PYSC 2150 Intro to Cognition Exam 2
Distraction
1 stimulus interfering with processing of another stimulus (i.e.: conversation in library interferes with cell phone game playing ability) --- Internal: mind drifts away External: environment
3 ways to code in WM
1) Acoustically 2) visually 3) semantically Have >1 thing at a time
Major methods to determine cog functioning & the brain
1) Analyze human/animal behavior after brain damage 2) Recording from single neurons in animals 3) Measuring activity of human brain 4) Recording electrical signals from the human brain
Idea of connectionism
1) Based on how info = repped in the brain 2) Explains may diff findings (how concepts = learned, how damage to the brain affects people's knowledge about concepts, etc.) Also called parallel distributed processing models Propose that concepts = repped by activity that's distributed across a network
3 ways to see how info = held in the brain
1) Brain damage: how damage/removal of prefrontal cortex affects ability to remember for short pds of time 2) Neurons: how neurons in monkey prefrontal cortex hold onto info during brief delay 3) Brain activity: areas of brain that's activated by WM tasks
3 problems with Quillian's model
1) Deoesn't explain typicality effect (predicts same rxn times, which is untrue) 2) Question cognitive econ bc people may store spec properties for that concept right at the node for that concept 3) Sentence verification results suggest that by using semantic network, "pig" is closer to "mammal" than "animal". Incorect
Significant problems with thinking memory tests = sensitive
1) Difficulty of recognition depends on difficulty of distractors - all semantics based = difficult 2) Make distractors similar to target, recognition is much harder 3) Type of cue (good vs shitty) also matters - experiment about cues to memory -cues must match way you encoded content (transfer appropriate processing) --> deep processing != better --> type of processing @ encoding = better for trying to retrieve meaning later. No cues that are absolutely good or bad Incompatible with memory strength & memory pride hypothesis
Problem with prototype model
1) Doesn't address idea of gradients between diff models 2) Can't create a prototype of what can be in your pocket, can still understand the way things work 3) Features = only as good as how you define them --- Ad hoc categories Don't have experience, but prototype model is general, tuned by making categorization jdugments Never don'e something before, now you're making up a def of category after the cast Not categorization judgement you are often Another would be the, "what would you rescue from a fire?" Don't have experience, can still make judgment, contradicts model
Measures of memory
1) Free recall 2) cued recall 3) recognition 4) savings in relearning
Problems with probabilistic models
1) How to select features? 2) Similarity & context 3) Categorization based on rules
Modal model
--> (rehearsal) --> Sensory memory --> STM -(storage)-> LTM --> | <-(retrieval)- LTM - | - output ---- Model proposing 3 types of memory (semantic, STM, LTM) = structural features of the model Places sensory & STM at beginning of the process of memory Also proposed control processes (i.e.: rehearsal)
Maybe they're terrible at encoding
-Focusing on wrong thing? -Get them to encode aloud to circumvent, make sure it's deep encoding - get patients to do deep encoding, no effect
Moving beam of attention has 3 separate processes
1) If attention = engaged, have to disengage 2) Moving attention 3) Re-engaging attention --- if 1 is damaged (posterior parietal cortex) - valid trials = cue points to direction we're looking at -->We can do this even when damaged - invalid trials = really bad on contralateral side -->difficult to disengage attention from the invalid side Subject understands something happened somewhere else. In order to do that, must disengage, which they cannot die If 2 is damaged (superior colliculus) - valid trials: slow to show advantage of the cue, do show it (key = length of the delay) - invalid trials: no benefit/cost with short delays (didn't engage attention). If longer delay & engaged attention, much higher cost Longer delay = good at showing you can engage. No real benefit from cue if delay = brief If 3 is damaged (thalamus) -Valid trials: don't seem to use the cue much bc you don't care about cues. Don't feel like trying. -invalid trials: don't seem to use cue much, easily distracted by invalid stimuli --> Have trouble shining attention on anything (huge bc can't engage attention in location they want it to be. It's engaged where you don't want attention to be) --- Sig bc neural evidence & describing a once thought single process into 3 distinct parts
What's the missing process?
1) Maybe they're terrible at encoding 2) Fast forgetting 3) Stuff gets into memory, just can't retrieve it
Connectionist approach to learning
1) Network responds to stimulus 2) Provided with correct response 3) Modifies weights to better match correct response --- Shows how you learn and what learning looks like Whatever output's closest to that's what you'd think it is At first, all weights = 0. Get input, tell which node should be activated, only activate nodes based on the weights --- Model's good for accounting how you generalize
Multitasking hypotheses
1) People who aren't good at regulating attention, you prefer multitasking (opposite also applies) 2) There's an effect of long-term multitasking, used to spreading things around -other things conflict with what you're doing (opposite = also true)
Properties of memory
1) Quantity and speed 2) Near miss 3) Relevant info 4) Resistant to faulty input
Activation of units in a network require
1) Signal that originates in the input units 2) Connxn weights throughout the network
Several ways to get info into LTM
1) maintenance rehearsal 2) elaborative rehearsal Better memory = associated with encoding based on meaning and making connections
Advantages to connectionism approach
1)Advantages of local rep (typicality) and graceful degradation 2)Works best with brain damage 3) Automatically generalizes - not accounted for in local model - grow to see them as separate with inc experiences
Properties of attention
1)Limited - not all sensory stimuli simultaneously get continued processing (can't fully process all the different sensations) 2) Selective - attention must be selective exactly bc it's limited 3) Informs/enables continued cog processing - can direct attention wherever you want
EXP: Godden
1/2 subjects went underwater, studied words Other 1/2 studied words on land. Later tested on words on land. People who studied on land performed best --- Best recall = when encoding & retrieval happen at same location Should study in same environment as test taking environment
Stem completion
1st part = same as tachistoscopic id 2nd part = give stem completion task. Complete stem so it makes a full word, do whichever word comes into mind 30% of people, when see list, finish stem with list of words
STM vs working memory
2 diff models of the same thing STM = storage WM = storage & thinking --- Working memory includes Central exec, which has VS, episodic buffer, phonological loop, semantics, episodic LTM STM = concerned with storing info for brief pd of time
EXP: Colin Cherry
2 different songs, one in each side of ear at equal volumes. When you receive messages in each year, do an experiment involving dichotic listening. Focus attention on words in one ear = attended ear, as you heard the words, repeat them aloud (shadowing). Next, without shifting attention from the attended ear, notice what you can take in from the other ear (was the speaker male/female? etc) --- Dichotic experiments confirm subjects = unaware of info presented in unattended ear People notice physical characteristics, don't know anything about meaning --- Idea that only attended stuff = processed for semantics
EXP: Hyde & Jenkins
2 factors to manipulate 1) Deep = rate pleasantness 2) shallow= does word contain Q or A? Gave incidental & intentional memory tests. --- 1) Depth has a big effect 2) intent doesn't matter Effort has impact only if it prompts you to do things consequential to memory (therefore indirect effect)
EXP: Craik & Tulving
4 levels of depth Structural (shallowest) --> phonemic --> category --> sentence (deepest) Presented word as a stimulus. Recognized more words as processing became deeper
Prototype
A "typical" member of the category Based on avg of members of a category that are commonly experienced Variations in category = rep diffs in typicality
Saccadic eye movement
A rapid, jerky movement from 1 fixation to the next
Procedural memory
Ability to do anything involving muscle co-ordination (riding a bike, etc.)
What are attention techniques concerned with?
