PSYC213 Midterm 2
Lecture 10: Memory part 1
[13 février]
10A: Basics
What memory does
Routines and habits
Sense of self
facts we have about self developed from experiences
Social functions
Solving problems
Clive Wearing
Worst case of amnesia ever seen
Episodic memory was impaired
Other memory forms remained intact
semantic memory: knew his wife
procedural memory: played the piano
Distincts memory systems supported by different neural circuits
10B: Stages of memory
The three stages
Encoding: learning new info, forming new “memory trace” as neural code
memory consolidation to storage
Storage: retaining encoded memory trace/neural code
Retrieval: activating a memory trace via a cue for a purpose
memory is in patterns
Neurology
Memory trace is formed (encoding) as a hippocampal-cortical activity pattern
Retrieval triggers pattern completion of the brain pattern
memory cue causes the completion
Memory is transformed into stable cortical pattern in consolidation
Memory as systems
Sensory input
Sensory memory
1 second
info not transferred to STM is lost
Short Term Memory
consolidated with rehearsal
30 seconds
info not transferred to LTM is lost
Long Term Memory
back and forth with STM
Sensory memory systems
Automatic reflections of a sense
Gustatory/olfactory memory
memory of the chemical compounds of taste/smell
smell has very close link to memory
Echoic memory
sound byte held for ~3 seconds
Haptic memory
very brief memory of a touch
Iconic memory
millisecond visual memory
vision persistence/afterimage
Afterimages
Positive afterimage: visual memory that represents the perceived image in same colors
helpful for seeing things smoothly
helps fills in holes in videos (24 f/sec instead of 75 f/sec)
Negative afterimage: visual memory in inverses colors of the perceived image
lasts slightly longer (few seconds) than positive afterimage
Length of sensory memory
Sperling (‘60): participants viewed for 0.5sec a visual display of 3x4 letters, asked to recall them after a tone
whole report: reporterd letters from the whole display
partial report: reported only one row of letters at a time over trials
Partial report conditions remembered + with - delay
STM
Attended info moves from sensory memory to STM
Located in prefrontal cortex
Limited time capacity of 20-30 seconds
Very limited capacity (7 ~+-2)
Serial position effects
Primacy effect: first item remembered +
rehearsal -> LTM
Recency effect: last item remembered +
effect eliminated is delay duration >30sec
10C: Multi-store and modal memory
Chunking
Strategy of grouping items together meaningfully so + info is represented at once
“cat dog pig” instead of “tdi ogg acp”
Increases with knowledge
ex: expert chess players recall + pieces on board than new players if in game, no difference if pieces are random
Working memory
Retention/manipulation of info not in our environment in conscious awareness
Guides behavior
Essential for many cognitive functions
Episodic buffer: integrates info from STM and LTM
“conscious awareness”
Phonological loop
Phonological store: passive store for verbal info
“inner ear”
Articulatory control loop: active rehearsal of verbal info
used to finger written material into sounds
specialized role in language
“inner voice”
Visuospatial sketchpad
Visual cache: info about visual features
Inner scribe: info about spatial location/movement/sequences
Separating stores
Differents brain areas active for visual/break STM tasks
Patient ELD has problems recalling visual/spatial but not verbal info short-term
Patient PV had problems recalling verbal but not visual material short-term
both examples of double dissociation
Implicit/explicit in LTM
Implicit memory: non-declarative, non-conscious
Explicit memory: declarative, conscious
Retention Interval Time between encoding and recall
Ebbinghaus
Learned nonsense syllables and tested me,Roy at various interval to see what was retained/forgotten
created over 2000 cards of fake syllables
Learned sets under strict conditions to remove confounds
read without any inflection
read at a consistent pace of 2.5/sec
did nothing else during experiments
Forgetting curve is exponential
memory loss is largest early on then slows down
Effects
Spacing effect: forgetting is reduced when learning is spread over time
repeated info + valuable
Testing effect: retrieving memory after a test leads to deeper encoding
active rehearsal
participants studied text passage, those who studied + had better ST recall, those tested had better LT recall
Levels of processing theory
Strength of memory (and forgetting potential) depends on processes used in encoding
Shallow: focus on sensory info
ex: memorizing vocabulary words in new language
Deep: integrate higher-level knowledge
things we know with learned info
ex: using new words in a sentence
Memory stronger with deep processing
+ elaborate memory traces
Mnemonics
Use deep processing
Organizational strategies to help encode info
Often involves linking new info to prior knowledge
new info to semantic info
Chunking strategies
ex: acronym to remember lists
Imagery and Method of Loci
use familiar image to link encoded info together
imagery helps memory
Forgetting
Decay theory: memories are lost over time due to disuse
like a muscle, memory not used gets weaker
Interference theory: interference responsible for majority of forgetting
encoded memories are labile, need to be consolidated into stable LTM
in pre-consolidation period, memories susceptible to disruption and interfering info
Effects of interference
Proactive interference: prior info interfere with encoding a new memory
ex: can’t learn new phone number bc of old one
Retroactive interference: newly learned info overwrites/interfere with a priorly encoded memory
ex: can’t remember old password after forming new one
Similarity effect: + alike something is to what’s already learned, the + it’ll mingle and interfere with memory
Lecture 11: Memory part 2
[15 février]
11A: Encoding specificity hypothesis
Basics
Memory retrieval better when there’s overlap in encoding context
Context can act as a retrieval cue
Context can be internal state or external environment
state dependent-learning: better retrieval when you’re in the same internal state (feelings, intoxication, etc)
external context: better retrieval when you’re in the same environment (ex: sea diving vs land experiment)
Transfer appropriate processing
Memory depends on relationship between learning/testing
Highlights importance of encoding context/retrieval cues
11B: Explicit memory
Episodic vs semantic memory
Episodic memory: specific events and episodes, with encoding context
ex: dancing with friends at prom
Semantic memory: facts and general informations, without context of learning
ex: prom happens at the end of secondary school
Episodic memory and hippocampus
Children with hippocampal damage had episodic memory impairment, but normal semantic memory
ex: can’t copy an image after a delay, but have a normal factual knowledge
Means episodic memory depends on hippocampus
this dependency decreases with time
Semantic dementia
Relatively spared at episodic memory tasks
Very impaired at word naming/picture matching task
scored worse than people with Alzheimer’s
Memory and consciousness
Anoetic consciousness: implicit memory, no awareness, no personal engagement
Noetic consciousness: semantic memory, awareness, no personal engagement
Autonoetic consciousness: episodic memory, awareness, personal engagement
called “mental time travel”
Reappearance hypothesis
Episodic memory trace recalled the same at each retrieval
reproduced, not reconstructed
Recurrent memories are unchanged from original events in PTSD
11C: Flashbulb memories
Basics
Vivid memories of significant events
emotionally arousing/shocking events
retrieve specific details about time/place of hearing the event
Flashbulbs over time
People tested about 9/11 recall after 1/6/32 weeks
DV: details used, recollection vividness, confidence
consistent detail and inconsistent details
No detail difference between flashbulbs and everyday memories
Flashbulbs ratings of confidence and vividness increase over time while accuracy decreased
Flashbulbs memories can change
ex: a flashbulb (OJ Simpson trial) recalled after 32 months had 70% inaccuracy and 40% major distortion
Flashbulbs vs episodic
Flashbulbs aren’t recurrent recording of events
Flashbulbs retrieval changes over time, aren’t resistant to distortion, even if memory feels very strong
distinction between subjective/objective memory
must accept theory that memories are reconstructed
Memory consolidation
Experiences are encoded then consolidated into LTM
Consolidation: STM to LTM
Recall: LTM to active memory
Reconsolidation: active memory to LTM
when a trace is activated it becomes de-stable
cortical connections can be strengthened/modified, altering how memory trace is reconsolidated
retrieval changes a memory trace
11D: Distortions in memory
Episodic memory construction
Constructing memories at retrieval mean it can be distorted
May use semantic memory/schemas to infer the way things “must have been” in a recalled memory
insert false info into constructed memory, affecting retrieval
Semantic knowledge affects retrieving detailed memories
Schemas and distortion
