato, Aristotle, and Jerome Bruner?
Plato: world possess natural partitions, concept representation involves learning the rules that govern the structure of partitions
Aristotle: concepts are mentally represented as definitions which provide the necessary and sufficient conditions for category membership (any member is equal to every other member, no in-between cases, either in the category or not)
Jerome Bruner: Categorization involves discovering the defining attribute in the environment
Do complex or simple rule take longer to learn?
complex rules take longer to learn
number of stimuli determines complexity
Are categories definite or can they change?
categories can change.
categories can be made on the fly (Ad hoc)
prototype enhancement effect: prototype is falsely remembered with same rate as old items
new items that have not been seen before are falsely remembered according to typicality
What is a “prototype”? + Person
Eleanor Rosch: average of all category members
Problem: average may not be an accurate representation. Examplar theory solves this
What is an “Examplar”? +Person
Rob Nosofsky: all category members, similarity based on distance.
In examplar theory, there would be no prototypical bird.
on the surface, it would appear that it did not reduce cognitive complexity, but it does if we allow for selection attention mechanisms to influence how similarity is determined across the category
What categorical theory best describes our understanding of how people categorise?
Examplar theory- most of the time.
but sometimes it changes.
How do we Learn?
Selective attention
neural mechanisms of learning
cognitive phenomenon
What is knowledge and how does it affect our interpretation of new information?
Knowledge = things we have already acquired
influences our interpretation of new info by explanation, fitting to our thinking, and learning about causes
What are the types of learning?
- Habituation: Decrease to stimuli after repeated exposure w/o reinforcement
- Sensitization: Increase in response to stimuli after strong stimulation
- Associative Learning: Classical conditioning (association of originally neutral stimulus with later salient event), Operant Conditioning (learning based on modification of behaviour due to consequences.
Who studied Aplysia and why did he study it?
Also name the features of a typical Aplysia.
WHO: Eric Kandel
WHY: Big neurons therefore easy to insert electrodes so can study neurons easier. studied Aplysia when they encountered danger by tickling their gills and siphon with paintbrush b/c they use it for respiration
WHAT: Sipon, Tail, Gill, Mantle
Describe an Action Potential and its related details.
-70mV resting membrane potential
Electrotonic conduction = depolarization = influx of Na+
AP is all or nothing
Sodium potassium pump = Na+ out, K+ in
Ca2+ facilitates neurotransmitter docking in synaptic cleft
Outline Habituation in Aplysia
Test gill tickling reflex
after 10 mins stimulation –> 15 mins habituation
after 4 days stimulation –> 1 week habituation
What they thought:
Sensory neuron does not react as strongly
motor neuron does not react as strongly
increase inhibition from interneuron
The actual reason:
with successive stimulation, less neurotransmitter released
different wiring so there are less and less connections between siphon and gill
Outline Habituation in infants
presented infants with 2 pictures each time, measure how often infant looked at repeated pattern. Discovered it manifests in humans as boredom or ignoring cue.
1-2 month old infants: looked at old pattern equal amount of time
Older infants: decrease in amount of time looking at old pattern, increase amount of time looking at novel pattern
What can memory be described as in terms of learning?
memory can be described as long term sensitization
What is a classic example of sensitization?
Learned food aversion
Know Classical conditioning + Person + Experiment
Pavlov, Dog, UCS, UCR, CS, CR
classical conditioning produces a series of molecular changes at the pre and post synaptic gap of the motor neurons
AP more likely to be produced because of change in cell membrane permeability
What is Latent inhibition?
Latent inhibition is the pre-amptive introduction to the conditioned stimulus before classical conditioning commences
Pre-amptive introduction does not transfer over contexts
slows down learning of conditioned response
Is learning context specific or broad? What evidence is there to show this? + Person
What percentage of people knowledge partition?
specific
Latent inhibition and habituation is evidence of context specificity
Roediger & Kaepicke (2006): matching context between self test and final test predicted better scores
Lewandowsky & Kirsner (2000): Firefighters partioning in fires and backburns
~ 2/3 of people knowledge partition
dependent on attention to cus or components of cues.
specific
Latent inhibition and habituation is evidence of context specificity
Roediger & Kaepicke (2006): matching context between self test and final test predicted better scores
Lewandowsky & Kirsner (2000): Firefighters partioning in fires and backburns
~ 2/3 of people knowledge partition
dependent on attention to cus or components of cues.
