Week 8 - Speech & Language Flashcards

1
Q

Hearing

A

Stimulus- sound waves

Amplitude & loudness:
High- big sound waves & loud sounds (SHOUTING)
Low- small sound waves & low sounds (WHISPERING)

Frequency & pitch:
High - many sound waves close together & high pitch
Low- few sound waves far apart & low pitch

Complexity & timbre:
A mixture of frequencies reflects a sound’s quality or character. A sound’s complexity reflects its timbre

Direction of sound:
Brain automatically determines this by calculating the slight difference in the time that it takes the sound waves to reach the two ears.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
2
Q

Sound localisation

Echolocalisation

Complexity

A

Identifying the source of the sound waves

Identifying the location of the sound waves

Detect and interpret complexity of the sound waves to enable speech & music perception

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
3
Q

Language

Definition

A

A form of communication that involves learning complex rules to make & combine symbols into meaningful sentences

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
4
Q

Language - knowledge

Schema theories

A

Schemas are organised packets of information stored our memory about the world, events, or people

Schemas allow us to form expectations & they influence comprehension

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
5
Q

Language - knowledge

Schema theories - Bartlett’s theory

A

Schemas play an important role in determining what we remember from stories

Conflict between presented story and relevant prior schematic knowledge

Rationalisation: tendency to produce errors in the recall of stories that conform to the cultural expectations of the rememberer

(Up) Errors (up) retention interval

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
6
Q

Language - knowledge

Kintsch’s construction-integration model

A

Propositions (that convey meaning of text) are constructed from each sentence

Each proposition leads to retrieval of associated propositions form long term memory

The two combine to form an elaborated propositional net

Spreading activation process selects propositions for text representation. Clusters of highly interconnected propositions attract the most activation & are more likely to be included in the text representation (episodic text memory). Irrelevant ones are discarded. This is the integration process.

Three levels of representations:

  • surface representations (the text itself)
  • propositional representations
  • situation representation (deep, conceptual, pragmatic, inferences)
How well did you know this?
1
Not at all
2
3
4
5
Perfectly
7
Q

ELAN: syntactic violations

A

ELAN (early left anterior negativity):

Indexes violations of syntax and grammar

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
8
Q

N400: semantic violations

A

N400:

Indexes violations of word meaning & world knowledge

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
9
Q

Speech production

McGurk Effect

A

A video shows a person repeating the sound gah

An audio plays the sound bah

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
10
Q

Neela says to the waiter, “I take coffee with cream and smog.”
Which ERP would be evoked for the type of language error in that statement?
A.MMN
B.N1
C.ELAN
D.N400

A

D. N400

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
11
Q

Speech perception

Ganong effect

A

Shows that word recognition processes use lexical information to guide perception when the speech signal is underspecified.

•Ganong (1980) constructed speech segments were to be ambiguous.
e.g., a sound designed to be acoustically
intermediate between /g/ and /d/

  • A pair of test items that began with this ambiguous stop consonant, with one member of the pair having “ulf” appended, and with the other member of the pair ending in “usk.” Listeners reported hearing “gulf” rather than “gusk” and “dusk” rather than “dulf”
  • Perception of the ambiguous segment is biased toward a sound that produces a word rather than a nonword.

The phonemic restoration effect, the Ganong effect, and the McGurk effect all show that speech perception cannot be understood solely in terms of the mapping between an acoustic (or gestural) event and a phonetic percept; speech perception is also guided by additional information sources available to the perceiver.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
12
Q

Speech perception

Dual route model

A

Dorsal stream:
Translating acoustic-based representations of speech signals into articulatory representations essential for speech production.

Ventral stream:
Processing speech signals for comprehension and meaning.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
13
Q

Speech production

Classic model: wernicke-lichtheim-geschwind (WLG)

A

Shared neural circuitry

Broca’s area - Speech production

Wernicke’s area - Speech comprehension

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
14
Q

Shared neural circuitry

Unification

A

Indicates the real-time assembly of the pieces retrieved from memory into larger structures with contributions from context.

●Unification operations take place at the syntactic, semantic and phonological levels. Conceptual and lexical elements are combined and integrated into larger structures.