Ability to focus attention on a particular image/task Often shift attention from place to place by moving eyes or shifting attention in your mind without moving your eyes
Cocktail party effect
Ability to focus on 1 stimuli while filtering out other stimulus ---- According to early filter model, this shouldn't be true
Attention
Ability to focus on specific stimuli or locations --- Limited. Can't process everything Is continued cognitive processing
Typicality effect
Ability to judge highly prototypical objects more rapidly Affects how quickly you categorize & how you reason When asked to generate examples of a category, always generate v typical examples
Reading span
Ability to juggle info in WM effectively = sign of higher thinking Switch to attention Manipulating info in WM quickly & effectively = sign of higher reading comprehension
Spreading activation
Activity that spreads out along any link that's connected to an activated node Spreads to other nodes in the network, and add concepts that receive this activation become "primed", can be retrieved more easily from matter --- If damage, you can't think about that node again (it gets completely wiped out)
Exemplars
Actual members of the category that a person has encountered in the past --- As we initially learn about a category, may average exemplars into a prototype. Later in learning, same exemplar info becomes stronger Early in learning, we'd be poor at taking into account "exceptions". Later exemplars for these cases would be added to the category
Retrieval cues
Aid memory Significantly more effective when they're created by the person whose memory is being tested --- Retrieval can be increased by matching conditions at retrieval to conditions that existed at encoding: 1) encoding specificity 2) State dependent learning 3) Transfer appropriate processing
Categories
All possible examples of a particular concept --- "Pointers to knowledge", when you know category, can focus energy on specifying what's special about this particular object Helps us understand behavior we may find baffling
Amnesic & repetition priming
Amnesic show benefits sim to reg patients Stem completion: Amnesic perform normally or terribly depending on the instructions 1) fill blanks with words you saw an hour ago = cued recall -Clive wearing would be terrible at this 2) Fill blanks with 1st word that comes to mind = free recall - shows same results as 30% of us People think of amnesia as memory minus some process, say some memories require a process. If there's a process, amnesic can't do that task. If not, amnesic can
Processing capacity
Amount of info people can handle and sets limit on their ability to process incoming info
Task irrelevant stimulus
An unrelated characteristic that's introduced (dog animation when trying to ID letters, etc). Ability to ignore task-irrelevant stimuli depends on load of the task you're trying to do and how powerful task-irrelevant stimuli is
Attenuator
Analyzes incoming message in terms of physical characteristics (high/low pitch), language (how message groups into words/syllables) & meaning (how sequences of words create meaningful phrases)
Saliency map
Analyzing characteristics such as color, orientation & intensity at each location in scene & combining them. Determining influences of saliency requires a saliency map of the scene (i.e.: something highly salient might be something that contrasts with its surroundings) One way to show where we look != determined only be saliency = checking eye movements Large variations in how people scan scenes
Central vision
Area you're looking at
EXP: Egly
As subjects kept eyes on the plus, one end of rectangle = briefly highlighted = cue signal that shows where target would probs appear. Subjects had to press button where target = presented anywhere on the display. Subjects responded most rapidly when target = where it had been presented. See advantage bc it's located within the object that receives subject's attention --- Example of same object advantage
EXP: Treisman & Schmidt with illusory conjunction
Asked subjects to ID the objects. When gave word bank of objects, illusory conjunctions = less likely to occur --- Subjects knowledge of usual colors of objects influenced their ability to correctly combine the features of each object
Studying central exec
Assess people with brain damage Frontal lobe plays sig role in WM (so people with frontal lobe damage have probs controlling their attention) Typical behavior = preservation (have trouble adjusting behavior when rules change) - breakdown in central exec's ability to control attention
Explicit memory
Associated with consciousness Aware of something you just learned Can directly test you
Default values
Assume something has property unless you're specifically told that it does not have that property Assume certain properties = true (occurs bc you have assumptions with that concept) Some concepts = inhibitory If 1 thing is true, then all other conditions must be false (if I eat meat, I can't also eat bugs and nectar)
Memory
Attending to memory doesn't mean you'll remember it (most of what you attend and perceive you don't remember) Some memories get lost, some = retained much longer --- Process involved in retaining, retrieving & using info about stimuli, images, events, ideas and skills after the orig info's no longer present Active any time some past experience has an effect on the way you think
Selective attention
Attending to one thing while ignoring others --- Animal that never diverts its attention from what it's doing in that moment can die/can't just shut off all distractions (i.e.: fire alarm)
EXP: Ericsson
Avg college kid with avg memory ability could memorize & repeat sequences <= 79 without errors Used chunking --- Interaction bn STM & LTM bc chunks = created from info stored in LTM
Differences between filter models
Awareness in early filter model doesn't always need conscious attention Load theory, if little attention = spent, you can understand things at a deep erlevel (meaning), unconsciouslly If something = cognitively tasking, takes a lot of your attention, you're gonna process things you're not attending to at a lower reservoir Only attend to it through physical characteristics (don't process for meaning) Late vs early is accounted for using load theory & attention
4 versions of implicit memory
Basal ganglia, cerebellum, amygdala, visual cortex
Connectionist network learning
Begins with incorrect connection weights, slowly modified in response to error signals Consists of initially week & undifferentiated activation of property units, with many errors Learn exp --> small change in connection weights, only after many repetitions, until canary = assigned correct properties through network Interleaved with other objects, small changed in connection weights made after many presentations
Point of automaticity
Being able to do more things at once (can read letters & understand meaning at once) Can switch attention between 2 tasks
EXP: Sensory memory (instrospectionist method)
Big bowl of black beans. Would scoop black beans and throw into white box on black tray Glance down, writes estimate, counts actual number of beans Meant to see how much info gets into consciousness simultaneously? --- 100% accuracy <= 5 beans 50% accuracy with 9 beans "Span of apprehension", span estimates = always the same
Continuity errors
Blindness to change in films --- Means we're constantly missing things in the environment bc perceptual system only focuses on a few things at a time
Attention can be influenced by stimulus salience
Bottom-up process: Depends solely on pattern of light & dark, color & contrast in a stimulus i.e.: to find blonde people, respond only to physical property of color without considering the meaning of the image
Plasticity
Brain can change, it''s not static Through experience, can get the brain to change so you can learn things
Patient H.M
Brain damage = circumscribed (know where it is) Developed anterograde amnesia post experimental surgery Him & other anterograde patients don't learn: -words -stories -faces -rhythms Has good short-term memory, something longer than 30 seconds Could spend hours with him, leave for 2 mins, come back and he would never remember meeting you
Bottleneck model
Broadbent's model bc filter restricts information flow Slows down flow of info, prevents large amount of info from going through Lets info in based on specific physical characteristics of that info (rate of speaking, pitch, etc.) Also known as early selection model Benefits: gave testable predictions about selective attention
Emotional condition
CR = an emotion (boat horn/rat to anxiety)
With regards to covert attention, some posit?
Can covertly attend to locations and specific objects. 1) Attention can enhance our response to objects 2) When attention = directed to one place on an object, enhancing effect of that attention spreads to other places on the object --- Don't move eyes, still shift attention
EXP: Badeley
Can remember short string of numbers while reading, etc. Concluded WM must be dynamic, consist of diff components that function separately Proposed: 1) phonological loop 2) visuospatial sketch pad 3) central exec
Clive Wearing
Can't get anything from STM into LTM, both mems = intact (knows english, how to play piano, etc.) Can't get knew experiences into LTM (something wrong with storage mechanism)//applies to people with Alzheimer's as well
EXP: Balirut
Cannot find target when conjunction search = required, can find target with feature search
Subordinate level
Categories = less abstract than basic level Don't operate here bc requires spec knowledge from you and other party
How prototypicality affects behavior
Category -Description -Experimental result --- Family resemblance: -Things in a category resemble each other in a number of ways -Hihger ratings for high-prototypical items when people rate how "good" a member of the category it is Typicality - People react rapidly to members of a category that are "typical" of the category - Faster rxn times to statements with high prototypical items than for low prototypical items Naming - People = more likely to list some objects than others when asked to name objects in a cateogry - high prototypical items = named 1st when people list examples of a category Priming - Presentation of one stimulus affects responses to a stimulus that follows - Faster same-different color judgments for high prototypical items
High typicality
Category member closely resembles category prototype (it like a typical member of the category)
Low typicality
Category member doesn't closely resemble a typical member of the category
Probabilistic/similarity models
Category membership = matter of probability, not all or non Sim bc likelihood of being member of category = calc by computing sim of exemplar to the category/concept
Deep processing
Close attention focused on an item's meaning & relating it to something else Harder --> doesn't always lead to better memory Ex: cake makes me think of birthday
Visuo-spatial sketch pattern
Coding by sight/visual-mental imagery Holds visual & spatial info by creating visual images in the mind in the absence of a physical visual stimulus Attached to central exec --- STM: People identify concrete nouns > abstract nouns May generate vis/mental image when seeing concrete nouns
Chunk
Collection of elements that's strongly associated with other elements in other chunks
Illusory conjunctions
Combos of features from diff stimuli. Can occur even if stimuli differ greatly in shape & size --- Happen bc in pre attentive stage, each feature exists independently of the others ("free floating" features) Can be incorrectly combined if 1+ objects
N-back performance test
Compare N number of things in memory to keep track of
Simple connectionist network in brain
Concepts & their properties = rep in the network by the pattern of activity in these units (= neurons found in brain) Lines = connections that transfer info bn units, equivalent to axons in the brain Some units = activated by stimuli from environment, some = activated by signals received from other units = input units
Semantic network approach
Concepts = arranged in networks Focuses on how categories/concepts = organized in mind
Skeletal conditioning
Conditioned response = overt response to motor system (air puff to eye & blinking)
Phonological similarity effect
Confusion of letters/words that sound similar
Trained network
Connectionist network needs to be trained for result to occur, which involved adjusting network's connection weights
Hierarchical model
Consists of levels arranged so that more specific concepts are at the bottom, more general concepts are at high levels
Dictionary unit
Contains words, stored in memory, each has a threshold for being activated
Masking
Contents of sensory memory = replaced by something else See letters, immediately see squiggles Letters = lost/overwritten/masked by weird symbols Knowing what to pluck out of sensory memory = physical cue
Attention and visual perception
Could you have stimulus on fovea that you're not seeing? Color changing card trick --- Just bc something's in visual field doesn't mean you see it Attn = nec for perception, seen as "perceptual glue" Some visual searches = easy, others = difficult
EXP: Rosh & hierarchical organization
Created 3 levels of categories: Superordinate/global level Basic level Subordinate/specific level Have to list as many features as you can that are common to all/most objects in the category --- Basic level = psychologically special bc going above (to global) --> large loss of info Going below (to specific) --> little gain of info
Connectionist networks & damage
Damage to system doesn't completely disrupt operation Graceful degradation often only happens in actual cases of brain damage in which damage to brain --> partial loss of functioning Studying ways networks respond to damage may suggest strategies for rehab of human patients
Multitasking
Data seems to indicate young > old at multitasking. Not true Young > old at working memory bc working mind peaks at 22 --- High multitaskers have worse attentional control Worse at regulating their attention --- Always carries a cost Switching, not sharing (not doing 2 things at same time)
EXP: Gray
Dear Aunt Jane. Subjects shadow message in one ear, ignore the other. Unattended ear was presented: 3, Aunt, 10. Attended ear was presented: Dear, 8, Jane. Subjects reported hearing "Dear Aunt Jane", not "Dear 8 Jane" --- Implies subject's attention jumped from 1 ear to another. Subjects were taking meaning of words into account = top-down processing. Conflicts with Broadbent's model
Definitional approach to categorization
Decide if something's a member of a category by determining whether a particular object meets the definition of the category Concept = list of necessary & sufficient features, defined --- PROB: For many items, defs don't work bc not all members of everyday categories have the same features
EXP: Bower
Decided to test whether visual imagery can create connections that enhance memory Used paired-associate earning where list of word pairs = presented - first word of each pair is presented, subject's task is to remember the word it was paired with One group silently repeats pair, next group forms mental image where 2 items interact --- People who created images remembered x2 as the other group
Imagery
Def helps memory If you engage vis-mental imagery @ encoding, more likely to put that info into your mind
EXP: Karpicke
Demonstrated advantages of testing through the testing effect
Ability to remember
Dependent on how material = programmed into mind Relating words to something meaningful & important (like survival) enhances memory
Connection weights
Determines show signals sent from 1 unit either inc/dec activity of the next unit Correspond with what happens at a synapse that transmits signals from 1 neuron to another
EXP: Neville Moray with Attention
Dichotic listening experiment where subjects shadow message to one ear, ignore the other ear. When Moray presented listener's name to unattended ear, 1/3 of subjects detected it. --- Info presented in unattended ear = processed enough to give listener some awareness of its meaning
Pop out phenomenon
Didn't need to look for target by end of training, could just rest eyes Not actively searching, felt that target letters "popped out". Only seen after training
Change blindness
Difficulty in detecting changes --- Have to actively hunt for a change (pay attention)
Perceptual load
Difficulty of a task based on high load or low load
EXP: Allen
Diggers and builders Told to memorize/figure out the rule itself After they've been trained, people = told to memorize using similarity, others using rule have 2 mechanisms bc difficult to categorize problems in world with inflexibility Lots of variability out there, which categorization doesn't handle very well
Pursuit rotor
Disk spins, you have to hold wand so that it stays in contact with disk the whole time Measures what % of timed interval you touch disk HM improved at same rate as someone without brain damage
Graceful degredation
Disruption of performance occurs only gradually as parts of the system = damaged
Transfer appropriate processing
Dog and the bamboo example When processes are same at encoding and retrieval, memory will be successful
Incidental memory test
Dunno test is coming
Control processes
Dynamic processes ass with structural features which can be controlled by the person & changed for tasks (strategies you use to memorize info better, strategies of attn to keep you focused, etc.)
