Week 3- Speech perception 1 & 2 Flashcards
Speech perception refers to….
how we identify the sounds of language (phonemes)
Spoken word recognition refers to …
how we recognize words
What is prelexical code?
–sound representations before a word is recognized
What is postlexical code?
-sound representations after a word has been identified
Phonemes in speech can be identified a lot faster than what type of sounds?
- __+___+___+___ could be identified at 1.5 sounds per second
- phonemes in speech can be identified at ____ per second or faster
- separate non-speech sounds
- buzz, hiss, tone, vowel
- 20 exemplars
Spoken words in ____ can be identified about ____ ms from their onset
context, 200
When are words recognized more easily and faster?
-words in a meaningful context can be recognized more easily and faster than in isolation
Why can phonemes not be easily separated?
-they slur together
What is co-articulation?
Give an example.
-the pronunciation of a phoneme depends on the preceding and following phonemes, e.g. nasalization of / I/ in “pin” compared to “pig”
Do both hemispheres play causal roles in speech perception? Use the Hickock et al. study and other studies to give proof.
- several studies show that even acute patients with unilateral left-hemisphere (LH) damage need manifest only moderate speech perception problems
- Hickock et al., shows that auditory word recognition is bilaterally organized and approx. 72% of cases have bilateral fairly symmetric lesions
Hypothesis: temporal integration windows differ between LH and RH.
- speech contains information on different ___ ____?
e. g., stop consonant discrimination approx. ___ ms, syllabic stress pattern recognition approx. ____ ms
- time scales
- 20-80
- 150-300
Hypothesis: temporal integration windows differ between LH and RH.
- speech contains information on different __ ___
e. g., stop consonant discrimination approx. ___ ms, syllabic stress pattern recognition approx. ____ ms
- time scales
- 20-80
- 150-300
Hypothesis: temporal integration windows differ between LH and RH.
Poeppel et al. (2003): both hemispheres initially create similar ____ representations of incoming sounds but later sample from these with different ____.
high-fidelity, rates
What is co-articulation?
Give an example.
-the pronunciation of a phoneme depends on the preceding and following phonemes, e.g. nasalization of /I/ in “pin” compared to “pig”
_______ makes faster speech possible
co-articulation
Explain Hickock’s and Poeppel’s Dual Stream Model of speech perception.
Main assumptions and components:
- bilateral speech perception
- difference between left and right sampling times
- ventral (what) and dorsal (how) streams
Do both hemispheres play causal roles in speech perception?
- test to match spoken word with one of four pictures (correct, phonological distractor, semantic distractor, unrelated distractor; e.g. bear, pear, moose, grapes)
- auditory word recognition bilaterally organized
- patients with word deafness have symptoms of variable severity
- approx. 72% of cases have bilateral fairly symmetric lesions
Hypothesis: temporal integration windows differ between LH and RH.
-speech contains information on different __ ___
-time scales
Hypothesis: temporal integration windows differ between LH and RH.
Poeppel et al. (2003): both hemispheres initially create similar ____ representations of incoming sounds but later sample from these with different ____.
high-fidelity, rates
Recognition of phonemes does not have to be based on the_____ of the phoneme in question alone
acoustic segments
Hypothesis: temporal integration windows differ between LH and RH. Abrams et al. (2008): study of 9–12-year-old children listening to “The young boy left home.” What were the two effects?
Two effects: LH and RH suppression to concordance (match with inner speech), LH enhancement to unpredicted discordance
Hypothesis: temporal integration windows differ between LH and RH. Abrams et al. (2008): study of 9–12-year-old children listening to “The young boy left home.” What were the results?
The left hemisphere showed greater sensitivity to phoneme/syllable-level discordance, the right to whole item discordance
If the right hemisphere samples speech in larger chunks, what could aphasia from RH damage look like? (Discussion question)
-may fail to abstract the main point contained in the information being shared
-Inability to interpret body language and facial expressions
-Problems with Conversational Rules
-disorientation to time and direction
-difficulty interpreting visual stimuli
??
If the right hemisphere samples speech in larger chunks, what might be a helpful therapy for LH fluent aphasia? (Discussion question)
-patients are asked to name pictures.
-Body Part Naming
-Oral sentence reading
??