Schemas organize/categorize info, provided expectations
War of Ghosts experiment (Bartlett, 1932): people read an unfamiliar Native folk story that don’t match western folk stories schema
they recalled a simplified version of story, became + conventional with each retrieval
omissions/alterations to match Western schemas
Study scene associated to schema-consistent items removed
Distortion effects
False memories: familiar feeling can lead to incorrect associated, details can be added to memories in retrieval
Misattribution effect: retrieving familiar info from wrong source, failure in source monitoring
not remembering where/when accurately
Misinformation effect: leading questions can cause false memory formation
how a question is framed can affect how info is remembered
Rashomon effect: memories are reconstitutions, it’s why people recall the same event differently
the elephant in the dark experiment
Implanting memories
People recalled childhood experiences recounted by their parents over three experimental sessions
A false memory was added to the list by the experimenter
20% of people had a false memory of the event by the end of the third session
Same processes that help us constructed the past help us imagine/plan the future
processes of hippocampal episodic memory
Summary
Episodic memories impacted by prior knowledge
memory driven by our expectations
Episodic memories reconstructed at retrieval
memories aren’t reflections of truth, they’re subject to distortion
Distortions can be false memories, but it reflects an adaptive characteristic
Same processes helping us construct the past help us imagine/plan future
processes of hippocampal episodic memory system
Lecture 12: Memory part 3
[20 février]
12A: Long-term memory
Basics
Explicit memory: semantic (facts) and episodic (events)
Implicit memory: priming, procedural, classical conditioning
priming: neocortex
classical conditioning: emotional responses (amygdala) and skeletal musculature (cerebellum)
procedural memory: skill and habits, basal ganglia
Procedural memory
Automatic behavior/actions
Pattern of movements encoded in the brain
basal ganglia for motor sequence
prefrontal cortex for organization
+ immune to forgetting compared to other types
Habits
First rely on explicit memory, with training/exposure, start relying on implicit memory
motor action sequences
Can form addictions or repetitive thoughts/emotions/OCD
Habit formation: requires the striatum
rats trained in T shape maze with tones to signal reward at left or right end
Breaking habits: need to inhibit prefrontal cortex
removing reward at one end, or making one reward non-rewarding didn’t break the habit
cue -> habit -> reward -> new behavior -> new reward
Priming
Prior exposure facilitates info processing without awareness
Word-fragment completion test
participants shown list of words, asked to complete words fragments
likely to use prior words to complete fragments without knowing it
Implicit emotional responses
Conditioned emotional responses
adaptive fear responses to scary stimuli
Amygdala: implicit emotional memory and modulating explicit memory
emotional enhancement effect
Semantic memory organization
Large units linked to properties with pointers
ex: “animal” unit linked to “bird” unit, which is linked to “feathers” property
general to specific
Spreading activation: automatic activation spread from activated concept to other interconnected aspects
spreading activation to features
ex: thinking about a crow will trigger activation in related bid concepts
Semantic priming: related ideas triggered at retrieval
train of thought may seem nonsensical
12B: Amnesia and dementia
Dissociations in LTM
Experimental neurosurgery to reduce seizure activity
bilateral medial temporal lobe, and hippocampus, removed
caused selective episodic memory loss
HM and the hippocampus
Had the neurosurgery that removed the hippocampus
Intact STM, could remember a short list of words for 30 seconds
Intact procedural memory, could learn new skill based tasks
Intact semantic memory, could recall major childhood event
Profound episodic memory loss, couldn’t learn new info and recalled his past in sparse details
couldn’t remember/describe details of past
couldn’t encode new events
couldn’t imagine the future
Amnesia type
Anterograde amnesia: can’t form new episodic memories
Clive Wearing
Retrograde amnesia: loss of memory from before the amnesia onset
temporally graded, recent memories + affected than old ones
Dissociative amnesia: very rare psychiatric disorder, usually a response to psychological/physical trauma, not injury
commonly retrograde amnesia for episodic memory and autobiographical knowledge
leads to lifestyle shifts link moving, assuming new identity
Dreams and amnesia
Dreams linked to memory
theory that dreaming helps process past experiences
Patients with focal bilateral hippocampal damage and amnesia woken up during night and asked about dreams
damaged reported - dreams, and they were much - detailed
Dementia
Progressive cognitive/functional impairments due to neuronal death
63% of all dementia cases are Alzheimer’s disease
Medial Temporal Lobe regions the first impacted by AD
Earliest symptom is a episodic memory deficit
MCI due to AD (first step): hippocampus and MTL, some memory loss
Mild: lateral/temporal/parietal lobes, poor object recognition/reading/direction sense
Moderate/Alzheimer’s dementia: spread to frontal lobe, poor judgment, impulsivity, short attention
Severe: widespread brain atrophy, language/function loss, basic motor skills
Semantic memory: starts in left anterior temporal lobe, convergence zone for semantic concept representations
deficit recognizing faces/words/uses of objects
12C: Differences in memory
Healthy aging
Volume loss of ~5% per decade after 40
not all areas affected equally
Implicit/semantic memory is impact, episodic/working memory is impaired
ability to remember associations is important for episodic
Cognitive aging theories
Older adults have déifier in general executive cognitive processes from frontal lobe atrophy
slower to process info, can’t inhibit irrelevant info
ex: have trouble focusing on one picture and ignore others
Associative deficit hypothesis
Older adults have problems encoding/retrieving associations in memory due to hippocampal atrophy
familiarity/single item (non-hippocampal) less affected
recollection (hippocampal) + affected
Older adults had + trouble with face-name associative recognition than younger
- trouble in simple face or name recognition
Memory test in scanner showed YAs and low memory old adults both used right PFC, but high memory old adult used bilateral PFC
neural compensation
Taxi drivers and extreme memory
Memory and space intimately linked
taxi drivers memorize 25 000 streets within 10km radius
Taxi drivers performed better on spatial memory tests and had r + posterior hippocampus grey matter volume
volume of posterior hippocampus related to years of experience as a driver
Highly Superior Autobiographical Memory
HSAM can remember every day of their life in detail
enhanced autobiographical memory
recalling very detailed daily memories
Doesn’t involve mnemonic strategies
HSAM people don’t remember a word list better
HSAM speific to personal memories
Consistency in recalling memories related to OCD symptoms
Goldilocks principle: memory words well with just the right amount, not too little or too much
Lecture 13: Concepts and knowledge
[22 février]
13A: Concept meaning and organization
Terminology
Concepts: general knowledge and mental representation of a category
Categories: items grouped to gueule according to concept
Exemplars: individual items in a category
Generalization: deriving a concept from specific experiences
Concept organization
Superordinate, basic and subordinate
super: - precise than the basic
sub: + precise than the basic
Goes from less precise to + precise
Concepts in development and disease
Child learn basic, than super, than sub concepts
Semantic dementia patients can use basic level concepts, impaired as disease progresses
later in illness, use super concepts
Cognitive economy
Organization allows efficient access of knowledge
Use simplest terms still meaningful for the situation
ex: using basic/super for children, sub for experts
13B: Concept learning and representation
Classic approach
Concepts involve forming rules about lists features
defining features: necessary/sufficient for category membership
characteristics features: common, but not essential for category membership
Feature comparison between encountered items and list
refines what a defining feature is for a certain concept
Works well for simple concepts, less for complex/variable/ambiguous
ex: a fur-less dog, a student, etc
Similarities
Concepts are defined by resemblance to a collection of features
not by defined features
Items are +/- part of a category, but an item can be categorized in + than 1
lines can be fuzzy
Prototype theory
Categories formed from overlap of exemplars
extracted from experience
Each category has an abstracted prototypes that’s pre stored in memory
represents most common features with other members
Exemplars included in category network around prototype
similar items stored closer together to the prototype than dissimilar items