Early vs Late filter
EARLY: Stimuli from environment --> processing physical characteristics [FILTERS OUT HERE] --> process semantics --> awareness LATE: Stimuli from environment --> processing physical characteristics --> process semantics [FILTERS OUT HERE] --> awareness --- Make different predictions about how much processing happens to stimuli you don't pay attention to
Fixation
Each time you briefly paused on one face
Attended ear
Ear you pay attention to
EXP: Attention with introspection
Early attention studies relied on introspection. Show a person a display consisting of diff patches of color and instructing them to pay strong attention to middle patch. Describe how attention affects clarity of the patch's color. Key characteristic of introspection: Variable results from person to person Behaviorists didn't focus on attention
Untrained network
Effect of activation on each unit depends on connection weights between the units If all connection weights = same, many (incorrect) units in the network would be activated Only correct concept & relation units should activate correct property unit Achieved through a learning process, occurs when error signals = sent back through network through back propagation to hidden & rep units. Give info on how connection weights should be adjusted so correct property unit = activated
Episodic vs semantic
Episodic: - tagged with spatial/temporal context ("I visited family in India 2 years ago) - Learned in single exposure - autobiographical feeling ("I remember this") Semantic: - Factual ("I know") - Not context (dunno when/where you learned what a faucet does) - Possibly learned in single exposure, not always - strengthened by repetition ---- Another example of diversity where there could be unity
Error signal
Erroneous response in property units
Scripts
Event schemas
Peripheral vision
Everything off to side of what you're looking at
Different kinds of knowledge
Exist in form of concepts, usually organized by categories
Worrying
Exp show that worrying takes up capacity Pressure dec ability to recall (errors inc when pressure ins) Writing down worries can free up WM capacity
Episodic memories
Experiences of past (picnics, long ago personal memories, etc.)
Influences on non-amnesic patients
Explicit mem = ass with awareness, implicit = outside awareness (i.e.: air puffs. Not aware that conditioning is occurring with puffs & tone) Tasks don't purely call on one memory system, task performance usually reps 1+ systems operating simultaneously
"Just in time strategy"
Eye movements occur just before we need the info they will provide
Memory failures
Failures of retrieval. You encoded memory, just can't retrieve it when you need it
Repetition priming & research
Feel separate, it's sep in brain, therefore overall should be separate (logic behind why researchers give up with brain and memory vs motor skills)
Early selection model
Filter eliminates unattended info right at beginning of the flow of information
Final output of Tresisman's model
Final output = determined by 2nd stage, when memory = analyzed by dictionary unit Threshold = smallest signal strength that can barely be detected. Common words, "important" words = easily detected (i.e.: listener's name) Words with low thresholds = easily detected. Words with higher thresholds != easily detected
Feature search
Find target by looking for a single feature (find the single - in the picture)
EXP: Derrick Parkhurst
First few fixations = more likely on highly salient areas. After 1st few fixations, were influenced by top-down processing that depends on observer's goals & expectations determined by past experiences
EXP: Keppel & Underwood
First few trials creates interference, diff to remember letters presented on later trials --- Rapid forgetting = due to interference caused by al the info subjects learned earlier Effective duration of STM, when rehearsal = presented = < 15-20 seconds
EXP: Change detection
Flash 2 scenes, spot the diffs bn each scene People often miss scene changes Also used to determine how much info a person can retain from a briefly flashed stimulus --- Perfect results = when 1-3 squares in array. Performance decreased with 4+ squares
EXP: Sensory memory & semantics
Flash digits & letters, report the vowel Can't point to sensory memory and expect it to know what things are and what they mean Sensory mem knows where things are, where they're stored, can diff by colors of letters, can't do vowels vs consonants --- Sensory memory can't do deeper meaning (pull letters out asap, then evaluate if it's a consonant or a vowel)
EXP: R. Conrad
Flash letters on screen, write them down in order they were presented Errors occurred when letters sounded like another letter Subjects saw the letters, mistakes made on how the letters sounded! --- Supports phonological similarity effect & idea that words = processed in phonological store part of phonological loop
EXP: Flash sentences
Flash sentences on screen, verify if true or false Gorilla = mammal, gorilla = animal Slower with mammal than with animal, model predicts the opposite --- Assume: traversing levels takes more time to verify if concept does/doesn't inherit from above
EXP: Sperling
Flashes numbers, say to report as many as possible Try again, report only 1 of 3 rows depending on tone - representative of what you know of the whole array Immediate report (= sensory memory/v complete representation of field that disappeared. Doesn't last long): gets 33% of the array Immediate tone cue for each row: Subjects get 3 correct (75% of row) Tone cue delayed for 1 second: Subjects get 1 correct (25% of the row) Idea of sensory memory = complete rep of visual field, doesn't last long disappears --- STM registers all/most info in our visual receptors, info decays in less than a second. Important bc reveals capacity (large) and duration (brief) of sensory memory Cued using spatial cue = physical characteristic (where in array is it?). Can use colors, etc.
EXP: Tulving
Free recall, then cued recall Cued recall --> 50% more remembered
Measures - difficulty
Generally, free recall = hardest, then cued recall, then recognition, then savings (which is easiest) Est by giving people material, testing them
Tone cue
Get stuff from sensory memory into a stable condition that won't fall apart They don't know the cue, have to guess which row Avg guess = incorrect, don't get many right
EXP: Hart
Give list of words, remember as many words, give test for words don't remember When free recall fails, recognize answer 50% of the time (recognition = more sensitive overall)
Models & brain structures
Gives testable predictions about brain, not precise structures within the brain Collins & Quillian used sentenced verification technique to measure reaction time, learned that "canary is an animal" took longer to answer than "canary is a bird"
Anatomic separability
Giving up on idea that there are two separate systems (implicit vs explicit) Not one chunk that's explicit and another chunk that's implicit Different things activate different areas. Different types of memory systems must be operating in the brain.
Episodic buffer
Goes in between VS & phonological loop Handles meaning for things. Part of working memory model --- Stores info (inc capacity) and is connected to LTM (allowing for interchange bn WM & LTM) Code in terms of meaning Things = filtered through this buffer. Go bn for other types of rep
Perceptual system
Good at taking in info we need to survive (only a small proportion of available size)
EXP: Datto
Had observer covertly shift attention to diff locations on a display while they measured the activity of his brain in an fMRI scanner --- Idea that attention enhances activity at locations on brain's topographic map that represent where subject = directing attention
EXP: Shiffrin & Schneider
Have target set with bunch of letters flipping very quickly Either 1 or 4 figures on each card Showed the pop out phenomenon
EXP: Morris
Heard sentence, one word = replaced by "blank" Meaning condition: answer yes/no based on the meaning of the word when it filled the blank Rhyming condition: answer yes/no based on the sound of the word Subjects had to process words diff Subject's retrieval performance depended on whether the retrieval task matched the encoding task --- Better performance when the type of processing matches in encoding & retrieval. Also shows that deeper processing at encoding != better retrieval as processing theory predicts
Hierarchical vs connectionist
Hierarchical: reps isa/is/can/has properties at network's nodes Connectionist: Has a "property" unit that contains these characteristics, by pattern of activity in rep & hidden units in the middle of the network When concept = activated, activation spreads along connections so some of the rep units & hidden units = activated Connection weights --> some units to be activated strongly, others weakly --- Concept = rep by pattern of activity in all the units of the network
Types of connection weights
High connection weight --> strong tendency to excite the next unit Low weights --> less excitation Neg weights --> dec excitation or inhibit activation of the receiving unit
EXP: Harrison & Tong
Holding info in visual cortex --- Flashed diff patterns with cue telling people which one to remember 11 sec delay, had to say if lines = rotated to left or right Brain activity = measured with MRI, used neural mind reading Could determine which orientation subjects were holding in their mind during delay --- Info about orientation a person is remembering = being held in visual cortex during delay & visual cortex = involved in WM
Sensory memory
Holds all the incoming info for a fraction of a second. Brief persistence of an image, retention of the effects of sensory stimulation --- Introspectionists = interested in how much info = in consciousness at one time --- Large capacity (up to 20 characters) Physical properties, probs little semantic info Semantic doesn't work --- Important bc shows how much info we can take in immediately, how much of that info remains half a second later
Short term memory
Holds info for 10-15 seconds, then transfers it to LTM
Long term memory
Holds info indefinitely
Modal model: LTM
Holds large amount of info for years or decades
What does Tresisman's model describe?