“Once you eliminate the impossible, whatever remains, however improbable, must be the truth”
What is abductive reasoning?
What experiment is a example of abductive reasoning?
if you have a hypothesis, you look for evidence to back it up.
otherwise, you rule it out and repeat the cycle
this reasoning is aided by existing knowledge
The Blicket detector experiment is an example of adductive reasoning.
What is Bayes rule? Which experiment is an example of Bayes rule in use?
Bayes rule is Prior belief X Liklihood of observed event
how much one should update their hypothesis after observing some event
Do people prefer simple or complex hypothesis in explaining reasoning or deduction?
simple hypothesis
What is Occam’s Razor?
Occam’s Razor = people prefer explanations that explain more data with minimal assumptions
the simplest hypothesis that fits the date is probably correct explanation
What did Frank Keil’s experiment on human knowledge show? (Hint: Helicopter)
showed that people thought they knew more than they did
people worse at explanatory knowledge than declarative knowledge
people were good at ad hoc explanation
people are good at knowing where to look for information rather than actually knowing information
Outline sensitization in Aplysia
tail shock when gill stroke
series of 4-5 shock produced sensitization for days
tail shock leads to greater influx of calcium
receptor that is normally blocked becomes unblocked b/c of increased calcium. Therefore there are more receptors for the neurotransmitter.
How do we induce sensitization?
by presenting 2 stimuli simultaneously
strong stimulation of 1 input makes response to other input stronger
In attention training, outline the Blocking procedure.
What did it show?
Blocking: During early training, attention shifted to A (Red light) as a important predictor of X (food on left). So there is no attention for B (Bell) when A.B are Paired. So cue D (Blue light) drives Final response in B.D
However, if B.D was paired, then there was a 50:50 chance on either side
SHOWED: if attention had been drawn to one previously, new stimuli would not affect it.
BUT: if 2 previously shown stimuli had been paired, then a 50:50 chance occurs
In attention training, outline Highlighting procedure.
What did it show?
Highlighting: A & B already paired, Attention shifted to D in A.D because D (Blue light) is new and alone predicts Y (Food on right). So cue D drives final response
SHOWED: If one stimulus is repeated with a different pair with response, pairing with new response will not predict the old response. The new cue alone will drive the new response.
What model does not support Occam’s Razor?
Rescorla-Wagner Model
Outline the theory/experiment of unidimensional rules? + Person
What did it show?
Kruschke (1993)
given box and line in it, and asked to guess which category it fit in, then given feedback.
Split into 2 conditions (stimuli same in both conditions): filtration condition (upper/lower or left/right filtered into different categories, therefore ignore irrelevant dimensions) & Condensation condition (no filtration- all stimulus in one category)
Filtration category performed better than condensation category
not cf with Rescola-wagner model which predicts the 2 should be learnt at the same speed.
What us attention?
it is the brain’s ability to self-regulate input from the environment
What are the 2 uses of the concept of “Attention” in psychology?
Sustained attention (alertness): Psychological arousal, vigilance performance declines over a long watch
Selective attention: limited amount of stimuli we can process, attend to some at expense of missing others. People have limited capacity and don’t treat stimuli equally
What is the “cocktail party” problem? + Person
Cherry (1953): how are we able to listen to one conversation among many?
because translation is selective
Use experiment on dichotic listening and shadowing to mimic cocktail party.
Explain Cherry’s findings from the experiment using dichotic listening and shadowing:
- No memory for unattended message
- Switch from English to German: not noticed (because acoustically similar to English)
- Switch from male to female: noticed
- Reversed speech: ‘something queer’
- Switch from voice to 400cps pure tone: noticed
ONLY SUPERFICIAL (PHYSICAL) FEATURES WERE NOTICED SEMANTIC CONTENT NOT NOTICED
Neisser (1967) said that this was because this was preattentive and did not require focal attention. Also said it was possible we are aware of unattended stimuli but only superficially.
