Neurolin quiz 3

serial processing

processes that take place sequentially one at a time

parallel processing

two or more processes that take place at the same time and overlap

bottom up

information is processed at lower levels without influence from higher levels, often a serial model

top down

information at higher levels can also influence processing at lower levels, often a parallel model

phoneme restoration

automatic phoneme identification while we look for possible words according to context.  A parallel process.  Influence of lexico-semantics on phonetics, top down processing

tip of the tongue state

the sound/orthographic traces of a signifier allow us to name its partially activated referent.  A parallel process.  Form/phonology influences lexical retrieval, bottom up process

sine wave speech perception

lexical/syntactic/semantic traces allow us to perceive phonemes in an otherwise unintelligible sine wave. A parallel process. Lexico-semantics influences phonological perception, top down processing

The McGurk Effect

visual perception is associated to specific articulatory features that changes our phonemic perception. A parallel process. Phonetics influences phonology, bottom up processing

brain regions involved in speech production

the cortical and subcortical substrates that govern motor skills:

• primary motor cortex

• supplementary motor area

• premotor area

• basal ganglia

• cerebellum

Some language regions:

• supramarginal gyrus

• inferior prefrontal gyrus

• area SPT (sylvian parietal temporal), a parietal-temporal part of the perisylvian tissues for sensorimotor transformations

processes of speech production

according to psycholinguistic models, three serial (top down) steps take place:

1. conceptualization - formation of the abstract message we want to convey, the core idea

2. planning - the message is specified concretely, words are selected and ordered, the phonological form of words is recovered.  The abstract message is converted to linguistic form

3. articulation - implementation of motor instructions, movement of the articulators.  Finalized sound sequence is converted to an acoustic sound sound

planning and memory circuit

regions involved:

• posterior cingulate cortex

• dorsal anterior cingulate cortex

• inferior parietal cortex

• inferolateral temporal cortex

• anterior insula

• superior temporal gyri

motor control and sequencing circuit

regions involved:

• posterior cingulate cortex

• dorsal anterior cingulate cortex

• inferior parietal cortex

• left dorsolateral prefrontal cortex

• pre supplementary motor area

• lateral premotor cortex

network architecture of overt speech production (supported by statistical modelling)

BA 44: location of lexical access, production model activation entry point

Insula: relay point between the cognitive level and the motor circuits

cerebellum/caudate: parallel activation for the selection of motor schemes (caudate) plus timing and smoothing (cerebellum)

premotor cortex: combinations of motor programs into a refined sequence

motor cortex: generation of final motor output to motor neurons

the articulators

the hypoglossal nerve (cranial nerve 12): innervates the muscles of the tongue

the facial nerve (cranial nerve 7): innervates the muscles of the face, mandibles, and lips

fundamental frequency

the lowest frequency component of a complex sound that determines the perceived pitch

gender differences: 200hz female, 125hz male

amplitude

intensity of the sound, our perception of how loud something is

physical properties of sound

frequency and intensity, can be measured

psychological features of sound

we perceive pitch and loudness

factors that make language processing difficult

signal varies enormously depending on multiple factors:

1. depending on the speaker (i.e. gender, accent)

2. depending on the context (formal vs informal)

3. depending on the coarticulation of sounds (i.e. allophones, phonological context)

what we perceive when we perceive language

1. information on the frequency and intensity: the spectral signature

2. information about time (temporal info): the speech envelope

spectral signature

Acoustic frequency is mapped early in the inner ear.

• when inner ear fluid vibrates it moves around the cochlear hair cells.  This movement generates electrical signals that are transmitted to the primary auditory cortex along the auditory nerve

• different sound frequencies stimulate different parts of the cochlea, low frequencies at the apex, high frequencies at the base, called tonotopic organization

• signals from the cochlea travel through several subcortical nuclei before reaching the primary auditory cortex, or Heschl’s gyrus

this takes about 50ms

• includes contralateral transmission and tonotopy

contralateral transmission

sounds from one ear are first processed by the opposite hemisphere’s primary auditory cortex.

Frequency mapping in the cochlea persists in the auditory cortex

acoustic information in represented with a spatial code both at the cochlea and in the primary auditory cortex

tonotopy

the spatial code for frequency, where neighbouring neurons respond to similar sound frequencies

different neuronal loci are activated according to frequency and intensity characteristics, as transmitted by the cochlea. Specific patterns of activation are dependent on the frequency and intensity of sound

speech envelope

our brain can map the changes in loudness over time very well. When an electrode is implanted into the left temporal lobe of a participant who is listening to a natural conversation, the recorded impulses correspond to this.