People learned to classified variants of a prototype in dot patterns, classified studied/new/prototype patterns
didn’t see the prototype
worse at classifying new patterns than old
equally able to classify prototype and old patterns
Prototype is an abstracted exemplar
other members remember prototype to certain degrees
+ obscure members are further away in network
Typicality effect: preference for processing items close to the prototype
The role of context
Prototype theory: concepts are context independent
don’t account for situation determining concept representation
Context affects typicality effect
faster/easier to recognize a concept if it’s in line with its context/surroundings
Exemplar theory
No single abstract prototype for a concept
every instance of a category is stored, not just prototype
Two steps to see if new item is part of a category
retrieve some/all exemplars of category members
compute similarity to new item at time of concept determination
Explains how context influence concept representation
experience/situational context used to form concepts at retrieval
Knowledge based theories
Explanation view of concept categorization
instead of similarity-based
Implicit intuitive knowledge used to assess new items
Essentialism: certain categories have an underlying reality/true nature that can’t be observed
13C: Embodied cognition and the brain
Embodied view of concepts
Concepts assessed as function of environment/current goals
Processed in different brain networks
shift depending on what’s required to be accessed from a concept
Can bring together dissimilar members into a single temporary category to meet a goal
Embodiment and the brain
Knowledge stored as sensorimotor neural representation
accessed representation as function of what info is required
Perceptual symbols system
Perception/conceptual knowledge as perceptual symbols
Activating a concept engage certain sensory-perceptions to engage mental stimulation as function of task’s goals
highlights importance of perception but also goals in storing/accessing knowledge
Property verification task: people faster to respond if a previous trial asked feature from same sense/percept
concepts represented via senses/perceptions
Brain representation
People passively read action words in MRI scanner
Specific brain regions that process movements associated with the words were active
Concepts rooted in motor/sensory neural activity
Neuropsychological case studies
Brain injury cases of people with category specific deficits
Some have selective impairment in naming living things, some in naming non-living things
ex: some can’t name animals, some can’t name tools
Sensory functional theories
Concepts represented by defining feature
Living things defined by visual features
visual processing regions
Inanimate objects defined by functional features
motor cortical regions
Lecture 14: Language
[27 février]
14A: Basics
Definition
Shared symbolic system for purposeful communication
symbolic: units to reference something else
shared: common among a group
purposeful: to communicate/translate thoughts
Flexibility
Language acts as a high level control system for the mind, allowing to sculpt mental representation of other and own
also sculpted by environment
Morphology/complexity decreases with languages spoken
Lexical tones partly determined by climate
tonal languages spoken + in warmer climates
14B: Language in the brain
Aphasia
Impaired language function, usually from brain injury
Broca’s (non fluent) aphasia: speech halted/difficult to produce, writing usually affected similarly
ranges from deficits in producing certain words to problems generating all forms of language
deficit depends on amount of damage to Broca’s area
expressive aphasia: intact language comprehension, impaired speech production/articulation
Patient Tan: large lesion in Broca’s area, could only say “tan”, tried to communicate with tone/inflection/gesture
Wernicke’s (fluent) aphasia: written/spoken comprehension affected, language content not meaningful/comprehensible
posterior superior temporal lobe damage, mostly left
verbal paraphasia: substituting a word with something semantically related
phonemic/literal paraphasia: swapping/adding speech sound
neologism: using made-up words
Conduction aphasia: impaired repetition, speech production/comprehension intact
load dependent
damage to neural pathway between Broca’s and Wernicke’s
Lateralization
Language often considered left lateralized
Not fully understood or linked to handedness
new data indicate up to 70% of left handed people still have left hemisphere language dominance
Broader aspects of language supported by right hemisphere
prosody, pitch, mood, attitude, gestures, etc
Right side seems important for higher-order non literal language use
right hemisphere lesion disrupt ability to interpret/express speech prosody
14C: Acquisition and comprehension
Nuturist/behaviorist view
Language acquisition is skill/associative learning
explicit training of language
Trial/error reinforcement and modeling other people’s language
Chomksy/naturist view
Language is not stimulus dependent or determined by reinforcement
Language is complex, acquired rapidly
Allows us to understand/speak what we haven’t heard before
Innateness hypothesis: grammar/syntactic structure separate from semantic meaning/cognition
Language Acquisition Device: entity supporting language
Universal Grammar: part of LAD that includes rules for all languages
children only need to learn language specific aspects to our on top of UG
Children exposed to different learning situations converge onto the same grammar
uniformity of (healthy) language development with age
Stimulus argument poverty
Child’s linguistic environment not sufficient to enough to allow the child to learn a language via reinforcement/rules/imitation
child doesn’t have enough language samples to acquire all language, not enough opportunities to learn from mistakes
must be something innate about language
Adult reformulation of child’s speech target the structure, not the meaning
Children extract regulations from experience to form rules
evidence that rules aren’t innate
Psycholinguistics
Buildings blocks of language
Phonemes: smallest linguistic unit (d, o, g)
english has a few dozen for morphemes
Morphemes/words: smallest meaningful units of language
Syntax: rules governing how words are arranged in sentence
Semantics: the meaning
Basics of comprehension
Understanding semantic from language
Resolving types of linguistic ambiguity using context/top-down processing
phonological: within a sound
lexical: within a word
syntactic/parsing: within a sentence
Ambiguity
Phonological ambiguity: determining phonemes depends on audio signal, which is often noisy
can use context/internal knowledge of speech sounds to hear
ex: noisy cafe over your conversation with a friend
Lexical ambiguity: single word can refer to more than one different concept
>80% of English words have more than one dictionary entry
basis of puns
homophones: words that sound the same with different meaning, correct one resolved the sentence context
Cross modal priming task
Word “bug” can mean two things
One group without context and one biasing context to “insect” meaning
Lexical decision shortly (short SOA) after hearing the word bug: both meaning active, bug primes both
Lexical decision with delay (long SOA) from hearing the word bug: into context biased meaning active
Both meanings initially retired, contextually inappropriate one quickly discarded
Parsing
Sentence parsing: diving a sentence into words and identifying them as nouns/articles/verbs/etc
Ambiguity can come bc we hear sentence incrementally or there’s often + than one way to parse a sentence
Garden path sentence: sentences with multiple syntax structures
interpreting a word one way leads to faulty interpretation
Syntax first theory: we use grammar rules to interpret a sentence as we hear/read it
local or specific
we parse with only grammar principles, in one direction, may get to the end and wrong meaning, must go back
Constraint based model: we use non grammatical info to help interest sentences and resolved any ambiguity
global or holistic
use + than grammar to parse sentence
can use semantic/thematic context, expectation, frequency
14D: Language and thought
Linguistic relativity
Language/thoughts are interconnected
Sapir whorf hypothesis: Language changes how we think/perceive
people who speak different languages think differently
ex: Inuits have + words for snow
Linguistic universalists: language/thoughts are independent
Findings
Russian, but not English, discriminate between lighter and darker shades of blue
Russian speaker faster to discriminate blues into divergent categories, English speaker showed not effect
language affects perception
English have more words for color than Indonesian Dani
suggest accessing color category without language labels doesn’t change across language
both groups match/categorize learned color patches the same
Dyslexia
Surface dyslexia: reads letter by letter, sounds out words
difficulty matching words to mental dictionary
impaired at producing irregular words (25% of English)
Phonological dyslexia: reads by comparing whole words to mental dictionary/lexicon