How an aspect of cognition = conceptualized as a problem of information processing (i.e.: info from environment passes through various stages of processing) Treisman's model = early selection model
Binding problem
How object's individual features become bound together EX: When person sees a red ball roll, cells sensitive to shape, movement, depth, and color fire in diff areas, but you still perceive ball as 1 object
Spreading activation model
How semantic memory = organized Have nodes and links, activation = all or none Activation & links = graded. Links only pass activation Heavy emphasis on activation and activation spreading through complex network Some nodes have v strong activation (activates with other nodes, lots of connecting links) Others have weakly weighted links, passes v weak signal to the next thing To access memory, see what's linked to what --- Model allows content addressable storage, default values, retrieval of properties, explains typicality through default values
How to deal with problems from probabilistic model?
Hybrid models Sometimes use similarity, sometimes use rules Use rules when features don't vary
Property inheritance
I inherit any property about me in a hierarchy, can store lots of info Good at lots of storage, good at relevant info Properties with exceptions need specifications for that spec property
EXP: Primary memory - representation
I say list of words, you say them to me After certain number of trials, performance starts to drop bc getting confused if in this trial or last one Demonstrates proactive interference
Proposal of flashbulb = intro of new mechanism with high emotion
Idea = you have qualitatively diff memory storage process that kicks in with high emotion Untrue, bc emotion has an impact, but no sep process from emotion exists Prone to errors/people's memories change after a period of time
Simons & Chabris
Idea that attention can affect perception within a dynamic scene (gorilla video) --- When observers = attending to one sequence of events, they can fail to notice another event, even when it's right in front of them
Family resemblance
Idea that things in a category resemble another in a number of ways --- Allows for some variation within a category Propose that categorization = based on determining how similar an object is to some standard rep of a category
Filter
Identifies message being attended to based on physical characteristics. Lets only this attended message pass through, everything else = filtered out --- Is movables, moves earlier when thing you attend to requires a lot of attention (not done consciously) If not very demanding, more attention to be spread around
Attracting attention
If faces in figure, spend time looking bc high info, attracts attention Must be meaningful
Recall cues & encoding
If recall cue is better at making you think as you did at encoding that a recognition cue, should be better at cued recall than recognition
Transfer from STM to LTM is impossible
If you can't get new things into LTM, how to explain HM's performance? Memory isn't all one thing
Relevant info
If you don't have answer in mind, memory system still produces some info Answer questions you've never thought of (Aristotle's hands) Gives you relevant info that might produce an answer
Repetition priming
If you experience something a second time, you're biased by 1st experience you had with stimulus Diff can be very wide: Tachisto: better at flash ID stem: see part of stimulus at completing word What I do know = influenced by what I did earlier, esp since it's the same stimulus EX: visual system = changeable so you're better at reading a word, measurable. The visual cortex actually changes --- Proved motor skills = diff from other parts of memory (not new idea)
Near miss
If you make mem error, error != random Noticeably related to what you try to remember "Tip of the tongue"
Factors that affect encoding
Imagery Emotion Repetition Thinking about meaning (depth) Effort/desire to learn (has no real impact on if you remember things)
Cued recall
Implicit contextual cue Experimenter gives extra hints relevant to content (might be semantic info) "Remember stuff I told you. I'll remind you all words started with a vowel", etc. --- Cues can include location, sound, smell, etc. (anything that returns you to same state of mind as when memories initially formed)
Distinctiveness
Important for memory
Similarity & context
In 1 context, things = v diff or v sim By changing a bit, similarity changes Models assume similarity = same, regardless of context
Attention as selection
Includes covert & overt attention Link between eye movements, attention and perception (i.e.: scanning photo to find Waldo) Scanning = necessary because good detail vision only happens for things you're looking at directly
Flashbulb memories
Incorrect idea Happen during v high emotion Resulting memory = v complete, accurate, you'll never forget it Mind makes exact copy of the whole scene bc you don't know what's important, just remember the whole thing
Event related potential
Indication of how much space = used in WM during a certain task
STM
Info that says in memory for brief pos of time (10 - 30 seconds) if we don't repeat it over and over again --- Stores small amounts of info (4-9 items). Most = lost, only some reaches LTM Measure capacity using digit span Describe with experiments dealing with recall to test memory Memory performance = measured as % of stimuli that's remembered, if there's a pattern to how items = recalled
Modal model: semantic memory
Initial stage that holds all incoming info for seconds/fractions of a second
Interpretation of memory & retrograde amnesia
Initially, memories = fragile, must be consolidated after they're encoded Insult to brain occurs Fragile memories = lost (for casual, like falling off a bike, and severe, like electroshock therapy) bike: lose last hour shock therapy: lose some consolidated memories
EXP: Broadbent's filter model of attention
Input [sounds of attended/unattended messages] --> Filter [lets attended message through, filters out unattended] --> Detector [Records info that gets through filter] --> To memory Used tape recorder to run experiments that tested people's ability to take in info under different conditions. Proposed filter model of attention, using results of experiments studying selective attention --- Only thing that makes it through = thing you pay attention to
Resistant to faulty input
Input that goes in I give you mem cues, one of them is wrong You still get to the right answer, even though input cue = incorrect
Proactive interference
Interference that occurs when info that was learned previously interferes with learning new info What I learned earlier = messing up what I'm trying to learn now Coding by sound, but meaning has an impact. As meanings = more distant from proactive interference, get better performance
Intention to learn
Irrelevant
Typicality in spreading activation model
Is robin a bird bid? Bird and robin = more active (& connected) bc bird activates typical properties and typical properties activate robin, which activates robin's typical properties which activates bird, etc. Idea of mutual reinforcement (not seen bn is hawk a bird)
Repetition
Just bc you look at it all the time doesn't mean it stays in your head (penny) Sheer rep != memory for the stimulus Needs to be good rep, then it really helps
Immediate tone
Know which row they have to look at and report Couldn't report all letters bc they rapidly faded as initial letters = being reported
Conceptual knowledge
Knowledge that enables us to recognize objects and events and to make inferences about their properties
Problems with Emotion
Lab studies see if people remember emotional photos better Not good bc photos people say give emotions have other properties Photos that don't illicit emotions = things you see everyday, distinctive things cause attention --- Vary from person to person (same materials, emotions vary)
Hierarchical organization
Larger, more general categories = divided into smaller, more specific categories --- To access other nodes, have to poke around more Properties closely associated with 2 sep ideas = tossed into working memory When node = active, surrounding nodes = also active. Activation doesn't spread through network, only goes to neighbors ---- Have to go up through links to access diff nodes (idea of property inheritance) Model = purely linguistic (doesn't deal with images) PROVEN WRONG bc doesn't account for typicality effect
How connectionist networks are created
Learning process that shapes the networks so info about each concept = contained in the distributed pattern of activity across a number of units Rep = more complex, involves more units for each concept Better model what happens inside brain Can sim normal cog functioning for processes such as language processing, memory, cog development
Working memory (WM)
Limited capacity system for temp storage manipulation of info for complex tasks such as comprehensive learning & reasoning
Shallow processing
Little attention to meaning, focus on physical features of word cake (font, caps, etc.)
Distributed representation in Connectionist model
Local rep Each node reps 1 though One node, one concept Need to look at all nodes and patterns or rep to see what's active --- Concepts = distributed across multiple nodes, can house more information
Content-addressable storage
Located items based on their content Concept = actively in use
Stuff gets into memory, just can't retrieve it
Logical problem: amnesic can retrieve older, long-term memories If retrieval = messed up, why can HM remember his 1st grade teacher?
EXP: Shepard & Metzer
Look at 2 objects, determine if they rep 2 diff views of the same object or 2 diff objects Rxn times inc for greater diffs in orientation, they inferred process of mental rotation --- Example of the operation of VS sketch pad bc it involves visual rotation through space
Load theory of attention
Low-load tasks that use few cognitive resources may leave resources available for processing unattended tasks irrelevant stimuli, whereas high-load tasks that use all of a person's cognitive resources don't leave any resources to process unattended task-irrelevant stimuli (think of the empty vs the full circles) --- When carrying out a hard, high-load task, no processing capacity remains, you're less likely to be distracted. If carrying out easy, low-load task, processing capacity that remains = available to process task-irrelevant stimuli Explains data discrepancy from Forster's experiments.
