Ch 11: Language

Question 1

anomia

Answer 1

inability to find words to label things in the world
no impairment in object knowledge i.e. knew what it was could not produce words

• related to the tip of the tongue effect

Question 2

areas important in language

Answer 2

mostly left lateralized!!

• Sylvian fissure
  Broca’s areea for speech production
  Wernicke’s for comprehension
  angular gyrus for memory

Question 3

aphasia

Answer 3

broad term for collective deficits in language comprehension and production that accompany neurological damage

Question 4

Broca’s aphasia

Answer 4

patients cna have various problems when speak or try to comprehen or repeat linguistic input provided

a) speech production is slow, effortful, lacks function words (resembles telegram)
b) accompanying problems w speech articulation bc of deficits in reg of the articulatory apparatus e.g. tongue muscles
c) hard time understanding reversible sentences

Question 5

Wernicke’s aphasia:

Answer 5

disorder in language comprehension
- damage to surrounding tissue or white matter causes it

Question 6

Wernicke proposed damage to arcuate fasciculus (connecting Broca and Wernickes areas) can result in
conduction aphasia:

Answer 6

can understand word they see/hear and can hear own speech errors but cannot repair them

Question 7

Lichtheim region (apart from Broca and Wernicke’s)

Answer 7

• that stored conceptual information about words
• once word retrieved from word storage, → sent to concept area, supplied all info that was assoc w word
• model does not correspond to what we now know but neuroanatomy nevertheless influential

Question 8

two properties of spoken language:

Answer 8

meaning and phonological form

Question 9

mental lexicon:

Answer 9

store of info about words including semantic info (words meaning), syntactic info (how words combine to form sentences) and details of word forms (spelling and sound patterns)

Question 10

Organization of Mental Lexicon

Answer 10

1. Morpheme: smallest meaningful unit of a language
2. Frequently accessed words are more quickly responded to than less freq accessed ones
3. Phoneme: smallest unit in sound that makes a difference in meaning
4. Representations in the mental lexicon are organized according to semantic relationships

Question 11

Collins and Loftus (1975) semantic model

Answer 11

strength of the connection and distance between the nodes determined by semantic or associative relations between the words (car → truck)
*still debated

Question 12

**semantic maps by Huth et al 2016

Answer 12

**E.x. panel a reveals major dimensions separating perceptual & physical categories (green, turq, blue, brown) from human-related categories (red, purple, pink)

Question 13

Wernicke’s patients show semantic paraphasias

Answer 13

use the word horse when meaning cow
- supports idea mental lexiocn contains semantic networks of words having related meanings clustered together

Question 14

**semantic category storage models

Answer 14

Elizabeth Warrington work → words stored and separated by biological vs human-made categories in the brain

Caramazza challenged this → semantic network organized along conceptual categories of animayc and inanimacy

Tyler et al. → examined patients & controls, processing of living vs nonliving

Question 15

**One model: living things → repped by many features, whereas nonliving → can be repped by only a few features that are distinct

Answer 15

**Patients w aphasia → harder to identify distinctive features for living things vs non-living objects

Question 16

brain areas for naing tiger at less complex domain level and naming same object at specific level

Answer 16

→ activity restricted to posterior occipitotemporal sites (green bars)

→ assoc w posterior occipitotemporal and anteromedial temporal lobes

Question 17

**see p79 and redraw this diagram

Answer 17

input can be spoken or written, then both are processed further up
inputs enter via auditory (spoken) or visual (written) modalities
bottom up flow in diagram → perceptual identification to “higher level” word meaning and activation
interactive models of lang → top-down influences to play a role
e.g. activation at word-form level influences earlier perceptual processes
i.e. the diagram would be bidirectional

Question 18

segmentation problem:

Answer 18

how do we segment all the auditory sounds we hear in diff words?