What happens when different messages using the same voice are presented simultaneously? + Person
Cherry
Binaural presentation: same voice, different content
need 10-20 presentations before they could shadow message 1
SHOWED: Source localization in space is important cue
What is one shortfall of Cherry’s experiments?
Even though he was interested to see what was perceieved, he was actually measuring what they remembered.
Sounds may be perceived and then forgotten
According to Broadbent’s (1985) filter theory, what role does attention play in memory?
- Filter after short term store, before limited capacity
- Prevent brain from being overwhelmed by all stimuli in world
- Meaning extracted in limited capacity store
- Decay quickly in STS if not quickly selected & stored
What evidence is there for filter theory?
Dichotic digit stream:
- Temporal recall (recall in certain order)- 3-4 correct
- Ear-by-Ear recall (Preferred recall order different for different people) all 6 correct
WHY?
Because 5 Switches need to be made in Temporal recall
and only 1 switch to be made in Ear-By-Ear recall
Therefore, Switching takes time & STS decays, so remember less
What evidence is there that suggests that Broadbent’s Filter theory may not be the whole story? + Person
1 . Gray & Wedderburn (1960): ‘Dear Aunt Jane’
- goes against what Broadbent said
- people preferred recall order was to follow semantic content: “Dear Aunt Jane” & “ 6,3,5”
- SHOWS: meaning extracted before filter switches
- Moray (1959): Dichotic listening task
- people could detect own name on unattended channel
THESE PEOPLE STARTED EARLY VS LATE SELECTION DEBATE
What is the early VS late selection debate?
Debate on where the filter is, before or after limited capacity channel where semantic content is analyzed
Treisman: Early selection supports Broadbent in placement of filter but not what it filters.
Treisman supports sensory analysis before filter, semantic content after filter
Late selection against Broadbent in everything. Everything to semantic analysis but filter before going into awareness.
What does Late selection say that differs from early selection?
Differs on where Semantic content is analyzed and differs on the amount of content that is analyzed
What does Norman (1968) say about late selection?
Everything, regardless of attention, gets into LTC system and content is semantically analyzed.
Bottom-up = stimulus driven
Top-down = Selection by “petinence” (our own attention to it because we thought of looking at it b/c it is distracting/caught your eye)
What is Treisman’s (1961) Attenuation model? How is this different from Broadbent’s Filter Model?
Broadbent’s filter completely blocks unattended stimuli
Treisman’s passes partially and strength get reduced, so only part semantic activation occurs
– Semantic activation was biased by context would get through the filter and shift attention
• Highly salient stimuli (name)
• Semantically related material (dear aunt jane)
What evidence is there for Early selection?
Treisman & Geffen (1967)
- target word ‘tap’ % detection higher in shadowed channel but not 0 on unattended channel
- cf. with Treisman’s attenuation theory rather than Broadbent’ Filter theory
Treisman & Geffen (1969)
- target word ‘tap’ detection was equal % on both because acoustically different and preattentive (Different voice). Meaning not needed.
What is one shortfall of Early selection?
Complexity of filter: need to distinguish related from unrelated stimuli to respond to highly salient or obvious semantic content
Late selection tries to rectify this by moving filter
everything semantically analyzed but not sufficient for awareness
- ES theory: activation = awareness
- LS theory: need to pass filter for awareness
What evidence is there to support Early Selection theory and who are the theorist?
Treisman & Geffen (1967): Poor in same voice
Treisman & Riley (1969): Good in different voice
What evidence is there for Late selection and who are the supporters?
McKay (1973);
VonWright, Anderson & Stenman (1975):
Semantic activation on unattended channel (but shown by indirect means)
Some people said neither LS or ES worked, why?
Some phenomena not predicted by either theory
Did participants have higher accuracy on selective or exclusive monitored channels when studying divided attention? + Person + Outline experiment
Moray (1970): Cost of diving attention
Auditory signal detection task
used pure tone rather than word or something else
54% - Exclusive OR (XOR)- monitor both, respond to one
52% - Inclusive OR (IOR) non-simultaneous- monitor both, respond to both
31% - Inclusive AND - simultanous- monitor both, respond to both, target on both came up simultaneously
What were the implication for the Moray (cost of divided attention) study?