difficulty reading letter by letter and sounding out words
impaired at reading non words and new words
Lecture 15: Bilingualism
[29 février]
15A: History
Traditional story in psycholinguistics
Bilinguals considered a special group of language users
like those with brain damage or language disorders
Thought learning two languages would be confusing
Bilinguals should be functional monolingual in two languages
late bilinguals: learnt second language later than childhood
Even successful late L2 learners spell with accent and fail to get subtle aspects of L2 grammar
New attitude
Publications about bilingualism jumped around 2001
Greater plasticity than previously understood
Language learning occurs at all ages
Languages processes are dynamic
Bilingualism provides lens to understand neurocognitive processes
alters structure/function of mind
Bilinguals not two monolingual in one
15B: Discoveries
Language coactivation
Bilingualism is joint activation of two language systems
cognates: words same in two language, facilitating
interlingual homographs: same word with different meaning, interference
Facilitates/interference effects stronger in L2
and in language from different writing system
also present in bimodal bilinguals
Bilinguals show effect of parable activation in reading
Initial stage of comprehension: first fixation duration (length of 1st time eye fixates on target)
Later stages of comprehension: length of all eye fixations
longer fixation durations -> + comprehension difficulty
Impact of L2 on L1
Bice & Kroll examined cognate effects in monolinguals and L2 Spanish learners with English lexical decision task
Cognates has no behavioral effect
Event-Related Potentials: brain voltage fluctuations time-locked to an event
Cognates has no ERP effect on monolinguals
Cognates in early L2 learners reduces N400 ERP
so native language is affected by L2
Cognitive control
The inhibitory control model: supervisory attentional system affects inhibitory system but not the goal
inhibition: triggered to presence of competition, proportional to coactivation level and prevent intrusion from not used L
N400 ERP shows no effect between L1 and L2
however, behavioral reaction time is affected
N200 ERP associated to early stage of speech planning to operation of inhibitory processes
can be observed as negative shift over front central sites
peak between 250/350 ms after stimulus onset
Strong cognitive demands needed for language control and brain/cognition are plastic and affected by experience
underlies that multi/monolinguals could use their cognitive control processes differently beyond language processing
Lifelong management of two languages -> language control enhancement -> enhancement in domain general control -> brain changes
Flanker task conflict effect very reduced in bilinguals
Bilinguals have greater bilateral frontal gray matter volume compared to monolinguals
Individual differences
Bi/multilingual advantage found in studies
other studies question the differences
Bi/multi advantages modulate but tasks used to assess executive functions
Adaptive inhibition hypothesis: interactional context -> language experience -> language/cognitive processing
Dual language contexts may be particularly demanding
both language control and general cognitive processes
Multilingual langages experiences a continuum
Language diversity across social context crucial to regulate language representation/control/accessibility
over basic measures like L2 age of acquisition and self reported ability
Summary
Bilinguals are not two monolingual in one, both languages are active and competing
two languages not separate
Bilingual language sustek permeable in both directions
L1 changes in response to L2 learning
bilingualism affects both languages
Effects on cognitive control, brain/cognition are plastic
Not all bilinguals are the same
language experiences are multifaceted
bilinguals differ by context demands and experience
Lecture 15: Choice
[12 mars]
15A: Heuristics and biases
Basics
Heuristics: mental shortcuts allowing to skip careful deliberation to draw interference
Slow system: serial logical analysis of info
effortful, non automatic
Fast system: heuristics based reasoning
easy, automatic
Availability heuristic
Estimate event probability based on ease at which it can be brought to mind
Liechtenstein (78) asked which was more deadly between tornadoes and asthma
falsely rated tornadoes as deadlier, bc deaths by tornado is easier to come to mind
Affect heuristic: tendency to overestimate risk of event that creates large emotional response
Sunstein (02): people rated sharks as deadliest animal, especially after being exposed to shark attack media
Representativeness heuristic
Tendency to make inference on basis that small samples resemble the larger population they were drawn on
Related to availability heuristic
relies on pre existing knowledge structures
People base their judgment of group membership based on similarity
results in biases
Base rate neglect: failing to use info about prior probability of an event to judge event likelihood
Conjunction fallacy: false belief that conjunction of two conditions is + likely than either single condition
likelihood of event is always higher than likelihood of that even and something else
Anchoring and adjustment
Anchoring: tendency for people to overweight initial info when making decision
Kahneman and Tversky (74): people saw roulette landing on big or small number, than asked to make an estimate
bigger number on roulette -> bigger estimate
Important to design self report scales
Regression to the mean
When a process is rather random, extreme values will be closer to the mean (- extreme) when measured again
Can’t always attribute changes in performance to manipulation
Bounded rationality
Bounded rational: theory that humans are rational relative to environmental/individual constraints
people are satisficers, we look for good enough
« making do » with our humans limitations
heuristics provide incorrect answers and lead to biases but they also work
Ecological rationality
Sees heuristic as optimal approach to solve problems
Given right environment, heuristic can be better than optimization than other complex strategies
Summary
Heuristics/biases arise from limitations we face
can sometimes produce correct responses (ecological rationality)
Applying heuristics often lead to biases
15B: Decision-making processes
Types of decision making
Perceptual: objective criterion for making choice
Value based: subjective criterion for making choice
depends on motivational state and goal
Decisions under risk: when outcomes are uncertain
ambiguity: when you have incomplete info out the consequences
Risky decision making
Needs to decide even when outcomes are uncertain
Both extremes in risk taking can be very harmful
stagnancy vs impulsivity
Risk premium: difference between expected gains of a risky option and a certain option
Risk averse: decision maker has positive risk premium
most people are risk averse
Risk neutral: decision maker has zero risk premium
no difference in options
Risk seeking: decision maker has negative risk premium
don’t need chance of winning + to take a risk
Classical economy theory: rational thing to do is choose option maximizing expected value
account for individual risk preferences
People are inconsistent in preference taken as bias
we don’t follow expected value
The framing effect
Display of inconsistent risk preferences depending on problem framing (loss vs gain)
people are risk averse when options are described as gains
people are risk seeking when options are described as losses
Gachter (09): PhD students could either get a early discount, a late registration fee, or both to cancel each other out
93% of students signed up early when told they would pay a penalty fee
only 67% signed up early when told they’d get a discount
Endowment effect
Tendency to ascribe higher value to owned objects compared to identical unowned objects
case in which decision framing matters
people are averse to give thing up once there’s ownership
Prospect theory
Psychological theory explaining how people make decisions under uncertainty
two major features: shape of utility function and shape of probability weighting function
Utility: subjective value assigned to an object
context dependent
assigned to monetary amount as function of current state, not in absolute value
losses loom + than gains
Probability weighting: how people understand likelihoods
tend to overestimate rare events and underestimate mundane events
links to availability heuristic
Fourfold pattern: two features together compared to risk
loss-high prob and gain-low prob: risk seeking
loss-low prob and gain-high prob: risk averse
Dual process theory
Broad scientific theory that humans possess two decision making systems
fast/automatic system vs slow/effortful
People gave higher death frequency estimate when in a negative mode
Prediction Error: difference between what you predicted will happen and what actually happened
positive PE: unexpected good outcome, increases positive affect
negative PE: unexpected bad outcome, increases negative affect
Mood changes predict risky decision making
when people are happy, they’re + likely to gamble
