PSY 401 Exam #2 Flashcards
what are the 2 main computational channels for speech processing (DM)?
dorsal stream: auditory motor functions
ventral stream: comprehension
what is spectrotemporal analysis? where is it conducted in the brain?
- SA: computation of which sound frequencies are present in the environ and their moderations over time; input received via signals from the thalamus
- conducted in Heschel’s gyrus [A1] + dorsal plane of superior temporal gyrus [STG]
DUAL STREAM MODEL BEGINS HERE
what are the 3 proposed steps of hierarchically organized speech perception?
- dorsal STG carries out early spectrotemporal analyses
- mid-posterior lateral STG represents subphonemic features and feature combos
- mid-posterior STS represents individual phonemes and sequential phonol. structures of whole words
where are sub-phonemic features of speech processed?
L mid-lateral STG; HALFWAY POINT of the stream
what is categorical speech perception? where in the brain has it been represented?
- different versions of the same general phoneme are perceived as being the same, even when they vary acoustically
- there is NOT a 1:1 mapping of acoustic signal and phoneme id
- mid-lateral STG
where in the brain are phonemes/phonological structures of words stored?
L mid-posterior STS
what is phonological neighborhood density? where in the brain are they activated?
- the # of words that differ from a target word by only 1 phoneme ex. BET - bit, bat, get, pet (dense)
CHURCH - ? (sparse)
- the # of words that differ from a target word by only 1 phoneme ex. BET - bit, bat, get, pet (dense)
- DENSELY POP WORDS TAKE LONGER TO PROCESS
- word neighborhood activation occurs in the MID-POST STS (step 3)
hierarchical pathway of speech perception: top down v bottom up processing
BOTH occur:
- bottom up: acoustic information from environ
- top down: expectations from experience! hypothesis, trying to prevent prediction errors
entrainment
process by which the frequencies of connected speech correlate with rhythms of neural activity / brain wave oscillations
- is this an aquired assoc? the brain waves match to any speech rhythm being perceived? - No sabemos
- INCREASED syncing assoc w INCREASED perception
do fMRI/PET experiments imply causal relationships?
- NO; just bc an area is active during a task DOESN’T mean it’s necessary for that task –> must look at neuropsych/brain damage data
is speech perception a uniquely L Hemisphere process?
- NO; studies show that R hemi is also active in speech percep processes
- patients w L hemi superior temporal brain damage acutely only have moderate SP issues –> R hemi can get the job done
Wada procedure + implications
- sedative administered that “immobilizes” 1 hemisphere in order to test LATERALIZATION of diff behaviors/brain functions
- i.e. if L hemi is inactive, will this behavior be impaired?
- -> most times NO
- i.e. if L hemi is inactive, will this behavior be impaired?
- further implies bilateral nature of early cortical speech perception
word deafness + implications
- neurolog. disorder where speech percep is disrupted but hearing is intact; caused by BILATERAL lesions to Mid-Pos STG (both hemis must b damaged)
[SPARES HESCHEL’s GYRUS/A1]
- further implies bilateral nature of early cortical speech perception
asymmetric sampling (AST) in time method
the left hemisphere may be dominant for detecting and categorizing rapidly changing signals (e.g., cues for place of articulation), whereas the right hemisphere may be dominant for dealing with longer signals (e.g., cues for manner of articulation)
- R HEMI: slow fts (manner)
- L HEMI: fast fts (place)
initial evidence for dual streams in speech percep: comprehension + repetition
- conduction aphasia [IMPAIRED REP, INTACT COMP] v transcortical sensory aphasia [IMPAIRED COMP, INTACT REP]
- suggests SEPARATE processing stream or comprehension + articulation/repetition
what are the 2 functional components of the ventral stream (DM)?
- lexical interface
2. combinatorial network
ventral stream: lexical interface - what does it do?
- a relay station for linking phonological structures of words —> their semantic structures/meanings during spoken language comprehension
- does NOT store meanings!! just is a middleman for connecting to them
where is semantic knowledge stored in the brain?
- all over! distributed across the cortex
what is a lemma?