Moderate cost of divided attention
large cost for simultaneous detection
ES & LS could account for one condition but not both
ES Predicts (OR < SEL): attentuation with divided attention but not (AND < OR): attentuation should not depend on identitiy of stimulus (because it should be preattentive- tone stimulus) (Simultaneous detection)
LS Predicts (AND < OR): (Simultaneous detection) because selected by pertinence & Compete to get through filter But not (OR < SEL): (divided attention) only 1 target so no competition so should be close to 100% but we only get 52%
What were the 2 new theories that arose from Early and Late selection debate? (1970)
Structural bottleneck theories:
- ES & LS are structural bottleneck theories.
- Performance limitation due to competition to pass through bottleneck
Capacity resource theories:
- people have limited capacity to activate brain structure
- can’t activate all at once
- therefore can’t process everything
- Measured by pupil dilation (More dialated-> more capacity required)
Who was the main proponent of Capacity theory?
What did he say/look at?
Kahneman, 1973
- Some tasks require more capacity that others
- We can flexibly allocate capacity to simultaneous tasks
- Allocation may be concious or unconcious)
- Easy task = little capacity
- Hard tasks = large capacity
This suggest a Attention Operatic characteristic (AOC) way of studying attention
What is the difference between detection and discrimination?
Describe the study which looked at this difference. + Person
Detection = say 'yes' or 'no' Discrimination = say 'increment' or 'decrement'
Bonnel & Hafter (1998)
- Visual and auditory task for discrimination and detection
- Measure under signal detection theory to say there is something and they detected it- measuring sensitivity
- tradeoff for driscrimination (limited capacity)
- no tradeoff for detection (unlimited capacity)
- cf with Cherry- preattentive
- extraction of meaning needed focal attention/further capacity
What are the pros and cons of capacity theory?
Pro:
- suggested new way of looking at attention
- emphasized divided attention
- showed flexibility of attention
Con:
- Vague, people could always come up with ad hoc explanation- always falsifiable
When looking at visual attention, What are researchers interested in measuring?
interested at looking at attention prior to eye movement.
hard to do back then, but now we have equipment for it
known at attentional orienting (1980)
Describe the findings of studies using the Spatial cuing paradigm. + Person
Posner
- Attract attention to A, present stimulus at A or B, then compare if A (in spotlight) is faster than B (out of spotlight)
- SOA (stimulus onset asynchrony) for eye movement = ~ 200ms
- benefit to cuing at valid location (80% of time - ~250ms) rather than invalid location (20% of time - ~300+ ms). cuing at neutral location (50:50 effect - ~270ms)
Reasons why this finding occured:
- Shifts of spotlight: time taken to shift/less time needed because there is no need to shift
- Capacity theory: Capacity allocated dependent on type of trial (neutral-50:50, invalid/valid- focus on 1 location)
- BUT: Hard to test between the 2 theories
What is Posner’s “Spotlight of Attention” referring to?
When we have our spotlight on an item, we process it more accurately than when your ‘spotlight’ is not on it.
Move spotlight when we want to, to where we want to
Describe the difference between top down and bottom up orienting in the natural world.
Top down and bottom up orienting systems are Attentional Orienting mechanisms
Bottom up orienting: something in environment calls attention to it. (out of corner of our eye)
Top down orienting: we want to look at it so we look to it
What does this suggest about the systems we have in our heads for visual orientation?
- need both to function properly
- seperate systems
- there are people who show failure to focus and failure to disengage attention
What are Endogenous and Exogenous attentional systems?
Endogenous = Voluntary (interpretation required - cognitive/symbolic/central)
Exogenous = Reflexive (interpretation not required - direct/spatial/peripheral)
Engage by different types of cues
What evidence is there for humans having separate orienting systems?
- Cuing effect
- peripheral (reflexive) peaks rapidly but transient (~100-150ms)
- central (cogntive) peaks slowly but stable (~300ms) - Jonides (1981): Different effect on loads
- cues are different (2 seperate systems) because one is affected by cognitive load (voluntary) and the other is not (reflexive). - Peripheral cuing effect- “inhibition of return”
- reflexive cues for miscued location improves for long SOA (over 300ms) and slower for cued locations.
- Said to be ecological/evolutionary - don’t need to search things twice