Hippocampus
Main structure for explicit memory Supports episodic vs semantic memory
Encoding specificity
Matching context of encoding & retrieval
State dependent learning
Matching internal mood present during encoding and retrieval EXP: Eich demonstrated this using music, mood & retrieval. Got the same results
EXP: Hiroyuki Shinoda
Measured observer's fixations and tested ability to detect traffic signs as they drove through driving simulator online More likely to detect stop signs at intersections than in the middle of block 45% fixations occurred close to intersections --- Learning about regularities in environment (stop signs = usually at corners) to determine when and where to look for stop signs
Prototype approach to categorization
Membership in a category is determined by comparing the object to a prototype that represents the category Good for large categories --- Prototype model deals with diff levels by suggesting you have diff prototypes for each level (it's clunky though, esp at superordinate level)
EXP: Hirst
Memories of 9/11 Took 4 surveys months and then years after 9/11 Asked where were you, who were you with, etc? Number of planes, airlines Took confidence measures Problems: rehearsing answers with friends over the years (discussing, telling stories, etc.) --- Inconsistent answers, confidence always stays high (make errors, still confident about their answers) Tell you dif things across time No effect of: residency during that time self-rated emotional intensity of event self-rated media attention self-rated ensuing conversation --- KEY: Memory confidence stayed high, accuracy dropped off with time
Semantic memory
Memories of facts (addresses, days, names of diff objects, etc)
Maintenance rehearsal
Memorize without thinking about meaning or making connections with other info Results in poorer memory
Retrieval & schemas
Memory = always a combo of actual event & relevant prior knowledge Memory = just a construction - good at telling you what probably happened Prior knowledge includes scripts & schemas Things that conflict with schema = attended to and remembered (occurs at encoding) Things that aren't part of schema but are irrelevant are not - not super regimented - everything in script doesn't have to happen - things can be omitted from script without a problem At retrieval, remember parts of script that didn't actually happen bc they're consistent with the script Memory gets mixed in with what actually happened (think event happened, even if it didn't actually happen)
Addressing system
Memory = like a storehouse Organizational system = addressing system (each entry associated with system has a corresponding number) Mind doesn't work this way bc: 1) should take longer to find stuff in mem when you're older 2) Small errors should give huge mistakes, not near misses 3) No "relevant info" if answer = unknown 4) No resistance to faulty input
Quantity & speed
Memory = limitless for humans V quick at retrieving memories
Levels of processing theory
Memory depends on depth of processing that an item receives Distinguishes between shallow vs deep processing Says that deep processing --> better memory
Word length effect
Memory for lists of words = better for short words than for long words
Self-reference effect
Memory improves if you can make it relate to yourself Statements that result in richer, more details reps in a person's mind --> better memory
Memory strength & test sensitivty
Memory strength Depends on how you encode things If deep encoding, memory = v strong Both factors of LTM Memory probe Free recall = insensitive, recognition = sensitive Can't pick up weak signals Measuring fever w hand (free recall - only works if v strong) vs thermometer (recognition) --- Seems like memory tests = sensitive to extent that they give you good cues - free recall: context (time & place) - cued recall: context & partial info - recognition: context & all info These are incorrect
Concepts
Mental representation of a class/individual, and the meaning of objects, events, and abstract ideas --- Provide rules for creating categories & process of categorization Allows you to generalize from prior experience
Anne Tresisman's model
Messages --> Attenuator --> attended message & unattended messages --> Dictionary unit --> to memory Proposed selection = in 2 stages (filter = replaced by attenuator). Attenuator reps a process, not identified with a specific brain structure Analysis of message only went far enough to identify the attended message. Once attended & unattended messages = identified, both pass through attenuator. Attended message emerges at full strength, unattended messages = attenuated (present, weaker than attended message) --- "Leaky filter model" bc at least some of the unattended message gets through attenuator
Late Selection Processing
Messages can be selected at a later stage of processing based primarily on their meaning
Free recall
Minimal info from experimenter Experimenter says remember, and context = usually implied, occasionally described Contextual cue
STM - Model 1
Modal model Largely agreed upon (not v accurate) Lots of stuff goes into sensory, only a bit goes into STM (stays longer if you continually refresh), which lasts for 30 seconds Lots of stuff in STM = eventually lost, most things doesn't make it into STM Duration of LTM = potentially forever/some stuff goes from STM into LTM. Info in LTM = chilling there, gets retrieved when you need it, you get to output it --- Brief sensory memory for visual stimuli (iconic memory) corresponds with do a modal sensory memory stage PROBLEM: too narrow to explain many research findings bc STM = only short-term storage mechanism Since STM = involved with dynamic processes (i.e.: transferring info) --> proposal of working memory WM considers how you hold & process info, which STM & modal model don't take into account
Delayed-response task
Monkey holds info in WM during a delay period --- Monkey sees food in tray, close door, guesses which tray holds food Can bc trained to do this//if prefrontal cortex = removed, pick correct food 1/2 of the time --- Supports idea that prefrontal cortex = important for holding info for brief pds of time
Impact of increased cues
More cues you get, more likely to have overlap with what you though about at encoding Free recall: can generate your own cues Not about # of cues, about probability of you thinking about cue in that context
Typicality categorization
More typical examples come to mind first, you're also better at confirming that they're a member of a category Faster to verify more typical instances
Basic level
Most inclusive, members still share most of their features --- People iD things at this level, indicate categories of objects fastest when category = basic level Not the same for everyone People with more expertise in a certain category tend to focus on more spec info ass with the specific level Ability to categorize = learned from experience//depends on which objects we typically encounter & what characteristics of objects we pay attention to
Late selection models of attention
Most incoming info = processed to the level of meaning before message to be further processed is selected
Multiple memory systems
Motor skill (pursuit roter, mirror tracing = procedural) Repetition priming (gollin figures, stem completion) Skeletal conditioning Emotional conditioning --- Systems designed to perform cognitive function The way they operate changed based on how you see/move Often conflict with each other, multiple systems act simultaneously EX: Afraid of snakes Explicit = this is fine Emotional conditioning = run for your life - conflicting & simultaneous
Diversity where there could be unity
Motor skills may not be same type of learning Develop theory of learning, say motor skills don't fall into this Not huge breakthrough, kinda verifies what you already inferred Done through: 1) Gollin figures 2) Tachistoscopic id 3) Stem completion All examples of repetition priming Amnesics = showing savings form relearning, even though they don't explicitly remember these things
Visual scanning
Movements of eye from one location to another (attempt to identify people across a room, look from one person's face to another)
2 factors that determine how people shift attention
Moving eyes (bottom-up) based primarily on physical characteristics of the stimulus. Top-down: based on cognitive factors such as observer's knowledge about scenes & past experiences with specific stimulus
Automaticity
Need consistency between practice & characteristic of task (same stimulus requires same response from you each time) If you change stimulus/include inconsistency, don't get automaticity --- Leads to fluency V important to our lives Everything would take forever otherwise, wouldn't have experience to look back on as references
EXP: McClelland
Network = presented with number of different concepts and relation statements and activity of units and connection weights bn units = calculated by the computer As learning progressed, patters = adjusted so they look diff/same depending on how sim/dissim categories = to each other
EXP: Quillian
Network with nodes that's connected by links Each node reps a category/concept, concepts = placed in network so related concepts = connected Number of properties = indicted for each concept Says there's an ass in mind bn diff things Proposed hierarchical model
Semantic Newtork Models
Networks of nodes From input, node = in diff state than it was before input
Beam metaphor
Neural evidence shows that damage to different parts of the brain support three separate processes: -Disengage -Move -Engage
Neural mind reading
Neural response to determine what someone perceives/thinks using fMRI, depending on task & nature of stimulus being perceived/remembered
EXP: Funahashi
Neurons = helpful in showing why prefrontal cortex matters with WM --- Recorded neuron activity as monkey completed delayed-response tasks --- Found that neurons that responded only when image = flashed in a particular location, these neurons continued responding during the delay
Retroactive interference
New learning interferes with remembering old learning
Issue with missing processes model
No single memory system which is missing a process Can explain why they're bad at things they're bad at, not why they're good at other things
Temporal gradient
Not all mems from past = equally affected Recent mems = v affected, distant memories not so much (due to consolidation) Ex: blow to the head, get retrograde, lose time up to the accident How to prove? Want a mom that's a tagged, public mom where everyone in culture could (and likely did) encode Famous faces test
Inattentional blindness
Not attending to something clearly visible --- Not paying attention to stimulus can cause us to miss things that's out of our field of view & things we're directly staring at
Preattentive stage
Objects = analyzed into separate features independently Before we're completely fully aware --- Analyzing features at a different level, not fully aware of it (part of us analyzes shape, other part levels, etc.) Combine when we get to focused attention
Differences between central & perceptual vision
Objects in central vision fall on fovea (has better detail vision than peripheral retina) on which the rest of the scene falls Parts include fixation & saccadic eye movement (i.e.: moving eye to observe another face) Even when freely viewing object/scene without searching for a target, still move eyes
Retrograde amnesia
Old memories lost Memories up to that moment = lost, weeks before that moment can be as well New memories encoded normally --- Temporal gradient Memories closer to trauma = more easily forgotten than ones farther in the past
Anterograde amnesia
Old memories mostly intact Memories up to that moment = lost, weeks before that moment can be as well Can't encode new memories --- Everyone with anterograde also has some degree of retrograde What's responsible for getting things into LTM also participates in consolidatation. Some mems = fully/partially consolidated. Won't be consolidated any further with anterograde amnesia
Superordinate level
One level more abstract Don't communicate a whole lot of info Don't use bc basic conveys more info about what you're looking at
Disjunctive search
Only paying attention to one feature of the object = parallel search Number of stimulus don't affect rxn times
High load tasks
Other tasks = difficult, not as well practiced. Use more of a person's processing capacity.