_____ reported in the 1970’s and 80’s for aphasic patients (e.g., Miceli, 1980):
-matching heard words to pictures vs. phoneme discrimination and identification
Give an example:
double dissociations
-e.g. matching /kæt/ with the animal and not the bed but failing to hear the difference between /kæt/ and /kɔt/
Hickock & Poeppel: phoneme tasks recruit some of the same neural substrates as ____, which are separate areas from those recruited by word comprehension tasks
repetition
What is lexical interface? Where is it located?
- a relay station between phonological structures of words (middle STS bilaterally) and semantic structures of words (widely distributed)
- located in pMTG and pITG in both hemispheres
Damage to pMTG and pITG results in what type of aphasia?
transcortical sensory aphasia
Lexical interface- Boatsman et al. (2000)
-six neurosurgical patients received direct electrical stimulation in implanted electrode arrays
Electrode pairs in pMTG resulted in symptoms of what type of aphasia?
-transcortical sensory aphasia
Lexical interface- Dronkers et al. (2004)
-64 chronic aphasia patients performed a battery of 11 sentence-picture comprehension tests
What kind of mapping?
What types of sentences were affected?
- voxel-based lesion-symptom mapping
- all but simple sentence types affected
Rogalsky and Hickock (2009): two tasks:
- detect syntactic anomaly (The plumber with the glasses were installing the sink.)
- detect semantic anomaly (The infant was spilling some carpet on the milk.)
Talk about the findings from Rogalsky & Hickock (2009):
-the whole ATL (anterior temporal lobe) ROI (region of interest) responded more strongly during active tasks than passive listening
-most of the ROI was equally sensitive to semantic and syntactic attention
-a small cluster of voxels showed semantic > syntactic pattern
-no voxels showed syntactic > semantic pattern
=> syntax and semantics processed together in ATL
The ventral stream: from what to what
The dorsal stream: from what to what
The ventral stream: from sound to meaning
The dorsal stream: from sound to speech
Sensorimotor interface.
-the most posterior portion of the supratemporal plane is called
-planum temporale (PT)
Sensorimotor interface.
- based on cell analysis, consist of __ different fields
- includes ____
- four
- area spt (sylvian-parietal-temporal)
spt (sylvian-parietal-temporal) appears to…
-it appears to translate between sound-based phonological network in the temporal lobe and motor-based articulatory network in the posterior frontal lobe
Based on cell analysis, the sensorimotor interface consists of __ different fields
-includes ____
- four
- area spt (sylvian-parietal-temporal)
Articulatory network includes areas involved in ___ and ___ speech and it provides what? …
- overt and covert speech
- provides neural substrate for auditory-verbal short-term memory (STM) = the phonological loop in Baddeley & Hitch’s working memory (WM) model
Articulatory network supports … and allows …
- supports digit span, repetition of pseudowords and nonwords
- allows recycling phonological material in mental rehearsal based on internal (covert) speech
Articulatory network supports…
-supports digit span, repetition of pseudowords and nonwords
Articulatory network includes areas involved in what types of speech? and it provides what? …
- overt and covert speech
- provides neural substrate for auditory-verbal short-term memory (STM) = the phonological loop in Baddeley & Hitch’s working memory (WM) model
Articulatory network supports…
-supports digit span, repetition of pseudowords and nonwords
Articulatory network is also involved in speech perception.
-data from aphasic patients (e.g., Miceli, 1980) suggest that tasks of phoneme monitoring or phonological awareness, such as identifying or discriminating between single phonemes rely on _____ and integrity of the ____ stream
auditory-verbal STM, dorsal
transcranial magnetic stimulation (TMS) allows…
selective enhancement or interference with processing in targeted brain regions
D’Ausilio et al. (2009).
-targeted lip and tongue areas in the left primary motor cortex
-there were two syllables with “lip sounds” /bæ/ and /pæ/ and two syllables with “tongue sounds”: /dæ/ and /tæ/ embedded in white noise
-TMS pulses were delivered either to lip or tongue motor areas 50 ms before consonant onset
-pulses to the lip area improved identification of labial consonants and TMS pulses to the tongue area improved identification of the two dental consonants
What were the results?
=> motor stimulation effects on speech perception