PSYC213 Midterm 2
Lecture 10: Memory part 1
[13 février]
10A: Basics
What memory does
Routines and habits
Sense of self
facts we have about self developed from experiences
Social functions
Solving problems
Clive Wearing
Worst case of amnesia ever seen
Episodic memory was impaired
Other memory forms remained intact
semantic memory: knew his wife
procedural memory: played the piano
Distincts memory systems supported by different neural circuits
10B: Stages of memory
The three stages
Encoding: learning new info, forming new “memory trace” as neural code
memory consolidation to storage
Storage: retaining encoded memory trace/neural code
Retrieval: activating a memory trace via a cue for a purpose
memory is in patterns
Neurology
Memory trace is formed (encoding) as a hippocampal-cortical activity pattern
Retrieval triggers pattern completion of the brain pattern
memory cue causes the completion
Memory is transformed into stable cortical pattern in consolidation
Memory as systems
Sensory input
Sensory memory
1 second
info not transferred to STM is lost
Short Term Memory
consolidated with rehearsal
30 seconds
info not transferred to LTM is lost
Long Term Memory
back and forth with STM
Sensory memory systems
Automatic reflections of a sense
Gustatory/olfactory memory
memory of the chemical compounds of taste/smell
smell has very close link to memory
Echoic memory
sound byte held for ~3 seconds
Haptic memory
very brief memory of a touch
Iconic memory
millisecond visual memory
vision persistence/afterimage
Afterimages
Positive afterimage: visual memory that represents the perceived image in same colors
helpful for seeing things smoothly
helps fills in holes in videos (24 f/sec instead of 75 f/sec)
Negative afterimage: visual memory in inverses colors of the perceived image
lasts slightly longer (few seconds) than positive afterimage
Length of sensory memory
Sperling (‘60): participants viewed for 0.5sec a visual display of 3x4 letters, asked to recall them after a tone
whole report: reporterd letters from the whole display
partial report: reported only one row of letters at a time over trials
Partial report conditions remembered + with - delay
STM
Attended info moves from sensory memory to STM
Located in prefrontal cortex
Limited time capacity of 20-30 seconds
Very limited capacity (7 ~+-2)
Serial position effects
Primacy effect: first item remembered +
rehearsal -> LTM
Recency effect: last item remembered +
effect eliminated is delay duration >30sec
10C: Multi-store and modal memory
Chunking
Strategy of grouping items together meaningfully so + info is represented at once
“cat dog pig” instead of “tdi ogg acp”
Increases with knowledge
ex: expert chess players recall + pieces on board than new players if in game, no difference if pieces are random
Working memory
Retention/manipulation of info not in our environment in conscious awareness
Guides behavior
Essential for many cognitive functions
Episodic buffer: integrates info from STM and LTM
“conscious awareness”
Phonological loop
Phonological store: passive store for verbal info
“inner ear”
Articulatory control loop: active rehearsal of verbal info
used to finger written material into sounds
specialized role in language
“inner voice”
Visuospatial sketchpad
Visual cache: info about visual features
Inner scribe: info about spatial location/movement/sequences
Separating stores
Differents brain areas active for visual/break STM tasks
Patient ELD has problems recalling visual/spatial but not verbal info short-term
Patient PV had problems recalling verbal but not visual material short-term
both examples of double dissociation
Implicit/explicit in LTM
Implicit memory: non-declarative, non-conscious
Explicit memory: declarative, conscious
Retention Interval Time between encoding and recall
Ebbinghaus
Learned nonsense syllables and tested me,Roy at various interval to see what was retained/forgotten
created over 2000 cards of fake syllables
Learned sets under strict conditions to remove confounds
read without any inflection
read at a consistent pace of 2.5/sec
did nothing else during experiments
Forgetting curve is exponential
memory loss is largest early on then slows down
Effects
Spacing effect: forgetting is reduced when learning is spread over time
repeated info + valuable
Testing effect: retrieving memory after a test leads to deeper encoding
active rehearsal
participants studied text passage, those who studied + had better ST recall, those tested had better LT recall
Levels of processing theory
Strength of memory (and forgetting potential) depends on processes used in encoding
Shallow: focus on sensory info
ex: memorizing vocabulary words in new language
Deep: integrate higher-level knowledge
things we know with learned info
ex: using new words in a sentence
Memory stronger with deep processing
+ elaborate memory traces
Mnemonics
Use deep processing
Organizational strategies to help encode info
Often involves linking new info to prior knowledge
new info to semantic info
Chunking strategies
ex: acronym to remember lists
Imagery and Method of Loci
use familiar image to link encoded info together
imagery helps memory
Forgetting
Decay theory: memories are lost over time due to disuse
like a muscle, memory not used gets weaker
Interference theory: interference responsible for majority of forgetting
encoded memories are labile, need to be consolidated into stable LTM
in pre-consolidation period, memories susceptible to disruption and interfering info
Effects of interference
Proactive interference: prior info interfere with encoding a new memory
ex: can’t learn new phone number bc of old one
Retroactive interference: newly learned info overwrites/interfere with a priorly encoded memory
ex: can’t remember old password after forming new one
Similarity effect: + alike something is to what’s already learned, the + it’ll mingle and interfere with memory
Lecture 11: Memory part 2
[15 février]
11A: Encoding specificity hypothesis
Basics
Memory retrieval better when there’s overlap in encoding context
Context can act as a retrieval cue
Context can be internal state or external environment
state dependent-learning: better retrieval when you’re in the same internal state (feelings, intoxication, etc)
external context: better retrieval when you’re in the same environment (ex: sea diving vs land experiment)
Transfer appropriate processing
Memory depends on relationship between learning/testing
Highlights importance of encoding context/retrieval cues
11B: Explicit memory
Episodic vs semantic memory
Episodic memory: specific events and episodes, with encoding context
ex: dancing with friends at prom
Semantic memory: facts and general informations, without context of learning
ex: prom happens at the end of secondary school
Episodic memory and hippocampus
Children with hippocampal damage had episodic memory impairment, but normal semantic memory
ex: can’t copy an image after a delay, but have a normal factual knowledge
Means episodic memory depends on hippocampus
this dependency decreases with time
Semantic dementia
Relatively spared at episodic memory tasks
Very impaired at word naming/picture matching task
scored worse than people with Alzheimer’s
Memory and consciousness
Anoetic consciousness: implicit memory, no awareness, no personal engagement
Noetic consciousness: semantic memory, awareness, no personal engagement
Autonoetic consciousness: episodic memory, awareness, personal engagement
called “mental time travel”
Reappearance hypothesis
Episodic memory trace recalled the same at each retrieval
reproduced, not reconstructed
Recurrent memories are unchanged from original events in PTSD
11C: Flashbulb memories
Basics
Vivid memories of significant events
emotionally arousing/shocking events
retrieve specific details about time/place of hearing the event
Flashbulbs over time
People tested about 9/11 recall after 1/6/32 weeks
DV: details used, recollection vividness, confidence
consistent detail and inconsistent details
No detail difference between flashbulbs and everyday memories
Flashbulbs ratings of confidence and vividness increase over time while accuracy decreased
Flashbulbs memories can change
ex: a flashbulb (OJ Simpson trial) recalled after 32 months had 70% inaccuracy and 40% major distortion
Flashbulbs vs episodic
Flashbulbs aren’t recurrent recording of events
Flashbulbs retrieval changes over time, aren’t resistant to distortion, even if memory feels very strong
distinction between subjective/objective memory
must accept theory that memories are reconstructed
Memory consolidation
Experiences are encoded then consolidated into LTM
Consolidation: STM to LTM
Recall: LTM to active memory
Reconsolidation: active memory to LTM
when a trace is activated it becomes de-stable
cortical connections can be strengthened/modified, altering how memory trace is reconsolidated
retrieval changes a memory trace
11D: Distortions in memory
Episodic memory construction
Constructing memories at retrieval mean it can be distorted
May use semantic memory/schemas to infer the way things “must have been” in a recalled memory
insert false info into constructed memory, affecting retrieval
Semantic knowledge affects retrieving detailed memories
Schemas and distortion