Which sentence is more difficult to understand?
“The reporter who the senator attacked admitted the error”
“The reporter who attacked the senator admitted the error”

The increased difficulty to understand (semantic) is due to differences in the structure of object and subject relative clauses (syntax).

Unification refers to the ongoing combinatorial operations in language, which include both context-driven predictions and integration of the lexical information into a representation that spans the whole utterance.

●Predictions work forward in time, whereas integration works backward in time; in the latter case, information is integrated into a context that preceded the currently processed lexical item.

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
15
Q

Aphasia

A

Aphasia is an impairment of language, affecting the production or comprehension of speech and the ability to read or write.

●Aphasia is always due to injury to the brain-most commonly from a stroke, particularly in older individuals.

●Brain injuries resulting in aphasia may also arise from head trauma, from brain tumors, or from infections.

MIDDLE CEREBRAL ARTERY
Stroke (Arterial Blockage/Bleeding)

Symptoms - Disorders of Production
Poor articulation, anomia/ problems with word finding, paraphasia / unintended words / phrases, loss of syntax / grammar, inability to repeat aural information, low verbal fluency.

Symptoms - disorders of comprehension
Poor auditory comprehension, poor visual comprehension

How well did you know this?
1
Not at all
2
3
4
5
Perfectly
16
Q

Fluent aphasia

A

Fluent speech (no articulatory disorder)
but difficulties either in
auditory comprehension or
in repetition of words, phrases, or sentences spoken by others.

Wernicke (sensory)

Transcortical (isolation syndrome)

Conduction

Anomic

17
Q

Nonfluent aphasia

A

Difficulties in articulation
but relatively good
auditory comprehension

Broca (expressive) severe

Broca (expressive) mild

Transcortical motor

Global

18
Q

Pure aphasia

A

Selective impairments in reading (alexia), writing (agraphia), or recognizing words in the absence of other symptoms of language disorders

Alexia

Agraphia

Word deafness

19
Q

Speech, language &. Reading disorders

A

In developing an oral language, one of the first tasks an infant must master is the perception & production of the speech sounds specific to the native language

●Although innate constraints influence some aspects of human language acquisition, mastering a particular oral, and especially written, language requires extensive learning, much of it implicit statistical learning

●Connectionist models, which implement statistical learning of speech, language, and reading, provide a useful framework for thinking about relations among LI, RD, and SSD at the cognitive level

Computational models address how young children learn the complex mapping between acoustic features and articulatory gestures, particularly when important aspects of the articulatory gestures made by adult models are not observable.

●Hidden units learn the particular abstract mapping between acoustic features and articulatory gestures that signal meaning differences for that child, and the representations of these abstract mappings are phonological representations.

●Lexical semantics is intimately involved in the determination of which acoustic and articulatory features are counted as relevant for a young child’s particular lexicon in a particular language.

●As the child’s vocabulary increases, the nature of her phonological representations also changes. Consequently, the development of phonological representations is protracted, and the weighting given to different acoustic features in speech perception changes with development.

Comorbidities between the disorders means that common or overlapping models have been proposed to explain the deficits seen across the disorders

20
Q

Language impairment. (LI)

A

Multiple models proposed for LI
a)Poor temporal resolution of perceptual systems (bottom-up auditory model)
b)Slowing of cognition rather than perception
c)Phonological short-term memory deficit
d)Procedural learning deficit also affecting non-linguistic motor functions
●Single deficit theories of LI are inadequate as SI associated with multiple deficits

21
Q

Reading disability (RD)

A

Children with dyslexia or reading disability have problems with basic printed word recognition.
●This is not the same as children who are poor comprehenders (good word recognition but poor understanding).
●Printed word recognition has two written language skills components:
a)phonological coding
b)orthographic coding

22
Q

Speech sound disorder (SSD)

A

Multiple causes put forward for SSD
a)a bottom-up problem in processing acoustic features
b)a motor problem in planning and producing articulatory gestures,
c)a problem learning the mapping between the two, a problem identifying which phonetic differences signal differences in meaning and which are equivalent (i.e., in learning phonological representations, a top-down problem in learning semantic representations that impedes the differentiation of phonological representations), or
d)some combination of these problems
●SSD associated with substitution of phonemes and not phonetic distortions