Week 8 - language Flashcards
Broca’s Aphasia
Broca (1861) – patient Tan
Autopsy -> impairment inferior frontal gyrus
Second patient with similar symptoms & site of damage
Localised to left area (1865)
Good understanding
Weak production
what is noticed in the patient (video)
Output of language diminished
Difficulty with grammar
Understanding relatively preserved
Recent work on Broca’s Aphasia
Understanding isn’t perfect; complex grammar is difficult
Injury localized to BA -> temporary aphasia (Dronkers et al., 2000)
BA is only one part of larger, more complex system
Long-lasting (6+ months) aphasia – lesions in neighbouring areas
BA = centre for syntactic processing
Processing syntax – what does it mean?
Syntax = structure of language to create meaning
‘Units’ of meaning
How do we recognize & understand speech?
Cohort Model
TRACE Model
Cohort model
Speech signal perceived -> /k/
Cohort of potential words activated ->
Cat?
Caterpillar?
Catapult?
Capture?
Comb?
Keyring?
More info available -> /kat/
Some competitors are eliminated
Cat?
Caterpillar?
Catapult?
NOT -
Capture?
Comb?
Keyring?
Precise recognition point!
Uniqueness point
Cohort model - phases
3 phases in language processing
- A word in the cohort is activated
- A word is selected from the cohort (sound)
- A word is integrated based on its semantic & syntactic properties (wider meaning using context)
Phases 1 & 2 bottom up, phase 3 top down
Revised model: context plays no role in stages 1 & 2
Re-revised model: integration happens continuously
TRACE model
- Bottom up and top down processes interact flexibly
- Individual and interconnected ‘nodes’ at 3 levels of speech: phonemes, features, words
- Connections between levels operate in both ways and are facilitatory
within levels = inhibitory - Nodes influence each other on level of activation and interconnectedness
- As activation spreads -> a trace / pattern emerges. Activated words compete to inhibit each other. Most activated -> wins
- words are recognised incrementally
Visual World paradigm
Participants hear beaker
Eye-tracking shows where they look
Predict:
1. Object named most fix.
2. Competitor – speaker – more fix. than unrelated – pram.
3. Competitor sharing first phoneme – beetle – more fix. than sharing last – speaker.
Visual world paradigm
Cohort model predicts
Beaker competes with beetle
Beetle declines once second consonant of Beaker is heard
Visual world paradigm
TRACE model predicts
Beaker activates a rhyme – speaker
As well as sharing initial segment with beetle
Interpreting the results
Competition from beetle was more important than speaker -> evidence for Cohort model
BUT both cohorts & rhymes compete for attention
Ppts more likely to fix on either a cohort or rhyme than non-competitor -> evidence for both models
Cohort activation more rapid & higher peak than rhyme -> evidence for TRACE
Results provide evidence that speech input is continually mapped onto potential representations over time -> evidence for TRACE
BUT caution that this is evidence for continual processing, rather than TRACE, as some assumptions were not supported
Can the models explain Broca’s aphasia?
Greater rhyme competition in BA carrot / parrot (TRACE model)
No difference for cohort-competition beetle / beaker
Wernicke’s Aphasia patient
speech is fluent
meaning is impaired
comprehension deficits
Wernicke’s Aphasia
Wernicke (1874) – two patients with comprehension difficulties, despite fluent speech
Autopsy -> posterior left supratemporal gyrus
Incorporated findings with Broca’s
Importance of arcuate fasciculus (band of fibres connecting Brocas area and Wernickes area)
Recent work on Wernicke’s Aphasia
Injury localized to WA -> temporary aphasia (Dronkers & Baldo, 2010)
Long-lasting aphasia – posterior middle temporal gyrus & underlying white matter
Densely interconnected with other brain areas – lesions disrupt widespread processes
Can the models explain Wernicke’s aphasia?
No difference in rhyme competition in WA carrot / parrot (TRACE model)
Larger cohort-competition effect beetle / beaker
Dissociation
Pattern of results on same task different for BA vs. WA
Broca’s = rhyme competition, but not cohort!
Wernicke’s = cohort competition, but not rhyme!!
Interpreting the findings
Two different impairments = different symptoms
OR single dimension, opposing patterns of disruption
Argue for impairment in response selectivity (rather than perception)
BA – response selectivity is reduced
Slow speech = difficulty selecting words for production
WA – response selectivity is increased
Chaotic speech = overabundance of words selected
Evaluation of the models
TRACE - advantages
Assumes bottom-up & top-down processes contribute to word recognition
Copes well with ‘noisy’ input
TRACE disadvantages
Allows too many theoretical inputs – can explain anything
Conceptual meaning can be activated earlier than model assumes
Cohort advantages
Accurate perception includes processing of competitor words
Processing is sequential & changes throughout presentation
Context can influence earlier than integration stage, esp. if input is noisy
When cohort contains many words, those with ‘high imaginability’ are selected first (semantic importance)
What’s the point of language?
Language = communication – vital for survival
Usually considered unique to humans
But non-human species have innate call systems & system of sounds based on learning & experience
Non-human primates:
Chimpanzees show similarly lateralized & enlarged Wernicke’s area (Holloway et al., 1998)
Capable of learning ASL
BUT Broca’s area 7x larger, more connections, greater activation in humans (Fitch et al., 2018)
Language in non-human primates
Washoe -
Spontaneously combined signs
Acquired a preferred word-order: SVO
Spontaneously taught signs to other primates
Meerkats
Highly sociable & vocal
Alarm calls, contact calls, social calls
Do dogs have language?
Fast mapping
Initial connection between word & referent
= partial understanding
With experience, understanding increases
Humans (especially toddlers!) – one new word / waking hour
Dogs do this too!
Direct teaching
Associative learning
Learning by
exclusion
Collie – over 1000
words!
Comparing human & non-human language
Language study hampered by:
Focus on structured aspect of language – grammar
Assumption that animal language is simple – codes based on environmental triggers
Methods used are not comparable