Schemas organize/categorize info, provided expectations
War of Ghosts experiment (Bartlett, 1932): people read an unfamiliar Native folk story that don’t match western folk stories schema
they recalled a simplified version of story, became + conventional with each retrieval
omissions/alterations to match Western schemas
Study scene associated to schema-consistent items removed
Distortion effects
False memories: familiar feeling can lead to incorrect associated, details can be added to memories in retrieval
Misattribution effect: retrieving familiar info from wrong source, failure in source monitoring
not remembering where/when accurately
Misinformation effect: leading questions can cause false memory formation
how a question is framed can affect how info is remembered
Rashomon effect: memories are reconstitutions, it’s why people recall the same event differently
the elephant in the dark experiment
Implanting memories
People recalled childhood experiences recounted by their parents over three experimental sessions
A false memory was added to the list by the experimenter
20% of people had a false memory of the event by the end of the third session
Same processes that help us constructed the past help us imagine/plan the future
processes of hippocampal episodic memory
Summary
Episodic memories impacted by prior knowledge
memory driven by our expectations
Episodic memories reconstructed at retrieval
memories aren’t reflections of truth, they’re subject to distortion
Distortions can be false memories, but it reflects an adaptive characteristic
Same processes helping us construct the past help us imagine/plan future
processes of hippocampal episodic memory system
Lecture 12: Memory part 3
[20 février]
12A: Long-term memory
Basics
Explicit memory: semantic (facts) and episodic (events)
Implicit memory: priming, procedural, classical conditioning
priming: neocortex
classical conditioning: emotional responses (amygdala) and skeletal musculature (cerebellum)
procedural memory: skill and habits, basal ganglia
Procedural memory
Automatic behavior/actions
Pattern of movements encoded in the brain
basal ganglia for motor sequence
prefrontal cortex for organization
+ immune to forgetting compared to other types
Habits
First rely on explicit memory, with training/exposure, start relying on implicit memory
motor action sequences
Can form addictions or repetitive thoughts/emotions/OCD
Habit formation: requires the striatum
rats trained in T shape maze with tones to signal reward at left or right end
Breaking habits: need to inhibit prefrontal cortex
removing reward at one end, or making one reward non-rewarding didn’t break the habit
cue -> habit -> reward -> new behavior -> new reward
Priming
Prior exposure facilitates info processing without awareness
Word-fragment completion test
participants shown list of words, asked to complete words fragments
likely to use prior words to complete fragments without knowing it
Implicit emotional responses
Conditioned emotional responses
adaptive fear responses to scary stimuli
Amygdala: implicit emotional memory and modulating explicit memory
emotional enhancement effect
Semantic memory organization
Large units linked to properties with pointers
ex: “animal” unit linked to “bird” unit, which is linked to “feathers” property
general to specific
Spreading activation: automatic activation spread from activated concept to other interconnected aspects
spreading activation to features
ex: thinking about a crow will trigger activation in related bid concepts
Semantic priming: related ideas triggered at retrieval
train of thought may seem nonsensical
12B: Amnesia and dementia
Dissociations in LTM
Experimental neurosurgery to reduce seizure activity
bilateral medial temporal lobe, and hippocampus, removed
caused selective episodic memory loss
HM and the hippocampus
Had the neurosurgery that removed the hippocampus
Intact STM, could remember a short list of words for 30 seconds
Intact procedural memory, could learn new skill based tasks
Intact semantic memory, could recall major childhood event
Profound episodic memory loss, couldn’t learn new info and recalled his past in sparse details
couldn’t remember/describe details of past
couldn’t encode new events
couldn’t imagine the future
Amnesia type
Anterograde amnesia: can’t form new episodic memories
Clive Wearing
Retrograde amnesia: loss of memory from before the amnesia onset
temporally graded, recent memories + affected than old ones
Dissociative amnesia: very rare psychiatric disorder, usually a response to psychological/physical trauma, not injury
commonly retrograde amnesia for episodic memory and autobiographical knowledge
leads to lifestyle shifts link moving, assuming new identity
Dreams and amnesia
Dreams linked to memory
theory that dreaming helps process past experiences
Patients with focal bilateral hippocampal damage and amnesia woken up during night and asked about dreams
damaged reported - dreams, and they were much - detailed
Dementia
Progressive cognitive/functional impairments due to neuronal death
63% of all dementia cases are Alzheimer’s disease
Medial Temporal Lobe regions the first impacted by AD
Earliest symptom is a episodic memory deficit
MCI due to AD (first step): hippocampus and MTL, some memory loss
Mild: lateral/temporal/parietal lobes, poor object recognition/reading/direction sense
Moderate/Alzheimer’s dementia: spread to frontal lobe, poor judgment, impulsivity, short attention
Severe: widespread brain atrophy, language/function loss, basic motor skills
Semantic memory: starts in left anterior temporal lobe, convergence zone for semantic concept representations
deficit recognizing faces/words/uses of objects
12C: Differences in memory
Healthy aging
Volume loss of ~5% per decade after 40
not all areas affected equally
Implicit/semantic memory is impact, episodic/working memory is impaired
ability to remember associations is important for episodic
Cognitive aging theories
Older adults have déifier in general executive cognitive processes from frontal lobe atrophy
slower to process info, can’t inhibit irrelevant info
ex: have trouble focusing on one picture and ignore others
Associative deficit hypothesis
Older adults have problems encoding/retrieving associations in memory due to hippocampal atrophy
familiarity/single item (non-hippocampal) less affected
recollection (hippocampal) + affected
Older adults had + trouble with face-name associative recognition than younger
- trouble in simple face or name recognition
Memory test in scanner showed YAs and low memory old adults both used right PFC, but high memory old adult used bilateral PFC
neural compensation
Taxi drivers and extreme memory
Memory and space intimately linked
taxi drivers memorize 25 000 streets within 10km radius
Taxi drivers performed better on spatial memory tests and had r + posterior hippocampus grey matter volume
volume of posterior hippocampus related to years of experience as a driver
Highly Superior Autobiographical Memory
HSAM can remember every day of their life in detail
enhanced autobiographical memory
recalling very detailed daily memories
Doesn’t involve mnemonic strategies
HSAM people don’t remember a word list better
HSAM speific to personal memories
Consistency in recalling memories related to OCD symptoms
Goldilocks principle: memory words well with just the right amount, not too little or too much
Lecture 13: Concepts and knowledge
[22 février]
13A: Concept meaning and organization
Terminology
Concepts: general knowledge and mental representation of a category
Categories: items grouped to gueule according to concept
Exemplars: individual items in a category
Generalization: deriving a concept from specific experiences
Concept organization
Superordinate, basic and subordinate
super: - precise than the basic
sub: + precise than the basic
Goes from less precise to + precise
Concepts in development and disease
Child learn basic, than super, than sub concepts
Semantic dementia patients can use basic level concepts, impaired as disease progresses
later in illness, use super concepts
Cognitive economy
Organization allows efficient access of knowledge
Use simplest terms still meaningful for the situation
ex: using basic/super for children, sub for experts
13B: Concept learning and representation
Classic approach
Concepts involve forming rules about lists features
defining features: necessary/sufficient for category membership
characteristics features: common, but not essential for category membership
Feature comparison between encountered items and list
refines what a defining feature is for a certain concept
Works well for simple concepts, less for complex/variable/ambiguous
ex: a fur-less dog, a student, etc
Similarities
Concepts are defined by resemblance to a collection of features
not by defined features
Items are +/- part of a category, but an item can be categorized in + than 1
lines can be fuzzy
Prototype theory
Categories formed from overlap of exemplars
extracted from experience
Each category has an abstracted prototypes that’s pre stored in memory
represents most common features with other members
Exemplars included in category network around prototype
similar items stored closer together to the prototype than dissimilar items