• includes entrainment and peridotopy

entrainment

match between perceptual and neural activity in the brain, key method the brain uses to lock on to auditory information

peridotopy

the spatial code for the speech envelope in the auditory cortex, different rates of periodic sounds map onto different brain areas

different neuronal loci activate for different temporal characteristics, we can organize temporal information in a spatial way

the neurogram

the neural representations of sounds, tonotopy and peridotopy, the axes of of tonotopia and periodopia are almost perpendicular, so we can encode spatial and temporal information at the same time

phonemic receptive fields

phonemes aren’t categorical but we interpret them categorically. There appear to be ‘phonemic receptive fields’ in the superior temporal gyrus that have feature maps. There isn’t a specific region for a phoneme, but rather a region for phonological features like voicing, place and manner of articulation.  The combination of activation of these regions makes it possible for us to interpret a phoneme.  Phonemic receptive fields have a ‘sample rate’, time windows at which chunks of information are integrated.

left heschl’s gyrus

specialization for short windows (spectral signature) for pitch and interpretation of fast, detailed sounds

right heschl’s gyrus

specialization for long windows (speech envelope) for prosody and slower syllabic interpretation of sounds

afferent motor aphasia

production deficit, the patient experiences a lack of kinesthetic feedback from articulatory movements.  They produce paraphasias unconsciously and have difficulty with articulatory movement.  They have difficulty coordinating speech movement to talk and write, though comprehension is largely preserved.  Typically caused by damage to the lower part of the precentral and supramarginal gyri.

acoustic-mnestic aphasia

production deficit where the patient has difficulty retaining audio verbal traces in memory. They produce paraphasias during naming and in spontaneous speech but their syntax and prosody remains intact.  They have difficulty with word finding and naming.

conduction aphasia

production deficit where the patient produces paraphasias during the repetition of spoken language. Comprehension and spontaneous expression are relatively good, the patient is aware of the deficit.  Typically caused by damage to the arcuate fasciculus and posterior temporal gyrus/supramarginal gyrus

agrammatism (deficit)

production deficit where the patient omits function words, morphemes and articles.  Some suggest that this is related to phonology, in that content words have more phonological salience and are therefore perhaps easier to produce since we need to start with a salient element and syllabic emphasis.  Generally caused by damage to the left inferior frontal gyrus.

cortical deafness

perception deficit where the patient “can’t hear” (is not aware of) sounds, even though their auditory system is intact, and they show observable auditory reflexes.  Generally caused by bilateral damage to the primary auditory cortex.

auditory agnosia

perception deficit where the patient hears consciously but is unable to identify nonverbal sounds.  They can usually identify phonemes and sounds.  Sometimes only one type of sound is affected (i.e. amusia).  generally results from bilateral lesions to the superior temporal gyrus.

verbal auditory agnosia

perception deficit where the patient hears consciously but is unable to identify or discriminate verbal sounds (words, phonemes, etc).  Language specific auditory processing deficit.  They generally retain capacity for production and reading, and perform better in conversational language where they can exploit contextual cues,  Impaired recognition or repetition of novel words,  Generally caused by bilateral temporal gyrus damage.

severe cases: speech sounds like flowing water or wind

less severe cases: difficulty distinguishing minimal pairs, words that differ by a single phonetic feature

phonagnosia

perception deficit where the patient hears consciously and can identify nonverbal sounds,  However the patient cannot recognize familiar voices (right parietal lobe lesion) and/or discriminate voices (left or right temporal region lesion).

auditory processing disorder

the patient hears consciously and knows that they are listening to language, however they experience difficulty interpreting speech, especially in non ideal listening conditions, Can be developmental, or a lesion to the corpus callosum or auditory cortex.

key aspects of sentence structure

different levels of coding are involved in formulating a sentence, our brains don’t just retrieve words, they make everything agree.

• affixation for tense, number or aspect

• grammatical case to show agent/object relations

• agreement for number, gender and/or person

• word order, the sequence of words to convey grammatical information

• movement for questions, subject auxiliary inversion

• ellipses or traces of a syntactic element that has moved

includes constituents, dependencies, and compositionality.

constituent

a group of words that function together as a single unit within a hierarchical syntactic structure.  Its composition depends on the syntactic category of the words.  It behaves as one chunk that can be moved or replaced.

dependency

an obligatory relationship between words or constituents in the hierarchical syntactic structure.  One concept relies on another for its interpretation.

Morphological, syntactic and semantic ones operate at different levels to express grammatical relations.

compositionality

constituents as well as dependencies are interpreted in this way, even if a sentence is heard one word at a time.

complex sentence composition

More activation in perisylvian regions (left inferior frontal gyrus, left superior temporal gyrus). More activation in the left anterior temporal lobe (temporal pole) when complex syntactic structures contain real words

simple sentence composition

The left anterior temporal lobe is strongly activated 250ms post stimulus, reflecting the process of syntactic and semantic composition.  But the activity of the left temporal lobe is modulated by the specificity of composition, when we combine words words with specific meaning it is more active.

neural network for morphosyntactic composition

left posterior temporal lobe - important for syntactic construction and grammatical framework

left inferior frontal gyrus - important for complex syntactic construction, probably due to increased use of working memory, only engaged when syntactic demands are high

left anterior temporal lobe - important for combining conceptual features of words (constituents)

the brain as a prediction center

our brains are constantly making predictions based on:

• constituents

• syntactic dependencies

• semantic context

• the identity of the speaker

The determiner preactivates nouns that are likely to appear next, and the mental lexicon updates and integrates this information with previously predicted words.  These predictions become more precise as we move forward in a sentence, and are continuous throughout a sentence so that we end up with a strong prediction at the end of a sentence.