- an abstract word node; intermediary stage between semantic meaning and phonolo. structure of a word, while maintaining syntactic info [ex. whether concept it’s a noun or verb]
ventral stream: combinatorial network - what does it do?
- links together the integrated meanings of words and sentences; word and sentence level compreh.
- sensitive to syntactic and semantic fts of speech [i.e. responds during passive listening to SENTENCES but not to just a list of nouns]
dorsal stream: sensorimotor interface - what does it do?
- maps phonological structures of words to vocal tract movement / speech motor areas of frontal lobe
- p21
dorsal stream: articulatory network- where is it located?
- located in PM, pIFG, ant insula
dorsal stream: sensorimotor interface - where is it located?
- located in area SPT in the planum temporale
- covers 4+ distinct cytoarchitectonic areas
dorsal stream: articulatory network - what does it do?
- needed for speech production
ventral stream: lexical interface - where is it located? what does damage here lead to?
- L Posterior Middle Temporal Gyrus (pMTG) and p Inferior Temporal Gyrus (pITG)
- stimulation here induced Trans Sens Aphasia [IMPAIRED COMP, INTACT REP]
ventral stream: combinatorial network - where is it located?
- L Anterior Temporal Gyrus (ATL), aMTG, aITS
which aphasias result from damage to the sensorimotor interface / area SPT?
- Conduction Aphasia:
- comp INTACT (no damage to ventral stream!)
- repetition IMPAIRED (bc area SPT is needed to translate what you hear —> how to say it)
- lots of phonemic paraphasias (bc motor programming of speech is disrupted)
- Logopenic PPA:
- anomia + same symptoms, less pronounced
what is auditory-verbal short term memory? which component of the Dual Stream Model facilitates it?
- the “phonological loop”, leads to maintained saliency of auditory info; what we hear is kept and “repeated” by our articulators, so we can continue to reference it for a time
- facilitated by the SENSORIMOTOR INTERFACE
critiques of the dorsal stream: articulatory network (4)
- L frontal lesions impair production but compr remains intact
- L hemi can be disactivated (Wada proced) and speech production/comp is fine
- Ppl who don’t develop speech production can still comprehend speech
- infants can perceive speech without the ability to produce speech
how does TMS stimulation of the frontal motor cortex affect perception?
- stim of areas for lips v tongue —> faster RTs in perceiving labial [lips] v dental [teeth] sounds
- takeaway: FRONTAL MOTOR SPEECH AREAS R ACTIVE DURING PRODUCTION AND PERCEPTION
what are the 2 segments of the Lemma Model?
- lexical selection - id’ing the most appropriate word in the mental lexicon for a concept you want to express
- subject to COMPETITION
- form encoding - figuring out the articulatory shape of a word
- NOT subject to competition
what is the first step in word production?
- mapping the idea u want to express onto a lexical concept
what is a lexical concept?
- a unit that considers/binds together all the different semantic fts that constitute the meaning of a word
ex. would be activated when seeing a pic of a horse (visual) and hearing a neigh (auditory)
what is the word-frequency effect? at what stage of form encoding are its effects shown?
- phenomena where common/”high frecuency words” [ex. dog] are more quickly retrieved than less common words [ex. chamomille]
- influence PHONO. CODE RETRIEVAL
what are the 3 characteristics of morpheme phonological codes?
- only the selected lemma spreads activation across the “rift” between the lexical selection subsystem and the form encoding subsystem; other lemmas that happen to be engaged to some extent during the process of lemma selection do not activate the corresponding morphemic phonological codes —-> es decir, COMPETITION FOR ACTIVATION CEASES ONCE LEMMA IS SELECTED, DOES NOT AFFECT PHONO CODE
- influenced by the word freq effect
- retrieved in parts, and the BEGINNING segment is most important
- [ex. when retrieving “banana”, RT is fastest when parts know it starts with “ba” / knowing it ends in “na” is less helpful]
are syllabification rules stored in long term memory?
- NOPE! syllabification occurs “on the fly” p9
what is the mental syllabary? about how many syllables are included in it? at what stage of the lemma model is it engaged?