Balint's Syndrome
Parietal lobe = damaged Inability to focus attention long enough on individual objects, difficult to combine features correctly Shows how breakdown in brain can reveal processes that aren't obvious when brain functions normally
Load theory
Part of late filter theories Attention = selective & limited if task = very attention demanding, don't have any more attention to monitor your surroundings --> dec cog fuel to devote to other unattended stimuli If task != very attention demanding, might be able to process things for meaning --> increased cog fuel as resources for stimuli
Conjunctive search
Paying attention to 2+ features of the object = serial search Number of stimulus impact rxn time --- In feature integration theory, can't just look at 1 color map (need to look at sev diff maps and know that they look at the same object) Need to know 2 features occupy same place in space, which requires attention
Divided attention
Paying attention to more than one thing at a time (listening on one conversation while simultaneously playing cell phone game)//distribution of attention among 2+ tasks --- Ability to divide attention depends on several different factors (practice, difficulty of the task, etc)
Code auditorily
People say to themselves to remember auditory, not visually, even though it's presented auditorily (change the code) Sound based code to keep info in memory
When people shift their attention
People shift their attention from one place to another as they're doing things and carrying out specific tasks Timing of when people look at specific places = determined by sequence of actions involved in the task Example: Patter of eye movements when someone makes a peanut butter sandwich (shift attention depending on if you're picking up knife or scooping out peanut butter) ---- Person's eye movements = determined primarily by the task. Person fixated on few objects not directly related to task, eye movement usually preceded a motor action by a fraction of a second
EXP: Wood
Phase2: Shock = associated with 3 city names (NY, SF, Dallas) Use galvanic skin response --> more nervous = sweatier palms Phase 2: Person shadows 1 ear, but in unattended hears a buch of city names including one of the 3 participants showed sweaty palms 38% of the time --- Unattended material = evaluated for the meaning Opposite for the unconscious response
Physically salient
Physical characteristic where object = only one in visual field that has it Some type of singular distinction
Stimulus salience
Physical properties of the stimulus (color, contrast, movement)
EXP: Lane & Finch
Present a cross stimulus, viewer IDs which arm flashed (repeat x5) Time 6: small square outline = added to display, only 10% reported seeing the square
EXP: Rensink
Presented 1 pic followed by a blank field followed by same pic but with missing item Alternated until observers could determine differences (had to do several times) --- Importance of attention in determining change blindness = shown when Rensink added a cue indicating which part of a scene had been changed, subjects detected changes much more quickly --- Lack of attention affects perception
EXP: Rosch
Presented subjects with a category title and a list of ~50 members of the category Rate how well they fit into category on a 7 pt scale Demonstrated diffs bn good/bad examples of a category When item's characteristics have large amt of overlap with other items in a category, family resemblance of these items = high Little overlap = little family resemblance --- Strong relationship bn family resemblance & prototypicality
EXP: Kathleen O'Craven
Presented subjects with displays where four & face = superimposed, asked to direct attention to one stimulus or the other (moving or stationary object) She measured activity in certain areas of face & brain. Activity in parts of brain changed depending if it was a moving/stationary house or a moving/stationary face --- Attention caused brain to take possession of the object being attended by increased activity in area specialized to perceive that object.
EXP: Tulving & levels of processing
Presented words to subjects, asked about physical features of the word, rhyming & fill-in the blanks Meant to create diff levels of processing (physical features = shallow, rhyming = deeper, blanks = deepest) Results: deeper processing --> greater memory --- Diff to determine what depth of processing is Can't define in way that's independent of memory test
Precuing procedure
Presenting a cue where test stimulus will appear enhances processing of the target stimulus.
Dichotic listening
Presenting different stimuli to the left & right ears
Priming
Primer will facilitate the subject's response to a stimulus if it has some of the info needed to respond to the stimulus
Memory without awareness
Priming Get priming without awareness = implicit memory If aware, get explicit memory, and performance completely changes
Shadowing
Procedure of repeating what you are hearing
Binding
Process by which features such as color, form, motion and location = combined to create our perception of a coherent object --- Attention = important determinant of what we perceive, brings things to our awareness, can enhance ability to price & to respond & to create binding
Consolidation
Process by which memories become more stable even if not practiced Takes years to be fully consolidated
Categorization
Process by which things = placed in categories Image --> compared to other concepts --> matched to a certain concept --> make inferences about the image (its internal properties, characteristics, etc.)
Encoding
Process of acquiring info & storing it into LTM --- Phase of an experiment when info might go into LTM (time when you think about something that might go into LTM)
Retrieval
Process of transferring info from LTM to working memory
Detector
Processes info from attended message to determine high-level processes of message (meaning). Processes all the info that enters it
Automatic processing
Processing that happens: 1) without intention 2) at cost of only a few cognitive resources --- Many trials of processing require this (going to Kroger, get to stoplight, can't remember going down the hill)
How to access hierarchical model?
Properties of starting point lead you to answer Enter model from wherever cue/environment prompts you to enter (can be from animal or bird node, work your way down to the correct concept from there)
Feature Integration Theory
Proposes 2 step theory to address binding problem (pre attentive stage & focused attention stage) Most involves bottom-up processing (some exceptions) --- Perception happens by tearing apart diff visual features Have parts of cortex that know about orientations, shapes, diff colors, etc Uses disjunctive search (by inspecting activity on map, sees if there's a particular characteristic out there)
Attentional capture
Rapid shifting of attention caused by certain stimulus such as loud noise, bright light, sudden movement, etc. Examples include interception of eavesdropping by noise of overturn book cart When attention due to stimulus science --> involuntary shift in attention Important if it's a warning of something dangerous coming towards us
EXP: Baddeley & articulatory suppression
Read list, recall as many words Read another list, say "the, the, the" out loud as you read --- Reduced ability to remember a list of words, eliminates word length effect Both short & long words = lost from phono store
EXP: Bransford
Read passage that makes no sense. If pic accompanied passage, people understood it better x2 --- Preventing organization from happening reduces ability to remember
Attention as "binder" of experience
Red circle, green triangle How to know which is which? Without attn, can't determine this. Uses conjunctive & disjunctive search --- As number of stimuli in array increase, reaction time also increases
Articulatory control process
Rehearsal that keeps items in phonological store from decaying
Elaborative rehearsal
Remember info by considering meaning or making connections to other information Better memory than maintenance rehearsal
Memory system uses organization to access info
Remembering words in a category might be a retrieval cue Recall list > orig list in terms of organization
How to deal with external distractions?
Remove distracting object from where you are, or remove self Drown it out with something meaningless
Schemas
Rep of gen type of event, characteristics generally true of the event, not of a specific event
Preservation
Repeatedly performing the same action or thought, even if it's not achieving the desired goal
Articulatory suppression
Repetition of an irrelevant sound Disrupts normal operation of phonological loop, happens when someone's prevented from rehearsing items to be remembered by repeating an irrelevant sound Redueces memory bc speaking interferes with rehearsal
Immediate report
Reporting as many letters immediately. As you report, sensory memory = decaying/falling apart. Couldn't see/remember the other letters as they were reporting them
Divided attention experiments
Require subject to carry out 2 tasks simultaneously. 1) holding info about target stimuli in memory 2) paying attention of a series of "distractor" stimuli & determining if at least one of target stimuli = present among distracting stimuli
Relationship bn encoding & retrieval
Respond diff to diff memory cues Mismatch bn what you think at encoding & retrieval, demonstrates "bank" doesn't work
Persistence of vision
Retention of the perception of light in your mind (i.e.: playing with sparklers, looks like there's a trail of light) Continued perception of a perceived visual stimulus even after it's no longer present Isn't obvious in everyday experience when objects = present for long pos, noticeable for brief stimuli
Mental rotation
Rotate an image of one of the objects in their mind
EXP: Peterson & measuring duration of STM
Say a bunch of letters than a number. When you hear a number, count by 3. When I say recall, tell me the last 3 letters before I said the number -counting for 3 seconds: remembered 80% of letter group - counting for 18 sec: remembered 12% of letter group --- Occurred bc memory trace vanished bc of delay from passage of time after hearing the letters
Conjunction search
Search for combo of 2+ features in same stimulus (find - line that is also green) --- Requires scanning to focus attention at a specific location
EXP: Rogers
See question, answer yes or no if word answered the question When testing recall, subjects remembered more words if they're descriptions of you instead of just the size, etc. --- Why? Bc the words = linked to something the subjects know very well - themselves
Explanation of filter model
Sensory memory transfers all the info to the filter through physical characteristics (speaker's tone of voice, pitch, taking speed, accent, etc.) Passes info to detector which is eventually sent to the short term memory. Then transferred to the long term memory
EXP: Vogel
Separate people without brain damage into groups of high & low working memory, tested using change detection procedure Given cue telling them where to direct attention, shown memory display, then had to indicate if test display = same orientation as memory display Recorded event-related potential as they were making this decision Added blue bar as a distraction --> small effect on response of high-capacity group, large effect on other group --- Same WM across the groups High working memory = good at ignoring distractions, central exec was functioning well for these subjects Some people's central exec are better at allocating attention than others More likely to perform well on reading & reasoning ability tests, IQ tests, etc.
Overt attention
Shifting attention from one place to another by moving eyes
Covert attention
Shifting attention from place to place while keeping eyes stationary (i.e.: sports - look to the left, actually pass to the right)
Attention maps
Show how directing attention to a specific area of space activates a specific area of the brain Could predict "secret" place observer was staring at without telling experiment the location
EXP: Treisman & Schmidt
Showed 4 objects flanked by 2 blacks numbers Subjects told to report black numbers first, then repeat what 4 shapes they saw were Many reported objects = combinations of 2 diff stimulus (illusory conjunctions) --- Features appear independently of each other early in perceptual process
EXP: Della Sala
Showed small-large patterns, with half squares shaded in. Subjects could complete patterns with avg of 9 shaded squares before making mistakes
Implicit memory
Shows importance of neuroscientific evidence Generated through study of amnesic patients --- Revealed through indirect testing Perform task, memory shown by performance (goes off of motor skills & priming) Show I remember by demonstrating the task, but don't say that I'm testing your memory
Back propagation
Signals = sent backward in the network starting from the property unit
How to select features?