• the words we predict are pre activated

• when there is a mismatch this is reflected in our brain activity, as an N400 or P600 to integrate unexpected information into the context

inferior frontal gyrus - pre activation even before the first word is finished

auditory cortex - the prediction is verified against actual input while the next word is heard

posterior medial temporal gyrus - the new word is semantically integrated with previous words and the combinations of their meanings

neologisms

morphosyntactic deficit where the patient replaces certain words with words that do not actually exist in the language. Two successive errors, grammatic and phonemic paraphasias, show some degree of lexical access to the word.  Typically results from superior temporal gyrus damage.

paraphasias

morphosyntactic deficit where the patient replaces certain words with words that exist, but are not semantically related, making the message difficult to interpret.  They produce unintended words or sounds.

lexical/verbal: photo becomes trail, results from damage to Broca’s area or the arcuate fasiciulus

semantic: fork becomes spoon, results from damage to Wernicke’s area

phrasal: the fish aquarium becomes the lion cage, results from temporal or parietal damage

logorrhea

morphosyntactic deficit where when the patient speaks, they seem to lose control of the flow of speech, and can only stop with difficulty.  The message can be incoherent, and there is a lack of conscious awareness.  Can result from left hemisphere dysfunction, or from a broad range of damage including Wernicke’s area, the thalamus, the frontal lobe, Broca’s area and the basal ganglia

circumloculation

morphosyntactic deficit where the patient uses an excess of words to convey a simple message.  This appears to be due to anomia or difficulty with lexical access.  They talk around it instead of being direct.  Controlled, and used to compensate for lexical retrieval difficulties.  Can result from lesions or damage to the left hemisphere, temporal, parietal, or frontal lobe.

paragrammatism

commonly seen in sensory/receptive/wernicke’s aphasia from posterior left hemisphere damage

• too much speech, word salad

• function words usually preserved

• inappropriate use of words, especially nouns, includes paraphasias and neologisms

• fluent speech

• rapid speech rate

• incorrect use of grammar, word order, morphology

agrammatism (symptoms)

commonly seen in motor/expressive/Broca’s aphasia form left hemisphere damage.

• too little speech

• function words often omitted, morphology often substituted incorrectly

• verbs are rare but nouns are preserved and well used

• non fluent, telegraphic speech

• slow speech rate

• deficits in grammatical function, not lexical access

• good comprehension for simple syntactic structures

why are nouns more well preserved among agrammatic patients?

syntactic explanation - verbs are more complex at the morphosyntactic level, they can inflect for tense, number, person.  Nouns are simpler and therefore easier to process, retrieve, and integrate

semantic explanation - verbs often represent actions, so frontal motor regions are involved, nouns often represent things we perceive, so sensory regions are involved

hypothesis 1 of agrammatism - agrammatism is a phonological deficit

• some suggest that agrammatism may be related to phonology: content words have more phonological salience are therefore perhaps easier to produce

• even during repetition, agrammatic aphasic patients begin their productions with a stressed word

• syllabic morphemes are less likely to be omitted, omission reflects the ease of producing phonologically salient morphemes

hypothesis 2 of agrammatism - mapping hypothesis

• postulates that Broca’s area is essential for syntactic-semantic mapping, “who’s doing what to whom?”, identifying agent and patient roles

• agrammatism would therefore be due to limitations in processing: syntactic analysis and semantic mapping cannot take place simultaneously

• agrammatic patients usually do know whether a sentence is syntactically well formed, but have difficulty assigning a thematic role to the constituents

hypothesis 3 of agrammatism - adaptation hypothesis

• postulates that syntactic structures are activated late or disintegrate too quickly

• agrammatism would therefore be a processing disorder, perhaps associated with verbal working memory, and they can’t maintain a syntactic structure long enough to produce it

• aphasic patients use coping strategies to get around this ‘reduced processing window’ to limit syntactic processing, including simplifications, reduced speech rate to give them more time to maintain syntactic activation, restarting to maintain activation, simplified grammar, less variety in sentences

Hypothesis 4 of agrammatism - trace deletion hypothesis

• a trace is left from moving constituents out of their original syntactic position

• postulates that Broca’s area is essential for the interpretation of syntactic trees

• if this region is damaged, the patient can no longer produce sentences that do not have the canonical word order of the language, and cannot produce sentences that require a trace and movement transformations