Documenti di Didattica
Documenti di Professioni
Documenti di Cultura
Karen Garrido-Nag
Hearing, Speech, and Language Sciences
Gallaudet University
Karen.garrido-nag@gallaudet.edu
Frontiers in Hearing
ACKNOWLEDGEMENT
Valerie Shafer, Ph.D.
Developmental Neurolinguistics Lab
The Graduate Center, CUNY
Hia Datta
Yan Yu
Nancy Vidal- Finnerty
2
MEASURING SPEECH & LANGUAGE
PROCESSING Motivation Motor skills
Attention
Language
Memory
Input Output
ERPs:
Paradigm: MMN?, MMR?
Age
Language Background
A8en9on
OVERVIEW
Background
Goals
Hypotheses
Experiment I: Tone
Experimental Design & Methods
Results
Conclusion
Experiment II: Speech
Experimental Design & Methods
Results
Conclusion
Conclusion and Discussion
*Maturational Changes in ERP components that measure
speech perception
Future Studies: Clinic
4
LANGUAGE DEVELOPMENT AND
SPEECH PERCEPTION
5
SPEECH PERCEPTION AND
ATTENTION
Models of infant speech perception suggest that
attention plays an important role in speech
perception and word recognition development
(e.g., Jusczyk, 1997; Werker & Curtin, 2005; 2011)
WRAPSA (Jusczyk, 1997) automatically focus attention on
relevant features of speech sounds
PRIMIR (Werker & Curtin, 2005) uses filters to direct attention to one
or more plane where speech is organized
6
SPEECH PERCEPTION AND
ATTENTION
Direct evidence of how attention affects speech discrimination
during infancy has minimally looked at.
Furthermore, the attentional mechanisms that an infant might
employ in these processes are not defined
The speech signal saliency probably influences the amount of
attention allocated to processing
Developmental changes in attentional mechanisms or resources
could influence what information an infant focuses on
7
HOW DO WE CONTROL
ATTENTION?
Adults: direct them behaviorally
Infants: ???
Explanation 2: Maturational
pMMR is a precognitive recovery from refractoriness
(Vaughan & Kurtzberg,1992; Friederici, Friedrich, & Weber,
2002)
nMMR is an immature MMN (Alho & Cheour, 1990; Pang,
Edmonds, Desjardins, Khan, Trainor, & Taylor, 1998)
Difference in polarity is a due to maturational changes in the
brain (Trainor, 2003)
GOALS
How does attention affect speech processing?
12
PARTICIPANTS: EXPERIMENT 1-
TONES
Visual Stimuli:
Deviant Deviant
0
significant interactions at
-2
Fz (f(10, 90)=3.73, p=.
65.0602ms;0.759229V 000)
-4
522.892ms; 3.00738V
277.912ms;-1.93603V
-6
-200 0 200 400 600 800
msec 6 Mismatch Response Fz Contingent
6 FZ4 S Fz Non-Contingent
GM_Contingent
LM Contingent
FZ4 S4 4
LMNon-Contingent
4 FZ4 D
FZ4 MMR 2
2 FZ4 MMR4
0
V
0
V
-2
-2
-4
478.715ms;-1.21719V
-4 109.237ms; -1.28523V
-6
350.201ms; -3.62588V
-6
-8
-200 0 200 400 600 800
0 200 400 600 800
msec msec
Visual Inspection: an even
1000-Hz 1000-Hz 1000-Hz 1200- Hz broader negativity from
6 250-400 ms with peak
GM_Contingent FZ4 S
FZ4 S4
amplitude of -3.6 V.