Sim depends on context At times, do seem to categorize via rules (as def view suggest) Made up arbitrary set of features, which affects sim/dissim, has an enormous impact on if they're the same
Connectionist networks can explain generalization of learning
Sim to how we learn bc learning about canaries enables us to predict properties of diff types of birds we've never seen Since sim concepts have sim patterns, training a system to recognize the properties of one concept also gives info about other, related concepts Limits to what connectionist networks can explain
Amygdala
Site of plasticity with emotional conditioning
Stroop Effect
Situation in which task-irrelevant stimuli = difficult to ignore (can ignore conversation in 1515, fire alarm would attract my attention)
Cerebellum
Skeletal condition Did eye puff
EXP: Cahill
Slide show about kid visit dad (surgeon at work) Kid got there when dad = operating on victims of a car accident Kid got there when dad practices drill of emergency procedures that included fake ops on realistic dolls Heart rate goes up, responds to what you think is real 2 weeks later, were tested again Emotional parts = remembered --- Emotion leads to a significant boost
Chunking
Small units (words, etc) can be combined into larger, meaningful words (phrases, sentences, paragraphs, etc.) --- Chunking in terms of meaning --> inc ability to hold info in STM Allows limited STM to deal with large amt of info involved in many tasks we perform everyday
Categorization based on rules
Some concepts = sensitive to rules Sim: take object 3 inches in diameter. Similar to a quarter or pizza? Categorization: obj = 3 inches in diameter. Is it a pizza or a quarter? There = some features not free to vary that people are sensitive to --- Low sim, high diagnosticity Some features don't happen v often, but when it happens, good sign that object = member of a category EX: drunk people
Advantages to Connectionist distributed rep
Some number of nodes gets you many more reps
Reduction in demands with practice
Some tasks = automatic (little/no attentional cost, with or without intention, not affected by motivation, can happen without awareness), others = controlled (attentional cost, occurs only with intention, affected by motivation, only with awareness) --- Reading letters = automatic for a v experienced reader/little attentional cost i.e.: Stroop effect (supposed to say green, read blue instead) Can go on autopilot for reading, driving, writing, eating, playing musical instruments, etc.
Low load tasks
Some tasks = easy, well-practiced, have low perceptual loads. Use small amount of person's processing capacity
Visual search
Something you do when we look for an object among a number of other objects (where's waldo) --- Can study role of attention through visual search
Bottom up selection
Sometimes things attract attn (color in black & white scale, movement in still frame, etc.) --- Something in environment causes us to think it's important, maybe we should give it some attention
EXP: Spelke, Hirst, Neisser
Sought automaticity for complex tasks Subjects read stories while taking dictation. Initially, huge interference (reading & writing another store someone reads to you) compared to when they're doing just one task Developed automaticity by doing this several times a week Weren't as good as doing single task, still v good//getting good at switching & knowing when to switch attention 1 task = self-paced (reading), the other = paced. When they talk and pause, shift back to reading 1-2 sentences from reading
Phonological loop
Sound based code to keep info in memory/stores things auditorily 2 parts: 1) Phonological store stores 2 seconds worth of aud information (not 4 items). Capacity in terms of time Info can enter store from environment 2) Articulatory control process Process of talking to yourself Another way of putting info into STM ----- Both = attached to central exec Holds verbal & auditory info
Echoic memory
Sounds also persist in mind only lasts for a few seconds after presentation of orig stimulus (when you ask "what?" but then hear it in your mind as soon as you ask)
Topographic map
Spatial map of visual stimuli in the visual context --- Each point on visual stimulus corresponds to activity at a specific location on visual cortex Points next to each other on stimulus cause activity at points next to each other on the visual cortex
EXP: Stroop Effect
Specify the color of the ink of a word not what color the word spells out. Difficult to do. Happens bc names of the words --> competing response, and therefore yields slow response to target. In Stroop Effect, task-irrelevant stimuli = very powerful because reading words = highly practiced & so automatic that's difficult to ignore
EXP: Dot example
Start with dot patterns = 2 prototypes Create exemplars by randomly moving some dots in diff directions & distance = family resemblance (no perfect rep, every exemplar shares some of the same characteristics) --- Prototype experiment: Deck of cards with exemplars A & B Study till correctly categorize 1 week later, categorization test for old, new & prototype Subjects = as good on prototype (which they've never seen before) as on old items, confident about answers
How do you know you're attending to wrong thing?
Stimuli from environment --> processing physical characteristics (loudness, pitch) --> process semantics (meaning, category) --> awareness (see a fox & its physical characteristics/combine the last 2 steps)
EXP: X or N?
Stimulus flashes on screen very briefly. 1) Have an N, put a bunch of o's around the N - fast & accurate response (this is easy) 2) Have an N, put a bunch of other letters around it - slow & inaccurate response (this is hard) Next put a cartoon graphic under. for 1: longer time, for 2: also longer time, not as great as a time diff as there was for 1 --- Fits with idea of load as the primary task Trying to focus attention on primary task. Easy: attn = spread around, captures attention Diff: less attn to monitor environment
Predictions about phonological loop
Store lasts 2 seconds --> people who can talk fast have larger capacity in WM Since store = 2 sec, everyone has larger capacity for short words & smaller capacity for long words (remember more short words than long words) If words sound alike, you'll get them mixed up If you busy articulators, articulatory phonological store can't put anything into store. Can't recode bc articulators = busy, find another way to code (meaning, group them together, visual mental image, etc) --- These effects disappear if you're not coding in sound
LTM
Stores info for long pos of time (minutes to lifetimes) Responsible for episodic, procedural, and semantic memories
Cognitive economy
Storing shared properties just once at a higher-level node Makes network more efficient, can add necessary exceptions at lower nodes
EXP: Forster & Lavie
Studied role of processing capacity & perceptual load in determining distractions. Subjects respond as quickly as possible when they ID a target (X or N), press the appropriate key Easy when n = surrounded by same letters, difficult when surrounded by a bunch of different letters Hard task --> longer reaction times Easy task --> faster except when task-irrelevant stimulus is introduced. Then, slower response for easy task than for hard task
Fast forgetting
Stuff gets in fine, disappears v quickly Can't look at how slowly/rapidly they forget bc they forget immediately (can't record) Reg get .25 seconds, amnesic get 15 seconds to look at pic - make sure it's deep encoding - recall @ 5 mins = control - forgetting happens bn 2 groups in parallel, doesn't account for what's going on
Distractor
Stuff you don't need to remember
Target
Stuff you're supposed to remember
EXP: Schneider & Shiffrin
Subject = shown memory set with 4 characters (target stimuli), followed by rapid presentation of 20 "test frames" with distractors (1/2 trials, frames had target stimulus from memory set) >600 trials later, went from memorizing memory set --> automatically completing task (subjects didn't have to consciously think before responding) --- Practice made it possible for subjects to divide their attention to deal with all of target's test items simultaneously Demonstrated for certain situations, if task difficulty = increased and change targets/distractors for each trial, automatic processing doesn't happen, even with practice (i.e.: driving & texting)
EXP: Morris, Bransford, Franks
Subject has 32 sentences, 16 rhyme & 16 meaning --> 24 hours delay 16 subjects take test on all 32 words with rhyming cues 16 subjects take standard recognition test on all 32 words When tested, those who learned in meaning had better scores when tested on meaning. Those who learned in rhyme and given rhyme cue had better scores than those who learned in recognition. --- Can't say deep processing works across the board Shows that transfer-appropriate processing matters
Savings in relearning
Subject learns material to a criterion, number of trials required = noted Later subject relearns the material to the same criterion, if fewer trials = required, that's savings Relearning occurs much faster than it did the first time around
EXP: Donald McKay
Subject listened to an ambiguous sentence with multiple interpretations in attended ear. Biasing words = presented to unattended ear, later presented with pairs of sentences, indicate which sentence = closer to the one they heard --- Meaning of the biasing word affect subject's choice even though they were unaware of biasing words Proposed word must have been processed to the level of meaning even though it was unattended
Recall
Subjects = presented with stimuli, after a delay asked to report back as many of the stimuli as possible
EXP: Strayer & William Johnson
Subjects = sim, hit brakes as quickly as possible in response to a red light Talking on phone --> 2x errors of people not talking on cell phone --- Talking on phone uses cognitive resources that would otherwise be free
EXP: Neville Moray
Subjects = unaware of a word that has been repeated 35 times in the unattended ear (known as the cocktail party effect)
EXP: Rosch and protoypicality
Subjects heard the primer = name of a color 2 seconds later, saw a pair of colors side by side, indicated if 2 colors = same/diff Side by side: colors = same, good examples of category Colors = same but bad examples of category Colors = diff with 2 colors from diff categories Priming --> faster "same" judgements for prototypical (good) colors than non prototypical (poor) colors Happens bc they hear "green", imagine a highly prototypical green --- Showed prototypical members of category = more affected by priming than non prototypical members Subjects create images of prototypes in response to color names
EXP: Michael Posner
Subjects kept eyes stationary, always looking at t in display. See an arrow cue that indicates where to focus attention while keeping eyes on t (example of covert attention). Press key as rapidly as possible when target square = presented to side Presented valid cues (direct attention to left, signal is on left) and invalid cues (direct attention to left, signal is on right) --- Subjects reacted to square more rapidly when attention = focused on location where signal is about to appear. Therefore, information processing is more effective at place where attention is directed --> attention = zoom that improves processing when directed towards a particular location --- Used precueing procedure to answer if paying attention to a location --> inc ability to respond to stimuli presented there --- BASICALLY: Beam metaphor Attention = "spotlight" that enhances perceptual processing Independent of eye movements Metaphor = only partly correct