People learned to classified variants of a prototype in dot patterns, classified studied/new/prototype patterns
didn’t see the prototype
worse at classifying new patterns than old
equally able to classify prototype and old patterns
Prototype is an abstracted exemplar
other members remember prototype to certain degrees
+ obscure members are further away in network
Typicality effect: preference for processing items close to the prototype
The role of context
Prototype theory: concepts are context independent
don’t account for situation determining concept representation
Context affects typicality effect
faster/easier to recognize a concept if it’s in line with its context/surroundings
Exemplar theory
No single abstract prototype for a concept
every instance of a category is stored, not just prototype
Two steps to see if new item is part of a category
retrieve some/all exemplars of category members
compute similarity to new item at time of concept determination
Explains how context influence concept representation
experience/situational context used to form concepts at retrieval
Knowledge based theories
Explanation view of concept categorization
instead of similarity-based
Implicit intuitive knowledge used to assess new items
Essentialism: certain categories have an underlying reality/true nature that can’t be observed
13C: Embodied cognition and the brain
Embodied view of concepts
Concepts assessed as function of environment/current goals
Processed in different brain networks
shift depending on what’s required to be accessed from a concept
Can bring together dissimilar members into a single temporary category to meet a goal
Embodiment and the brain
Knowledge stored as sensorimotor neural representation
accessed representation as function of what info is required
Perceptual symbols system
Perception/conceptual knowledge as perceptual symbols
Activating a concept engage certain sensory-perceptions to engage mental stimulation as function of task’s goals
highlights importance of perception but also goals in storing/accessing knowledge
Property verification task: people faster to respond if a previous trial asked feature from same sense/percept
concepts represented via senses/perceptions
Brain representation
People passively read action words in MRI scanner
Specific brain regions that process movements associated with the words were active
Concepts rooted in motor/sensory neural activity
Neuropsychological case studies
Brain injury cases of people with category specific deficits
Some have selective impairment in naming living things, some in naming non-living things
ex: some can’t name animals, some can’t name tools
Sensory functional theories
Concepts represented by defining feature
Living things defined by visual features
visual processing regions
Inanimate objects defined by functional features
motor cortical regions
Lecture 14: Language
[27 février]
14A: Basics
Definition
Shared symbolic system for purposeful communication
symbolic: units to reference something else
shared: common among a group
purposeful: to communicate/translate thoughts
Flexibility
Language acts as a high level control system for the mind, allowing to sculpt mental representation of other and own
also sculpted by environment
Morphology/complexity decreases with languages spoken
Lexical tones partly determined by climate
tonal languages spoken + in warmer climates
14B: Language in the brain
Aphasia
Impaired language function, usually from brain injury
Broca’s (non fluent) aphasia: speech halted/difficult to produce, writing usually affected similarly
ranges from deficits in producing certain words to problems generating all forms of language
deficit depends on amount of damage to Broca’s area
expressive aphasia: intact language comprehension, impaired speech production/articulation
Patient Tan: large lesion in Broca’s area, could only say “tan”, tried to communicate with tone/inflection/gesture
Wernicke’s (fluent) aphasia: written/spoken comprehension affected, language content not meaningful/comprehensible
posterior superior temporal lobe damage, mostly left
verbal paraphasia: substituting a word with something semantically related
phonemic/literal paraphasia: swapping/adding speech sound
neologism: using made-up words
Conduction aphasia: impaired repetition, speech production/comprehension intact
load dependent
damage to neural pathway between Broca’s and Wernicke’s
Lateralization
Language often considered left lateralized
Not fully understood or linked to handedness
new data indicate up to 70% of left handed people still have left hemisphere language dominance
Broader aspects of language supported by right hemisphere
prosody, pitch, mood, attitude, gestures, etc
Right side seems important for higher-order non literal language use
right hemisphere lesion disrupt ability to interpret/express speech prosody
14C: Acquisition and comprehension
Nuturist/behaviorist view
Language acquisition is skill/associative learning
explicit training of language
Trial/error reinforcement and modeling other people’s language
Chomksy/naturist view
Language is not stimulus dependent or determined by reinforcement
Language is complex, acquired rapidly
Allows us to understand/speak what we haven’t heard before
Innateness hypothesis: grammar/syntactic structure separate from semantic meaning/cognition
Language Acquisition Device: entity supporting language
Universal Grammar: part of LAD that includes rules for all languages
children only need to learn language specific aspects to our on top of UG
Children exposed to different learning situations converge onto the same grammar
uniformity of (healthy) language development with age
Stimulus argument poverty
Child’s linguistic environment not sufficient to enough to allow the child to learn a language via reinforcement/rules/imitation
child doesn’t have enough language samples to acquire all language, not enough opportunities to learn from mistakes
must be something innate about language
Adult reformulation of child’s speech target the structure, not the meaning
Children extract regulations from experience to form rules
evidence that rules aren’t innate
Psycholinguistics
Buildings blocks of language
Phonemes: smallest linguistic unit (d, o, g)
english has a few dozen for morphemes
Morphemes/words: smallest meaningful units of language
Syntax: rules governing how words are arranged in sentence
Semantics: the meaning
Basics of comprehension
Understanding semantic from language
Resolving types of linguistic ambiguity using context/top-down processing
phonological: within a sound
lexical: within a word
syntactic/parsing: within a sentence
Ambiguity
Phonological ambiguity: determining phonemes depends on audio signal, which is often noisy
can use context/internal knowledge of speech sounds to hear
ex: noisy cafe over your conversation with a friend
Lexical ambiguity: single word can refer to more than one different concept
>80% of English words have more than one dictionary entry
basis of puns
homophones: words that sound the same with different meaning, correct one resolved the sentence context
Cross modal priming task
Word “bug” can mean two things
One group without context and one biasing context to “insect” meaning
Lexical decision shortly (short SOA) after hearing the word bug: both meaning active, bug primes both
Lexical decision with delay (long SOA) from hearing the word bug: into context biased meaning active
Both meanings initially retired, contextually inappropriate one quickly discarded
Parsing
Sentence parsing: diving a sentence into words and identifying them as nouns/articles/verbs/etc
Ambiguity can come bc we hear sentence incrementally or there’s often + than one way to parse a sentence
Garden path sentence: sentences with multiple syntax structures
interpreting a word one way leads to faulty interpretation
Syntax first theory: we use grammar rules to interpret a sentence as we hear/read it
local or specific
we parse with only grammar principles, in one direction, may get to the end and wrong meaning, must go back
Constraint based model: we use non grammatical info to help interest sentences and resolved any ambiguity
global or holistic
use + than grammar to parse sentence
can use semantic/thematic context, expectation, frequency
14D: Language and thought
Linguistic relativity
Language/thoughts are interconnected
Sapir whorf hypothesis: Language changes how we think/perceive
people who speak different languages think differently
ex: Inuits have + words for snow
Linguistic universalists: language/thoughts are independent
Findings
Russian, but not English, discriminate between lighter and darker shades of blue
Russian speaker faster to discriminate blues into divergent categories, English speaker showed not effect
language affects perception
English have more words for color than Indonesian Dani
suggest accessing color category without language labels doesn’t change across language
both groups match/categorize learned color patches the same
Dyslexia
Surface dyslexia: reads letter by letter, sounds out words
difficulty matching words to mental dictionary
impaired at producing irregular words (25% of English)
Phonological dyslexia: reads by comparing whole words to mental dictionary/lexicon