4 FZ4 D
FZ4 MMR Presence of MMR: Two-
2 FZ4 MMR4 way Stimulus x Time
ANOVAs at each site
V
0
revealed significant
-2 interactions at all three
478.715ms;-1.21719V frontal sites F3 (f(10, 90) =
-4 109.237ms; -1.28523V
2.58, p = .008), Fz (f(10, 90)
350.201ms; -3.62588V = 2.76, p = .005), and F4
-6
(f(10, 90) = 2.70, p=.006)
-200 0 200 400 600 800
msec
6 6 Mismatch Response Fz Contingent
Task Differences: No
4
GM_Non-Contingent FZ4 S
FZ4 S4 4
Fz Non-Contingent
LM Contingent statistically significant
FZ4 D LMNon-Contingent
FZ4 MMR
2
difference between the
2 FZ4 MMR 4
0 contingent and non-
V
0
V
-2 contingent tasks
-2
-4
65.0602ms;0.759229V
-4
522.892ms; 3.00738V-6
Individual
17
277.912ms;-1.93603V
-6
-8
-200 0 200
msec
400 600 800 0 200
msec
400 600 800
Variability: ???
Number of infants showing negative or positive MMRs for each condition
8 FZ S IA07_Contingent FZ S
IA05_ Non-Contingent 6
FZ D FZ D
6
FZ MMR FZ4 MMR
4
4
2
2
0
V
V
0
-2
-2
-4
-4
-6 446.586ms;2.05698V
-6 354.217ms;5.12932V
-8 -8
-200 0 200 400 600 800 -200 0 200 400 600 800
msec msec
ERP waveforms for two infants are displayed. The positive peak
following the negative MMR is shown by arrows.
18
Topo Map: Experiment 1 (Tones)
Non-Contingent Contingent
The MMR difference were greatest at the left frontal sites near F3
(site 13) and midline frontal Fz (site 4)
Conclusions: Experiment 1
Attention enhances the negativity, and/or eliminates
the positivity
20
HYPOTHESES
a. If the pMMR is similar to the adult P3a
component, for all participants, we
expect to see a pMMR to the deviant
stimulus for both conditions and
stimuli.
21
EXPERIMENT 2: BP -BIP
Purpose
22
Stimuli: Experiment 2
Visual Stimuli:
Black and white patterns:
Deviant Deviant
31
HYPOTHESES
a. If the pMMR is similar to the adult P3a
component, for all participants, we
expect to see a pMMR to the deviant
stimulus for both conditions and stimuli.
amplitude, if attention to the stimuli
strengthens a weak representation of the
relevant information.
32
DISCUSSION AND CONCLUSIONS
Attention modulates the MMR responses to speech
34
DISCUSSION AND CONCLUSIONS
pMMR: may either be a recovery from refractoriness or an orienting
response.
The presence of a predominant pMMR when attention is inconsistently
or randomly focused on the auditory stimuli may indicate that the
infants are actually not discriminating the differences in speech
contrasts
35
MATURATIONAL
CHANGES: P1
LATENCY &
MMR LATENCY
36
Maturational changes: T-complex latency
T7/left T8/right
3 months
6 months
12 months
18 months
3 yrs
4 yrs
6yrs
37
125 -198 ms 128 -241 ms
FUTURE DIRECTIONS: CLINIC
1. With the maturational factors involved in speech
perception development, it would be an advantage to
test participants both cross-sectional and
longitudinally to document the effects of attention in
speech perception (speech and language measures)
39
MMN and Attention
Some studies suggest that MMN is not attention-dependent
(Naatanen, 1990)
Others indicate that MMN magnitude can be modulated by
attention (Muller, 2002; Dyson, 2005; Muller-Gass, Stelmach, &
Campbell,
2006; Gomes, 2000; Hisagi, Shafer, Strange &
Sussman 2006)
It follows:
a) Modulation of attention in the visual modality does not interfere
with sound processing
b) Attention to the auditory channel has an effect on MMN size, if
the deviants are difficult to discriminate
40
EVENT RELATED
POTENTIALS (ERPS)
Mismatch Response (MMR)=deviant - standard
With or without overt attention
indexes automatic pre-attentive discrimination of auditory
information
often seen positive MMR(pMMR) in young children and negative
MMR (nMMR, or MMN) in older children and adults (Shafer, Yu &
Datta, 2010)
CONTROVERSIES ON MMR IN
YOUNG CHILDREN
6 months
12 months
18 months
3 yrs
4 yrs
6yrs 500 ms
44
125 -198 ms 128 -241 ms
TO SUM
P100, MMR and T-complex latency decreases as age
increases.
