Silvia Part 2 Flashcards
Semantic memory background
Meanings linked to concepts and world knowledge. Can think about concepts as a collection of features but doubt everything’s we know about something is stored in one area/temporal lobe. Disputes triangle
Cog theories of conceptual reps - hierarchical
Collins & quinlan hierarchical models: start w superordinate category like animals, then get more spec as you go down like birds, then ind at bottom. Things in each category have the same attributed. Evidence: time to respond whether something belongs to category or not takes longer the more categories away it is. Or prototype model, hierarchical can’t explain exception like ostriches can’t fly-doesn’t fit typical features of bird category.
Prototype model
Each concept is a Mental construct, they are graded in how similar they are to the prototype e.g. degrees of similarity, reason for longer rt e.g. in birds, swallow closer than penguin. Need categories to erected about use
Embodies concepts
Concepts grounded to body perceptions and action e.g. word jug stim mental sim of jug, activates planning of things you do with it. Action verbs activate mirror neurones- in scanner, hearing verbs activates motor neurones to do them yourself, somatotropic- even stronger is if see someone do action, you do it in head
Models in the brain
Network organised around sensory motor function legal think about what it looks like via visual cortex, activate motor cortex for action, auditory for sounds. Object id is in visual cortex, ventral part. Diff regions for animal shapes vs object as object tactile but animals less so- senses become attached to representations
Category spec impairments
Before imaging, ppl claimed spec areas related to spec concepts e.g. p JRB: had stroke, couldn’t name living things but could non living, other ps show opposite pattern e.g. couldn’t give info about features when given the na,e- therefore could just be about features e.g. loving things known for sensory perceptual properties but non living more functional so has actions. JRB impairment can be explained like hierarchy like start w damage, selective damage, sensory, knowledge reps?
SD and herpes part 1
Affects anterioir temp lobe- leads to non category specific semantic impairments, inconsistent w distributed view. SD have progressive loss of meaning, damage greater on L. Can’t recognise or understand words, objects, sounds, touch. Not affected initially are grammar, pronounciation, episodic, spatial , EF. Lose low level distinctive features first e.g. know chicken but lose ostrich so say chicken, layer say all animals are chickens, reflects temp as diff sections for more abstract concepts.
SD part 2
When draw animals after 5s delay, lose distinctive features e.g. all become blobs w four legs and bends even if frog. Highly familiar and frequently used concepts lost later, age of acquisition affects as things learnt earlier remembered better. Mayberry: ps make more errors when naming pictures from less typical for category than more, say substitutions like butterfly for bird as overgeneralise- links to prototype theories
Explaining SD
Temp pole forms modality independent hub where features of meaning are brought together from different stores (spokes)- hub and spoke model e..g multi model integration, disputes JB? Explains distributed and convergent nature of concepts but not semantic decisions and context dependent lang
Semantic processing
We think about concepts in a context e.g. normally thin piano cool but when asked to lift, think about its weight. Selection of relevant features may require different areas. Studies like fam fortune e..g say more averag eanswer or pointless, say most unique answer
Background
Have associative knowledge in the brain when you see objects but not enough to account for meaning processing so need to bring together in the anterioir temporal lobe. Conceptual processing studies have to select competing alt based on context# suggests other hubs
LIFG and PMTG
Thompson 97: ps had to generate action based on object. Easy was simple link like scissors where only one option but hard was something more abstract like wheel. When ppl perform actions, more work in condition w more alt so more activity in brocas and pmtg- always responds to word gen/concepts. No activity in anterioir as noise from scanner and close to bone
Ambiguity and the oifg and pmtg and difficulty
E.g. asked links to word bowel- could be object or action, when as, ppl it’s 50/50 so both strongly linked, means more activationhappens less in context, if stroke in this area, hard to interpret ambiguous words . E.g. ask which word more related to top one, weakly related means more activity than strong- have to work harder to find link. Also more activity for verbs over nouns as more complex/diff situations-action more than word
Convergence
Have distributed areas like motion, action, sound, shape but see when using concept to do specific task- use the two regions so called convergent zones or processing regions. Receive features from other areas and do manipulations in different contexts. Don’t see in fmri as too slow/no temp detail. But involves anterioir temp lobe, borax as, and medial temp gyrus
Combining words into sentences
In sentence comp: need to select relevant features in context of other words e.g. the large wine glass- brining together concepts in one object, same for actions. Lifg and pmtg also used in comprehension, also anterioir temp. When look at related motor meaning there’s activity in motor cortex, more activity , same time as lifg and pmtg. For speech production- more activity in dorsal motor and lifg pmtg (findings words, putting them together and pronouncing them). Hu
Humphreys and gennari 2014:
Given easy and hard to complete sentences like the movie that the director and the director that the movie. See more activity in harder as links meaning, processing and production
Disorders of semantic processing- aggrammatism - lifg damage from stroke - background
Ps can’t create grammatical sentence, have halted speech of individual words, takes lots of effort to produce. It’s the ability to convert thoughts to sentences. Fine understanding but struggle to build sentence around action verb
Agrammatism tests
Bishop 82: test for the reception of grammar. Shown image like girl chasing horse, have issues in comp only when subtle diffs in meaning e.g. right answer picture for girl chases horse as active but if swap to the girl is being chased by the horse (passive) then get wrong image . Often fine in real life as get meanings of words from context
Sentence comprehension
Semantic knowledge helps interpretation so hard when can’t talk from word order or general knowledge e.g. do best in control sentence girl is kicking green ball, then semantically constrained- wagon that the horse is pulling is green. Then reversible the cat that the dog is chasing is black. Worst at semantically improbable - the lion that the baby is scaring is yellow. Issues w relationships between words and verbs, functions like is, by, the, inflectional endings like Ed and ing, base understanding on ind words
Atypical lang networks - sign language
In ppl born deaf, left hemi assumes diff roles based on nature of input and output. Compared deaf and control in scanner, if comprehending signs compared to pantomime, see activity in auditory cortex as well as visual, also lifg. For controls, when hearing spoken sentences compared to meaningless words, see activity in auditory and lifg- lifg role resembles hearing people’s
Reading network in blind
Visual cortex used for reading braille, same as visual word form area. From birth have most plasticity, but light blind similar to controls. Listening in born blind activates posterior visual areas despite intact auditory- late blind don’t. Have links to words and sounds to occipital
Innate vs expereince
Can reorganise lang system from examples but not completely as left lateralisation for right handed stays and lifg plays similar role and damage leads to similar issues. Network changes from sensory expereince is more posterior. Basic dorsal and ventral paths remain (constraint) and only sensory parts redeployed. So part genes part expereince