Question 19

prosodic info

Answer 19

what the listener can derive from speech rhythm and pitch of speaker’s voice

Question 20

**Anne Cutler et al. → English listeners use syllables that carry an accent

Answer 20

e.g. “lettuce” stresses first syllable (heard as single word)
“invests” stresses last syllable (heard as two words)

Question 21

neural substrates of sound perception

Answer 21

superior temporal cortex important for sound perception (red spot)
- aka auditory association cortex

acoustic sensitivity (gen sounds) and speech sensitivity (speech sounds)
- acoustic sensitivity decreases moving anteriorly, inferiorly and posteriorly from primary cortex, while speech sensitivity increases

Question 22

**panel b: left hemisphere network most involved in comprehension of spoken words

Answer 22

centered on: posterior middle temporal gyrus (PMTG), linking to middle temporal gyrus (MTG), inferior temporal gyrus (ITG), and posterior cingulate cortex (PCC)

• ITG also directly linked with PCC

Question 23

how can words be symbolized? (3)

Answer 23

alphabetic, syllabic or logographically

Question 24

Selfridge’s first computational model of letter recognition:

Answer 24

image --> feature --> cogn --> decision (IFCD) Selfridge’s model consists of: Image demons: rep raw visual input e.g. letter image feature demons: detect specific features like vertical lines, curves, or angles Cognitive demons: repr specific letters and “shout” if their features are present Decision demon: listens to shouting and selects loudest cognitive demon, identifying the letter key features of the model: bottom-up, feed-forward relies hevaily on feature detection simple and intuitive, explains letter recogn based on visual components ** model cannot explain all phenomena (e.g. top-down processing) and doesn’t allow for parallel letter processing

Question 25

McClelland & Rumelhart developed a new computational model for visual letter recogn

Answer 25

three layers: 1) features, 2) letters, 3) representation of words - allows top-down info (from higher cognitive levels) to influence earlier processes happening at lower lvls of representation - also explains bottom-up processing - allows for parallel processing (letters processed at same time) - excellently explains real life phenomena (word superiority effect: words recognized better than non-words) → words not perceived letter by letter

Question 26

brain regions for viewing letter strings and faces

Answer 26

faces → activated region of lateral fusiform gyrus (yellow) letter strings → region of occipitotemporal sulcus (pink; a.k.a. visual word form area (VWFA))

Question 27

alexia

Answer 27

condition whereby patients cannot read words though other aspects of language are normal damage to VWFA

Question 28

**DTI → identify projections of white matter tracts from VWFA and fusiform face area (FFA) to other brain ares

Answer 28

**VWFA more strongly connected w left perisylvian language system areas **FFA more connected to medial temporal and occipital areas

Question 29

3 classes of models to explain word comprehension:

Answer 29

1. modular models - lang comprehension executed within separate and independent modules → higher level reps cannot influence lower-level ones i.e. flow is bottom-up 2. interactive models - all types of info can participate in word reocgn → context can have influence even before sensory info available 3. hybrid models (between modular and interactive models) - lexical access is autonomous and not influenced by higher level info - but lexical selection can be influenced by sensory and higher-level contextual info

Question 30

syntactic parsing

Answer 30

when brain assigns a syntactic structure to words in a sentence - higher order semantic processing important to determine right sense of word meaning in context of a sentence

Question 31

N400 ERP and semantic processing

Answer 31

Van Berkum et al → N400 elicited for words inconsistent with meaning of entire story P560 observed suggesting N400 really about semantic violations and not surprise ERP N400 in mild aphasia and healthy controls → normal N400 in more severe aphasia → reduced and delayed N400 fMRI → N400 likely elicited from left temporal cortex

Question 32

What is P600?

Answer 32

ERPs from Pz, a midline parietal scalp site, in response to each word of sentence that’s syntactically anomalous (dashed waveform) vs syntactically correct (solid waveform) in anomalous → pos shift (shaded) emerges in waveform 600ms after syntactic violation → called syntcactic positive shift (SPS) or P600 see p 87

Question 33

what brain area is activated when processing complex syntactical structures?