45
ERP COMPONENTS
One single ERP component may not be
sufficient to index speech processing.
46
CONCLUSIONS
pMMR is a more primitive processing
nMMR/MMN indexes the robustness of speech
sound representation
Increased negativity in bilingual girls is probably due to more
attention allocated to processing the speech contrast.
CONCLUSIONS
The latency of P1, MMR, and T-complex
decreases as age advances.
For both monolingual and bilingual children,
the positive MMR dominants at younger
age, and negative MMR/MMN emerges
around age 4 years old.
Amplitude of MMR correlates with age only
in monolinguals, possibly due to attention
modulation effects in bilingual females.
48
Measuring speech & language processing in
infants
Input Output
Behavioral: Brain measures are aected by a number of
factors:
High-amplitude sucking
Age
Heart-rate measures Cor9cal matura9on
Gender
Visual habituation/dishabituation
Language exposure
paradigms
Language performance
A8en9on
CONCLUSIONS (CONTD)
Language group differences observed as early as six months
of age.
T-complex response appears sensitive to English language
measures (e.g., PPVT).
50
ERPS: AUDITORY AND
LANGUAGE PROCESSING
Phonology
Syntactic
Auditory processes Reanalysis
Syntax Semantics
_________________________________________________________________
Time (ms)
0 150 300 450 600
ABR N50 P1 N1 ELAN P2 N2 P300 LRP N400 LAN P600
PN MMR LR
obligatory cognitive
responses responses
Neural Basis t0 t1 t2 t3 t4
of ERPs
t1
0 0
t0 t1 t2 t3 t4
EEG VS. ERP
Electroencephalogram (EEG)
spontaneous brainwaves
4 P2 229 ms
-2
Voltage
+ ENVIRONMENT Time
EVENT RELATED POTENTIALS (ERPS)
Cz
S S S S S
4 P2 229 ms
V 0
Voltage
-2
N1 119 ms
-4
0 200 400 600 800
milliseconds
Time
EVENT RELATED POTENTIALS
13- 4- 62
17 Cz (63)- 54
LM -26
RM -51
161-490; 30 ms
Background- neurophysiology
65
SPEECH PERCEPTION AND
LANGUAGE DEVELOPMENT
Infants learn to weight cues according to importance in the
ambient language (Kuhl; Curtin & Werker, 2005).
Infants learn from the statistical distribution of phonetic
variation in the speech (Maye, 2002)
Infants learn to automatically attend to relevant acoustic
cues in the ambient language (e.g., Jusczyk, 1997)
Selective Perceptual Routines
(SPRs, Strange & Shafer, 2008)
These abilities develop over the first few years
of life (Nittrouer & Miller, 1997)
66
TONE
67
68
BIP-BEP
69
70
THE PRIMARY
PURPOSES
Investigate the relationship between brain maturation and
language experience
the developmental changes of ERP responses from infancy to early
school age
role of bilingual exposure in speech processing
whether ERP measures correlate with language measures
71
EVENT RELATED
POTENTIALS
(ERPS)
ELECTROENCEPHALOGRAM
(EEG)
Cz
Voltage
+ ENVIRONMENT Time
EVENT RELATED POTENTIALS (ERPS)
Cz
S S S S S
Voltage
Time
TOPOGRAPHY
TOPOGRAPHY
Latency: 226 ms
CONCLUSIONS
Bilingual infants and children show more dynamic changes,
with interjec=on of adult-like paAerns of speech percep=on
than monolingual children modulated by certain factors
Brain measures of speech percep=on in infants and children
are aected by a number of factors:
Age
Cor9cal matura9on
Gender
Language exposure
Language performance
A8en9on
Attention & MMR
Attention to a random set of sounds: pMMR Attention to sound-change: : nMMR