Speech Production Flashcards
How to investigate speech production
Observation of movements of speech e.g MRI, recordings of natural speech, analysis of natural and experimentally induced speech errors (TMS), neuroimaging (FMRI), genetic e.g FOXP2 KE Family
Source filter model (Fitch, 2010)
Air produced by pulmonary pressure-> larynx where vocal folds vibrate -> filtered by superlayrngeal gyrus
Phonetics is the
Study of speech sounds
What are consonants classified by
Voicing, Place and Manner e.g plosives- b, g, t: compressed air forces open closed vocal folds in larynx
Manner= stops and burst release followed by vowel
Vowel classification
Horizontal/vertical and high/low/ middle e.g ‘E’= front (horizontal) and mid
MRI to measure speech production
Allows study of whole vocal tract with good soft tissues contrast and resolution
Support for MRI to measure speech production
Ventura et al (2009): provided useful morphological data about position and shape of different speech articulators
EPG to measure speech production
Displays timing and location of tongue contact with roof of mouth during continuous speech
How is EPG used?
Custom made artificial palette which contains electrodes that register tongue contact
Goal of measuring speech production
Further our understanding of normal and abnormal speech production and improving QoL for individuals with impaired communication
Fisher et al (1998) KE Family study
Fisher et al (1998): KE family
- 15 members across 3 generations could understand spoken words but struggled to string words together to form response
- mutated FOXP2 gene
Genetic basis of complex vocal articulation- family finding
Mutations in FOXP2 gene interfered with the part of the brain responsible for language development learning problems
Watkins et al (2002) study on family members with aphasia
Family members and patients had similar profiles of impairment
Stages to speaking
- activation of a concept
- selection of correct syntactic structure and words
- activation of a phonetic code
- coordination of many muscles to produce articulatory output
Partial reason for no single all-inclusive model of speech production
Limited access to process as it occurs with our consciousness
All models of speech production agree
- How linguistic components are retrieved and assembled during continuous speech
- linguistic information represented by distinctive units and hierarchy of levels
Simple model of speech production (cognitive- levelt et al., 1999)
Conceptualisation, formulation and articulation
Schriefers et al (1990): Evidence for simple model of speech production
Semantically related EARLY= slower to name picture
Phonologically related later= SLOWER
(Semantic interference)
Evidence for simple model (Levanthal et al., 1999) suggests
Meaning is processed earlier than sounds
Types of speech errors
Exchange, preservation, addition, subtraction, blend, deletion
Speech error at semantic level
May occur when semantic unit is activated e.g ‘Blonde eyes’ instead of blonde hair
Speech error at phonological level
A wrong phoneme is activated e.g hats and dogs instead of cats and dogs
Criticism of sequential model (simple model) of speech
What about phonetic and semantic word errors e.g. Virtue of chastity rather than charity
Parallel models of speech
Stages are simultaneous and not sequential
Support for parallel models
Lexical bias effect- people more likely to make phoneme errors that end up in actual words than non-words
Dell et al (1986) interactive model of speech production
4 layers=
- Semantic
- Syntactic
- Morphological and phonological
- Selected word
Levelt et al (1999) criticism of parallel models/speech error evidence
Internal error monitor can explain speech errors, so do we really need a model?
Speech production in the brain. Which area is most associated?
Broca’s area
Indefery and Levelt (2004) how long for speech production
600ms in total
A meta-analysis of 82 studies found which areas involved in speech production?
PET and FMRI majority
- superior, medial parietal lobe, right anterior, medial frontal lobe, anterior temporal lobes
Motor theory of speech production argues
Motor representations are the primary unit of speech production
Lieberman (1950) on speech motor system:
Involved in perception of speech and listening is a motor act
Wilson et al (2004) FMRI study- motor theory
Listening to speech activated bilaterally a superior portion of the ventral pre motor cortex- largely overlapped Brodmann’s area 4a and 6
Limitation of Wilson et al (2004) FMRI study (motor theory)
Scott et al (2009): no control group
Pulvermuller et al (2006) motor theory finding
Overlap between areas involved in speech perception and speech production is articulatory-specific
Meister et al (2007)TMS on pre motor cortex
% correct ID’s impaired due to disruption of human premotor cortex
Yuen et al (2010) EPG & heard match/mismatching the syllable & distractor congruent/incongruent
Found subtle change in producing syllable (mismatching)
Incongruent distractors= greater alveolar contact vs congruent
Scott et al (2009) critique of role of motor cortex in speech production
Neurological data show no correlation between speech categorisation
Speech production and perception dissociate in development
Motor cortex can be linked to speech in other ways e.g. Time taking essential in joint speech