EXP: Li
Subjects looked at T on fixation screen, saw central stimulus (array of 5 letters where all were same or 1 = diff from other 4) Immediately followed by peripheral stimulus at corner of the screen Central task = indicate if all letters in central stimulus = same/peripheral task = what corner image was --- Possible to take in info about some objects but not others in scene, even when attention is focused elsewhere
Lexical decision task
Subjects read stimuli that are/aren't words They indicate if each entry is/isn't a word
EXP: Baddeley
Subjects remember 77% short words, 60% long words (demonstrating word length effect) --- Occurs bc takes longer to rehearse long words & to produce them during recall Also proposed that people who speak faster can remember more
EXP: Shepherd & Recognition memory
Subjects see 512 words, take 2-choice recognition test See 612 Sentences (brief), 2 choice recognition test Get 88% on both tests
Recognition failure of recallable words
Subjects see pair of words on at a time, memorize words in all caps Get 2 tests 1) Recognition test: circle words you recognize 2) cued recall test: see lowercase, just write down as many you remember Key finding: fail to recognize some words, get the cued word correct - bias people to think about odd aspect of a common word when paired with lowercase letters ---- Cue = better hint to though you had at encoding rather than uppercase letter alone
EXP: Jenkins
Subjects spontaneously organize info as they recall them
EXP: Slameka
Subjects study words by reading in group of related words or by filling in words related to the first group (generate group) Read group < generate group at reproducing word pairs --- Displays generation effect: Generating material vs passively receiving it --> enhanced learning & retention bc giving connections bn material to be remembered & other material in memory
Gollin figures
Taken pic of object, erased lines, guess what actual object is, give more complete figures over time Writing down which image you identify object by With inc trials, can ID with less complete versions
Same object advantage
The faster responding that occurs when enhancement spreads within an object
Digit span
The number of digits a person can remember
Thinking about meaning (depth)
Think about how it relates to things you already know Deep vs shallow Degrees of depth
Tachistoscopic ID
Time 1: see a list of common words, asked to do something with them Time 2: hour later, done 3 other tasks. Didn't make any ref to 1st task Flash word v quickly, adjust time so you're getting 50-60% of words correct When I should you world on list, should ID it better Memory component bc showing some benefit to a word
Episodic memories turn into semantic memories
Time and place keeps changing, info stays the same, can't remember time & place of when you learned it As number of sim events occur, strength of memory increases. As strength of memory increases, more likely that memory = semantic
Recognition
To be remembered info = presented, along with other stuff (distractors) & subject must distinguish new from old
EXP: Multitasking
Told you'll see letter/digit combo Do diff acts for each letter depending on quadrant, press button --- Shows you can't make multitask v well, even with simple rules. Multitasking = switching, can't keep multiple rules in mind
Scene schemas
Top-down processing = also influenced by scene schemas. Observers knowledge about what's contained in typical scenes influence how people direct attention --- People look longer at things that seem out of place. Attention = affected by knowledge of what's usually found in scene
Mirror tracing
Trace around a star as quickly as you can, get timed Mirror perpendicular to table, see hand in mirror (can't see actual hand though) and trace looking at mirror With repeated practice, get better & much faster. HM performed like a typical 30 year old on this task
Input units
Units activated by stimuli from environment or stimuli presented by the experimenter Send signals to hidden units, which send signals to output units --- Stimulus presented to the input units = repped by the pattern of activity that's distributed across the other units
EXP: Alvarez
Used change detected used by Luck and Vogel (colored squares & complex stimuli) Would have to say if 2 scenes with diff number of cubes = same or not Basically, subject's ability to make same/diff judgement = dependent on complexity of stimuli --- Greater amount of info in an image, the fewer items can be held in visual STM
EXP: Meyer
Used idea of lexical decision task Showed 2 words stacked above each other, subjects press yes if both = words, no if atleast one = nonword Key variable = ass bn pairs of real worlds (closely/loosely ass) Rxm time = faster when 2 words = ass --- May have occurred bc retrieving 1 word from memory triggered spread of activation to other nearby locations in network More activation spreads to words that = related, response to related words = faster than response to unrelated words
EXP: Edward Smith
Used sentence verification technique to see how rapidly people could answer q's about an object's category Subjects responded faster for objects high in prototypicality than objects low in prototypicality, they demonstrated typicality effect --- Shows connections bn prototypicality and behavior
EXP: Brooks
Visualizing F & its corners, pointing when it's an outside (inside corner or saying out or in) Pointing = more dif bc pointing & holding image = VS sketch pad tasks, becomes overloaded, while speaking uses phono loop which doesn't interfere with VS sketch pad --- Exp demonstrates how interference can affect the operation of the VS sketch pad
Top-down selection
Wanna pay attention to this (force yourself to pay attention to something bc it's important) --- Overt attention (moving your eyes) & visual search Unlike bottom up: 1) Have specific thing in mind you want to attend to (not usually attending to it) 2) Searching by color = relatively quick Single perceptual characteristic that help you ID stuff (finding blueberry in the snow, etc) --- Searching for meaning = difficult to do (red & white stripes in a pool of red & white - where's waldo)
Switching
Way of dealing with limited attention --- Can't practice and get better at it. It's about working memory Always carries a cost
What do these experiments focus on with regards to selective attention
What types of info = used for selection (physical characteristics of meaning, etc) No single answer to what's early & late (early selection = demonstrated under same conditions, later selection under others, depending on observer's task & types of stimuli presented) Idea that ability to selectively attend to a task can depend on distracting stimulus & on nature of task being studied
Basic level category = psychologically privileged
What we use when we usually think Use in language, faster to verify on object name at basic level than other levels Levels of categorization don't mesh well with rules
Role of prior knowledge in retrieval
What you think about at encoding = v important for future memory & retrieval MC: Need to know distractors & testers, while in essays just need to integrate info from other sources Whether you remember depends on match bn retrieval & how you thought about things at encoding. What you already know about topic = also important for encoding Memory is more than remembering. It's also about knowledge. Infer this most likely happened, fill in gaps bc memory = fallible
Focused attention stage
When "free floating" features = combined & we can perceive the object --- Observer's attention plays important role in combining features to create the perception of whole objects Focusing attention specifically eliminates illusory conjunction
Prototypicality
When asked to list as many objects in a category as possible, tend to list more prototypical members of the category first (robins come before penguin) Prototypical objects = more affected by priming Follows closely in time
Basal ganglia
When motor skill learning/procedure occurs Important for movement/control
EXP: Nelson
When recognition fails, subjects relearn the material faster than new material
EXP: Brown & Kulik
When something happened in culture (assassination, 9/11, etc.) they'd ask how it felt and for spec details. Decades later would ask the same question --- 8 years after event, people reported a lot of details and were v confident about answers In realty, recalled details that have nothing to do with event
Sensory store
When something's presented briefly, perception continues for a fraction of a second (can perceive movies with this)
Fan effect
When you don't know much about an area, only have one link to activate When you know a lot about princeton, have many ass facts with concept People with lots of knowledge = slower to verify there's a uni than someone with less knowledge --- Only seen in lab, bc IRL you have lots of connections linking concepts with princeton uni and town of princeton (this road leads you to uni, etc.) Activation starts to go to nodes, then flies around bn nodes bc nodes = connected to nodes --- Fan effect predicts experts = slower to produce knowledge than novices. Not true bc of all the interconnections bn the nodes In lab, can teach you facts so that nodes don't interact, not true in practice
How a prototype model works
When you see new exemplar, compare to the prototype. More sim to prototype --> more confident you are that it belongs to that category When you see prototype dot pattern, you're v confident bc perfectly matches prototype in memory --- Judge sim to other prototypes you have in memory, tweak what you have in memory a bit
Central executive
Where major work of WM occurs (control center) Pulls info from LTM, co-ordinates activity of phono loop & VS sketch pad by focusing of spec parts of a task & deciding how to divide attn bn diff tasks Helps you ignore messages from outside so you can focus your attn Attn controller: focuses attn, splits it bn tasks, switches it bn tasks ---- Tells other components what to do WM = sustained activation of reps Maintains info in WM stores/important as coordinator of other cog processes --> not well specified/developed
How many windows in your house example
Why do your eyes turn away? 1) Avoid distracting stimuli -Faces = socially important, hard not to attend to them 2) Habituate -Attention demanding now, after a while it won't bug you -More effective for stimuli that don't change much (harder to habituate to music than to refrigerator hum) -complexity != important as long as it's simple
Central characteristic of attention
Withdrawing from some things in order to deal effectively with others. Doesn't capture diversity of phenomena
Modal model: STM
holds 5-7 items for 15-20 seconds
Exemplar approach to categorization
involves many examples Takes into account the wide variation among items that belong to a particular category Involves determining if an object = similar to other objects using exemplars Explains Rosch's results about typicality Same idea as family resemblance ("better" objects have high family resemblance) Advantage: Uses real examples Takes into account individ cases --> doesn't discard info that might be useful later deals better with variable categories Good for small categories
Intentional memory tests
know test is coming
Rehearsal
repeating a stimulus over and over again (repeating telephone number to self)