difficulty reading letter by letter and sounding out words
impaired at reading non words and new words
Lecture 15: Bilingualism
[29 février]
15A: History
Traditional story in psycholinguistics
Bilinguals considered a special group of language users
like those with brain damage or language disorders
Thought learning two languages would be confusing
Bilinguals should be functional monolingual in two languages
late bilinguals: learnt second language later than childhood
Even successful late L2 learners spell with accent and fail to get subtle aspects of L2 grammar
New attitude
Publications about bilingualism jumped around 2001
Greater plasticity than previously understood
Language learning occurs at all ages
Languages processes are dynamic
Bilingualism provides lens to understand neurocognitive processes
alters structure/function of mind
Bilinguals not two monolingual in one
15B: Discoveries
Language coactivation
Bilingualism is joint activation of two language systems
cognates: words same in two language, facilitating
interlingual homographs: same word with different meaning, interference
Facilitates/interference effects stronger in L2
and in language from different writing system
also present in bimodal bilinguals
Bilinguals show effect of parable activation in reading
Initial stage of comprehension: first fixation duration (length of 1st time eye fixates on target)
Later stages of comprehension: length of all eye fixations
longer fixation durations -> + comprehension difficulty
Impact of L2 on L1
Bice & Kroll examined cognate effects in monolinguals and L2 Spanish learners with English lexical decision task
Cognates has no behavioral effect
Event-Related Potentials: brain voltage fluctuations time-locked to an event
Cognates has no ERP effect on monolinguals
Cognates in early L2 learners reduces N400 ERP
so native language is affected by L2
Cognitive control
The inhibitory control model: supervisory attentional system affects inhibitory system but not the goal
inhibition: triggered to presence of competition, proportional to coactivation level and prevent intrusion from not used L
N400 ERP shows no effect between L1 and L2
however, behavioral reaction time is affected
N200 ERP associated to early stage of speech planning to operation of inhibitory processes
can be observed as negative shift over front central sites
peak between 250/350 ms after stimulus onset
Strong cognitive demands needed for language control and brain/cognition are plastic and affected by experience
underlies that multi/monolinguals could use their cognitive control processes differently beyond language processing
Lifelong management of two languages -> language control enhancement -> enhancement in domain general control -> brain changes
Flanker task conflict effect very reduced in bilinguals
Bilinguals have greater bilateral frontal gray matter volume compared to monolinguals
Individual differences
Bi/multilingual advantage found in studies
other studies question the differences
Bi/multi advantages modulate but tasks used to assess executive functions
Adaptive inhibition hypothesis: interactional context -> language experience -> language/cognitive processing
Dual language contexts may be particularly demanding
both language control and general cognitive processes
Multilingual langages experiences a continuum
Language diversity across social context crucial to regulate language representation/control/accessibility
over basic measures like L2 age of acquisition and self reported ability
Summary
Bilinguals are not two monolingual in one, both languages are active and competing
two languages not separate
Bilingual language sustek permeable in both directions
L1 changes in response to L2 learning
bilingualism affects both languages
Effects on cognitive control, brain/cognition are plastic
Not all bilinguals are the same
language experiences are multifaceted
bilinguals differ by context demands and experience
Lecture 15: Choice
[12 mars]
15A: Heuristics and biases
Basics
Heuristics: mental shortcuts allowing to skip careful deliberation to draw interference
Slow system: serial logical analysis of info
effortful, non automatic
Fast system: heuristics based reasoning
easy, automatic
Availability heuristic
Estimate event probability based on ease at which it can be brought to mind
Liechtenstein (78) asked which was more deadly between tornadoes and asthma
falsely rated tornadoes as deadlier, bc deaths by tornado is easier to come to mind
Affect heuristic: tendency to overestimate risk of event that creates large emotional response
Sunstein (02): people rated sharks as deadliest animal, especially after being exposed to shark attack media
Representativeness heuristic
Tendency to make inference on basis that small samples resemble the larger population they were drawn on
Related to availability heuristic
relies on pre existing knowledge structures
People base their judgment of group membership based on similarity
results in biases
Base rate neglect: failing to use info about prior probability of an event to judge event likelihood
Conjunction fallacy: false belief that conjunction of two conditions is + likely than either single condition
likelihood of event is always higher than likelihood of that even and something else
Anchoring and adjustment
Anchoring: tendency for people to overweight initial info when making decision
Kahneman and Tversky (74): people saw roulette landing on big or small number, than asked to make an estimate
bigger number on roulette -> bigger estimate
Important to design self report scales
Regression to the mean
When a process is rather random, extreme values will be closer to the mean (- extreme) when measured again
Can’t always attribute changes in performance to manipulation
Bounded rationality
Bounded rational: theory that humans are rational relative to environmental/individual constraints
people are satisficers, we look for good enough
« making do » with our humans limitations
heuristics provide incorrect answers and lead to biases but they also work
Ecological rationality
Sees heuristic as optimal approach to solve problems
Given right environment, heuristic can be better than optimization than other complex strategies
Summary
Heuristics/biases arise from limitations we face
can sometimes produce correct responses (ecological rationality)
Applying heuristics often lead to biases
15B: Decision-making processes
Types of decision making
Perceptual: objective criterion for making choice
Value based: subjective criterion for making choice
depends on motivational state and goal
Decisions under risk: when outcomes are uncertain
ambiguity: when you have incomplete info out the consequences
Risky decision making
Needs to decide even when outcomes are uncertain
Both extremes in risk taking can be very harmful
stagnancy vs impulsivity
Risk premium: difference between expected gains of a risky option and a certain option
Risk averse: decision maker has positive risk premium
most people are risk averse
Risk neutral: decision maker has zero risk premium
no difference in options
Risk seeking: decision maker has negative risk premium
don’t need chance of winning + to take a risk
Classical economy theory: rational thing to do is choose option maximizing expected value
account for individual risk preferences
People are inconsistent in preference taken as bias
we don’t follow expected value
The framing effect
Display of inconsistent risk preferences depending on problem framing (loss vs gain)
people are risk averse when options are described as gains
people are risk seeking when options are described as losses
Gachter (09): PhD students could either get a early discount, a late registration fee, or both to cancel each other out
93% of students signed up early when told they would pay a penalty fee
only 67% signed up early when told they’d get a discount
Endowment effect
Tendency to ascribe higher value to owned objects compared to identical unowned objects
case in which decision framing matters
people are averse to give thing up once there’s ownership
Prospect theory
Psychological theory explaining how people make decisions under uncertainty
two major features: shape of utility function and shape of probability weighting function
Utility: subjective value assigned to an object
context dependent
assigned to monetary amount as function of current state, not in absolute value
losses loom + than gains
Probability weighting: how people understand likelihoods
tend to overestimate rare events and underestimate mundane events
links to availability heuristic
Fourfold pattern: two features together compared to risk
loss-high prob and gain-low prob: risk seeking
loss-low prob and gain-high prob: risk averse
Dual process theory
Broad scientific theory that humans possess two decision making systems
fast/automatic system vs slow/effortful
People gave higher death frequency estimate when in a negative mode
Prediction Error: difference between what you predicted will happen and what actually happened
positive PE: unexpected good outcome, increases positive affect
negative PE: unexpected bad outcome, increases negative affect
Mood changes predict risky decision making
when people are happy, they’re + likely to gamble