Nina Last Lecture

Front 1

Explain why understandingg spoken speech is difficult?

Back 1

Spoken speech is continuous and has weird places for breaks

Front 2

How many syllables do we say a second in speech?

What is sinewave speech?

Back 2

6 syllables

A modified speech signal that is ambiguous between speech and non speech

Front 3

If you sequence speech to phoneme mapping wrongly what might you get?

Back 3

CAT instead of ACT

Front 4

There is no one to one mapping between what?
What does this make difficult?

Back 4

Waveform and phonemes
This makes it difficult to work out phonemes in speech.

Front 5

What are formants?

Back 5

Two main frequencies in the vowel waveforms.

Front 6

Regarding vowels, again what is another reason speech is difficult to work out?

What does this suggest?

Back 6

Everyone has variance in their vowel formants. And therefore, some people that say O, will run into the same formant space that others use to say U. So how do you tell what is what

THere is more information needed to work it out.

Front 7

Regarding consonants, why as far down as these can't you get a one to one mapping between sound and understanding?

Back 7

There is variation in the formant transition into the consonant, each different.

Front 8

How do we perceive notes in music?

Back 8

Linear perception.

Front 9

Explain a categorical perception curve?

When we perceive speech sounds what kind do we show?

Back 9

A bunch of notes all percieved as short. And another bunch of durations only perceived as long.

When we perceieve speech sounds we exhibit a categorical perception curve.

Front 10

What is the voice onset time?

Back 10

The very start of the consonant as the air starts coming out.

Front 11

Whats the difference between ta and de?

Back 11

The voice onset time. Ta has longer VOT.

Front 12

Where does the boundary between consonants ta and de lie? What does this map?

Back 12

about 30ms of voice onset time. This maps the categorical perception curve and demonstrates it.

Front 13

How do we work out the differences between consonants?

Back 13

Differences in voice onset times makes our perception categorical.

Front 14

Why is categorical perception useful?

Back 14

It selectively highlights phonemic distinctions that are important for distinguishing word meanings and therefore, increases the speed of speech processing

Front 15

Is categorical perception a process that occurs early, like a bottleneck, or later?

Back 15

Sharma et al 1999

Front 16

What is mismatch negativity? Who created it?

Back 16

Mismatch negativity (Naantanen 1978)

Recorded brain responses in an experiment where people heard sounds in an oddball paradigm.

Front 17

What is an oddball paradigm?

Back 17

Lots of the same sound, and occasional infrequent sound.

Front 18

What did Naanaten find?
What did he control for and how?

Back 18

He found the first 100ms of both waveforms were similar and then they diverged. (Mismatch negativity)

He controlled for the physical different in sounds (so it can't be due to that). They swap standards and deviants and then average the physical differences out.

Front 19

What is the mismatch negativity measuring?

Back 19

Change detection

Front 20

Where is mismatch negativity shown in the brain?

Back 20

The auditory cortex

Front 21

Who answered the question how early can categorical perception take place?

Back 21

Sharma et al 1999

Front 22

Explain what Sharma et al 1999 did?
Results?

What did Sharma et al imply?

Back 22

Recorded brain responses to 3 syllables:
Da (VOT 20ms)
Ta (VOT 40ms)
Ta (VOT 60ms)

Results: MMN was there for the Da ta contrast, but not the Ta, Ta contrast.

Categorical perception takes place early within 150ms, overriding objective truth about the sound.

Front 23

What study shows us how we understand the meaning of speech quickly? And what is this called?

Back 23

Categorical perception
Sharma et al 1999

Front 24

How fast is speech encoded and how?

Back 24

Speech is encoded immediately using categorical perception.

Front 25

Is speech processing special?

And a supportive double dissociation?

Back 25

Yes. Auditory input canbe processed by two different processors, a dedicated speech processor and a non speech processor. (Rand 1974)

Pure word deafness vs amusia (Poeppei 2010)

Front 26

Name two theories of speech perception?

Back 26

Motor theory of speech perception
General auditory approach to speech perception

Front 27

Explain the motor theory of speech perception

Back 27

When trying to understand speech we subvocally recruit our own vocal apparatus to try to understand whats being said.

Front 28

What is the gist of the motor theory of speech perception?

Back 28

Intended motor gestures are invariant. So for pa you must first close your lips, thenopen your mouth to say ah, then put your tongue on the top of your mouth. The intended moto gestures are the same, so you can decode other peoples words more easily.

Front 29

What gave the motor theory of speech further evidence? (2) (1 reference)

Back 29

Mirror neurons.

When someone is istening to speechl, speech production and speech perception areas are active. (Wilson et al 2004)

Front 30

Is the moto area a neccessity for speech perception?

Damage to what should lead to what?

Back 30

No. If it was you'd expect someone who has very sever emotor deficits would not be able to perceive speech. Muscular dystrophy, or brocas aphasia.

Damage to motor areas should lead to impaired speech recognition but it doesn't in brocas aphasia. Damage to the left frontal regions.

Front 31

Explain the gneralised auditory approach to speech perception.

Back 31

Listeners recover the spoken mesage from the actual acoustic signal.

If you look for multiple features within the acoustic signal you can find a unique combination of acoustic features to identify phonemes.

Front 32

There is a close link between speech perception and speech production systems.
Motor theory or generalised auditory approach correct?

Back 32

Motor theory

Front 33

Some properties of speech perception (e.g.categorical perception) general auditoryproperties. Motor theory or generalised auditory approach correct?

Back 33

Auditory right

Front 34

Speech perception probably not innate speciesspecific. Motor theory or generalised auditory approach correct?

Back 34

Motor theory wrong

Front 35

Name one thing motor theory is correct about?
Name one thing auditory theory is correct about?
Name one thing motor theory is wrong about?

Back 35

There is a close link between speech perception and speech production systems.

Some properties of speech perception (e.g. categorical perception) general auditory properties. Motor theory or generalised auditory approach correct?

Speech perception probably not innate speciesspecific.

Front 36

Name 3 things speech perception is aided by?

Back 36

Lexical and sentential context (Ganong effect)
Visual context
Regularities from the listeners native language

Front 37

Explain what lexical and sentential context is? What is the ganong effect?

Back 37

The ganong effect

If you have an ambiguous VOT, between say B and T... your brain will use whichever letter makes sense for the sentence.

Front 38

When the sound is ambiguous such as D and P, what do you do?

Back 38

You use the context of the sentence/ganong effect

Front 39

Explain how we get help from visual context? What effect?

Back 39

Mcgurk effect McGurk 1976
Ba.. ta

Front 40

Explain how we can get help from the regularities from our native language?

Back 40

You can use lexical information, sublexical (chunks/regularities in language smaller than words) or suprasegmentall information (bigger chunks of words)

Front 41

Explain a suprasegmental cue?

Back 41

The stress you put in words. So for instance in finnihs and hungarian you always stress the first syllable.

Front 42

Explain an example of a sublexical cue?

Back 42

If you heard a word like tmin, you'd know its not a real word because t and m don't go together so you'd discount it.