Answer 33

left inferior prefrontal cortex - PET activations in anterior portion of superior temporal gyrus (STG) related to syntactic processing thus network of inferior frontal gyrus and superior temporal cortex involved in synaptic processing

Question 34

language comprehension model Hagoort

Answer 34

three elements: memory, unification, and control memory (yellow) in left temporal lobe unficiation (blue) in left inferior frontal gyrus control (pink) in lateral and medial frontal cortex

Question 35

Memory

Answer 35

inguistic knowledge encoded and consolidated in neocortical memory systems - knowledge about building blocks of lang (e.g. phono-, morphological, syntactic units) is domain specific and coded diff from visual feature or object information

Question 36

Unification

Answer 36

he integration of lexically retrieved phonological, semantic, and syntactic info into an overall rep of whole utterance unification can occur in parallel

Question 37

Control

Answer 37

relates lang to social interactions and joint action (e.g. bilingualism and in taking turns during a convo)

Question 38

Friederici → proposed network of lang Networks of the left-hemisphere Perisylvian Language System

Answer 38

ventral pathways important for comprehending words dorsal pathway connect to premotor cortex (PMC) involved in speech prep other dorsal pathway connects Broca’s area (specificlaly BA44) w superior temporal gyrus (STG) and superior temporal sulcus (STS) involved in syntactic processing

Question 39

*Levelt model of speech production*

Answer 39

word production: stages of conceptual prep, lexical selection, morphological and phonological encoding, phonetic encoding (phon rep into motor output), and articulation speakers monitor own speech by using comprehension system see diagram p91 Example see a goat pic and need to indicate what it is concept of goat activated (but also similar concepts) activated concepts → activate rep in mental lexicon (also to access syntactic info e.g. noun not verb) lexical selection when syntactic properties of word retrieved morphological encoding (pluralizing morpheme s is added: goats) modified word contain phonological and metrical info (info about syllable # and stress pattern) e.g. goat single syllable, stressed phonological encoding → ensures phonol info mapped onto metrical info [serial model: output rep is used as input for next stage]

Question 40

**evo origins of lang unknown

Answer 40

**human lang possibly began w gestures (important for other species) **Washoe (chimpanzee) learned form of sign lang (approx 350 signs!)

Question 41

Kanzi (chimp) able to match 378 lexigrams to many things and spoken words they represent

Answer 41

able to combine lexigrams intro protosyntax and freely uses keyboard for his demands generalize a specific ref (e.g. lexigram for bread to mean all breads, incl tortillas) understands signs in symbolic way

Question 42

**Left Hemispheric Dominance and Specialization

Answer 42

humans have robust interconnectivity of inferior frontal cortex, inferior parietal cortex, and temporal cortex, esp lateral and inferior temporal cortex by massive projections of arcuate fasciculus white matter tracts of chimp and macaques show greatly reduced or absent projections into lateral and inferior temporal cortex show that during evo, humans were dramatic changes in connectivity → rise to human lang in all its complexity

Question 43

Key Take Home Points

Answer 43

Language relies on distributed left-hemisphere networks Broca’s areas → speech prod Wernicke’s area → compr white matter tracts (e.g. arcuate fasciculus) connects regions into integrated lang system Mental lexicon is structured and multi-layered words stored w semantic, syntactic, and form-based info semantic relationships, freq, and category structure all influence how quickly and accurately we access words Spoken and written language use different entry routes but converge on shared meaning systems auditory inputs rely on hierarchical processing in superior temporal regions, while reading involves VWFA (visual word form area) before linking into lexical and semantic networks Language comprehension is shaped by both bottom-up input and top-down context evidence from segmentation, semantic priming, and N400/P600 components → contextual and syntactic info infl how we interpret words and sentences Syntactic processing depends on a frontal-temporal network inferior frontal gyrus (Broca’s region) and superior temporal cortex jointly support understanding grammatical structure, w deficits arising when either part of network is damaged Language production is a multi-stage process models such as Levelt → speech emerges from interacting conceptual, lexical, phenol, and motor planning processes w internal feedback loops Human language capacity emerged from evolutionary changes in brain connectivity comparative estudies → humans devved dramatically expanded fronto-temporal white-matter pathways (e..g arcuate fasciculus) → rich symbolic comm, syntax, and flexible lang far beyon other species