- a repository of motor programs/instructions for producing regularly used syllables in a given language
- up to 500 syllables
- engaged during PHONETIC ENCODING
where in the brain are lexical concept nodes represented?
- repped in the Anterior Temporal Lobe (ATL) maybe
[activated 230ms after picture onset] - ppl w semantic PPA have lesions here [they r impaired at pic naming tasks]
what brain area resolves activation conflicts in the ATL?
- the Inferior Frontal Gyrus; selects the most relevant lexical concept out of all that are activated after hearing a sound/seeing a pic
p12
-
lesion sites of which condition are assoc with housing lemmas?
- pure anomia: have a disconnect between semantic content and phonol. content of words —> tip of the tongue states
where in the brain are phonological codes retrieved? (Step 1 in form encoding)
- L posterior STG
DIVA Model:
- Directions Into Velocities of Articulators
DIVA model: what is a somatosensory target representation?
- specifies the tactile and proprioceptive (i.e., body-part positional) sensations that are expected to be felt in the vocal tract whenever the phrase is produced; Once this representation is in place, it serves as a kind of template against which incoming somatosensory signals are compared
DIVA Model: feedforward control sys - what does it do?
- sufficient for producing well-learned speech sounds under normal circumstances
- evidenced by the fact that
DIVA Model: feedforward control sys - what does it do? how do we know it exists?
- sufficient for producing well-learned speech sounds under normal circumstances
- evidenced by the fact that humans can produce and modify their speech even when they can’t hear themselves[ex. hay white noise blocking]/
spastic dysarthria - which part of DIVA model does its damage correspond to?
characterized by reductions of the speed, strength, range, and accuracy of articulatory movements;
- MUSCLES R INTACT, CONTROL is disrupted in ventral MC Articulator Map
inverse models: what do they do?
- their function is to estimate in advance the motor commands that are required to achieve particular sensory effects
lemma model: what are the 2 segments?
- lexical selection
2. form encoding
lemma model: lexical selection - wht r the 2 parts?
- lexical concept selection -
- lemma selection
* BOTH SUBJECT TO COMPETITION**
lemma model: lexical concept selection - where does it occur?
- L ATLs
lemma model: phonetic encoding where does it occur?
- L BA44, ant insula, vPMC
lemma model: lemma selection - where does it occur?
- anterior MTG, ITG; may be topographic, category specific
lemma model: phonol. code retrieval - where does it occur?
- L pSTG
lemma model: syllabification - where does it occur?
- pIFG
cranial nerves relevant to speech (5) - which dysarthria does damage lead to?
- trigeminal V- jaw mvm, proprioception of skin, face muscles
3 . facial VII- facial exp, lips
- vagus X - throat, larynx
- hypoglossal XII- tongue mvmnt
- glossopharyngeal IX - elevation of pharynx & larynx, swallowing, posterior 3rd of tongue
- damage leads to FLACCID dysarthria
DIVA model: auditory maps - where are they located?
- p Auditory Cortex
DIVA model: somatosensory maps - where are they located?
- v post somatosensory cortex
DIVA model: articulator map - what does damage here lead to?
- ventral MC –> Spastic Dysarthria
DIVA model: initiation map - where is it? what does damage lead to?
- Supplemental Motor Area –> SMA syndrome, akinetic mutism
- BASAL GANGLIA INPUT*
- hyperkinetic / hypokinetic dysarthria
DIVA model: speech sound map - where is it? what does damage lead to?
- left vPMC —> Apraxia of Speech (AOS)
DIVA Model: auditory state map
- sends signals to —> Auditory ERROR map signals are EXCITATORY
DIVA Model: auditory target map
- sends signlas to –> AUDITORY ERROR MAP, signals are INHIBITORY
DIVA Model: feedback ctrl map - where is it? what does it feed into?
- R vPMC, IFG / feeds into ARTICULATOR MAP –> leads to corrected speech
DIVA Model: Auditory + Somatosensory maps - damage here leads to?
- ataxic dysarthria –> cerebellar damage, poor speech timing, stressing, loudness