Test 3 Flashcards
“compensatory or functional communication” target
day-to day- communication abilities
“compensatory or functional communication” goal
improving functional communication and quality of life> reducing impairment
“impairment-based” or “restorative” target
language impairment/processes (phonological)
“impairment-based” or “restorative” goal
generalisation beyond trained items/tasks, and to the communicative environment of the person
- reducing impairment> improving success of communication
Compensatory approach Pros
- focus on functional strategies for daily life
- easy for client to see impact
- easier for clinician to administer
- fit into limited number of sessions
compensatory approach cons
- functional strategies often limited in their overall impact on impairment
Impairment-based pros
greater improvement on overall language processes (generalisation)
impairment-based cons
- less immediate impact for client
- more difficult for clinician to administer
- may require more treatment sessions (potentially exceeding what insurance will cover)
generalization
the transfer of treatment gains to untreated items and every day, non-clinical environments
- ultimate treatment goal
Neuroplasticity
- use, improve, or lose it
- specificity rebuilds targeted networks
- salience is essential
- repetition and intensity promote learning and consolidation
- promote generalisation; avoid interference
- complexity enhances learning and generalisation
Models of language processing
- classic model of language processing
- modern models of language processing
Classic Model hypothesis
brain mechanisms are involved in word access/storage
Wernicke-Geschwind- repetition of spoken words
auditory cortex-> Wernickes area-> arcuate fasciculus-> Brocades area-> motor cortex
Wernicke-Geschwind- reading aloud of written text
visual signals-> angular gyrus-> Wernicke’s area-> arcuate fasciculus-> Broca’s area-> motor cortex
Strengths of Wernicke-Geschwind model
- aid framework for modern models
- started our thinking about pathways involved in comprehension and production of speech and language
- comprehension: acoustic-> concept
- spontaneous speech: concept-> motor
Problems with Wernicke-Geschwind model
- too simplistic
- only models lexical processing, not language processing
- many aphasias don’t fall into these categories
- single lesion sites don’t correspond to single deficits
functional anatomical model
- model (typical speaker) of two auditory speech processing networks, connected to BOTH motor and conceptual systems
functional anatomical model- dorsal stream
mapping sound onto sub-lexical and articulatory-based representations
- ex. speech perception and production
functional anatomical model- ventral stream
mapping sound onto meaning
- speech recognition
speech perception (dorsal stream)
- activate and maintain sub lexical items such as phonemes and related articulation
- perceiving speech sounds
- ex. “rhughet” I perceive speech sounds (they are familiar)
speech recognition (ventral stream)
- access the lexicon/stored words
- recognising words
- ex. “rhubarb” “I recognise that word”
Dorsal stream
- left-hemisphere dominant
-acoustic information received - auditory sensory/phonological + motor interface
- articulatory network
- measures of speech motor impairment
Ventral stream
- bilateral
- acoustic information received
- phonological + semantic-> lexical
- higher-level syntax/ combinatinatorial network;
… before moving on to articulation… - measures of impaired speech comprehension
representation
how language is represented- phoneme, morpheme, grapheme/letter, lemma/word form, lexical item/word, concept
process
how language is activated/accessed
modular language models
- also called box-and-arrow, local, or discrete
modular language models representation
- language units are stored locally (together)
modular language models processing
- language units are processed serially (one at a time, in order)
Distributed language models
also called connectionist
Two connectionist models:
1. parallel distributed processing (PDP)
2. interactive activation
Distributed language models representation
language is represented in terms of learned patterns/networks of activation between different knowledge units
Distributed language models processing
knowledge is processed in an interactive manner; not sequential, serial manner
Basic connectionist model principles
- knowledge (language) can be described by interconnected networks of units
- these networks or patterns of units are activated depending on the strength of the network/patterns
Basic connectionist model principles- spreading activation
- when a neutron/unit is activated, it spreads to other neuron’s/units connected to it
Basic connectionist model principles- learning
- ex. speech therapy
- learning is modifying the connections between units
Parallel Distributed Processing- Connectionism
-interconnected networks> representation
- language must me PROCESSED (not just retrieved)
Parallel Distributed Processing
- strength of representation depends on network strength
- more experience/familiarity with representation = greater network strength (connection weight)
Highlight of lecture 9
- aphasia is an impairment of language processing
- models help us understand the specifics of impairments and support clinical decision-making
Modular Models highlight
- localised representations, serial processing
- useful to consider some speech/language phenomena
Connectionist Models highlight
- interactive networks of sub lexical units
- networks connect with and between modalities
- strength of networks based on experience (learning)
Aphasia and Learning
- principles of neuroplasticity
- Learning in aphasia rehabilitation
Learning in aphasia rehabilitation
- errorless learning
- knowledge of results vs performance
Treatment/rehabilitation is learning, learning requires:
- attention
- memory
- executive function
-language - visuospatial skills
attention
attend to treatment tasks
memory
remember instructions
executive functions
work towards a goal
visuospatial skills
to attend to visual stim, facial expression, gesture
important learning components
- multimodal approach to more richly encode info
- intentional cueing framework
- metalinguistic instruction
Feedback: presence, absence, timing
low frequency feedback
may promote learning better than high frequency feedback in many context
high frequency feedback
may promote learning when training complex skills
Performance
- feedback about performance/process
- helpful during training on specifics of performance
Results
- feedback about correctness of response
- helpful in retention of trained skills