Electrophysiological

1
Tone and Intonation
Electrophysiologica
l investigations of semantic, phonemic and
tone processing in Cantonese  
Mireille Besson1, Daniele Schön1, Annett
Schirmer2, Elain Ng3 Trevor Penney3
1Mediterranean Institute for Cognitive
Neuroscience, CNRS Marseille 2 MPI,
Leipzig 3Chinese University, Hong-Kong
Supported by the Procore
International Program
International Conference
on Tone and Intonation
Santorin Sept 9-11 2004
2
Previous experiments on tone languages
Does the
perception of pitch (F0)
vary as a function of linguistic
relevance ?
Cross linguistic
approach Behavioral evidence quite
mixed YES (Van Lancker Fromkin, 1973 Lee
Nusbaum, 1993) NO (Repp Lin, 1990 Cutler
Chen, 1997) Brain imaging more
homogeneous (PET, fMRI)
linguistically relevant
larger left
hemisphere activation
(Gandour et al,
1998 2000 2002 Klein et al, 2001) Pitch
variations
linguistically irrelevant
larger activation of
right frontal regions
(Klein et al, 2001)
3
Overall aim of the present experiment
Examine the temporal dynamics of phonemic,
tone and semantic processing in
Cantonese Event-Related brain Potentials
(ERPs) methodology Excellent
temporal resolution



4
Methods
Event-Related brain Potentials
M. BESSON - CRNC - CNRS - Marseille - France
5
(No Transcript)
6
(No Transcript)
7
(No Transcript)
8
(No Transcript)
9
Summation and average
EEG
s
s
s
s
s
20 à 30 essais
5 µV
Evoked Potentials (ERPs)
0
2000 ms
1000
200
400
600
800
1200
1400
1600
1800
s
A. BENRAÏSS
10
Specific aims
Aim 1 Characterize the N400 effect in
Cantonese in terms of
latency, amplitude and
scalp distribution Aim 2
Examine the relationship between phonemic
and tone processing
Spoken sentences rather than isolated
words



11
SEMANTICS
VISUAL MODALITY
AUDITORY MODALITY
(From Holcomb Neville, 1990, Lang. Cog.
Processes)
(From Kutas, Lindamood Hillyard, Nature, 1984)
12
Methods
130 sentences (CP gt 80) ending with monosyllabic
nouns 1. Semantics and tone
congruous (40) CW (Hong Kong
farms mostly feed / chicken  gai 55) 2.
Phonemic, tone and semantic incongruous (30)
PIW (This cinema at same
time shows three set / movies hei 33 / knife
dou 55) 3. Tone and semantic
incongruous (30) TIW
(Winter weather compare autumn / colder dung
33 / east dung 55) 4. PseudoWord /
tone unexpected (30) TMW
(Milk contains large amount / calcium koi 33
/ calcium koi 55) Participants 22
right-handed native speakers of Cantonese Task
Is the sentence semantically acceptable ?
Base your judgment both on segmental and tonal
information
13
Behaviour
Percent errors very low (4 ) Not
different between conditions Reaction Times

TIW TMW lt PIW lt CW
14
N400 effect
Congruous
Incongruous
15
Right frontal lateralization of the N400 effect
16
CW vs TIW
CW TIW
17
TIW vs TMW
TMW
TIW
18
All conditions
CW
PIW
TMW
TIW
19
Tentative conclusions
Right frontal lateralization
of the N400 (CW vs PIW)
Linked to tone language ?
Attention focussed on tonal
changes ?
Further experiments
CW vs TIW Unexpected
change in tone (from 33 to 55)
Early tone mismatch detection
process (MMN)
Beginning of an N400 component (attempt to access
meaning) Late
positivity (decision that there is a tonal
change) PIW vs TIW Tone
violations in both cases but
phonemic information OK in
TIW
wrong in PIW
Increased N400 in PIW
Interaction between phonemic
and tone processing

20
Thank you !
21
(No Transcript)
22
Tone languages tonal patterns allow the
differentiation of lexical meaning
from otherwise identical
strings of phonemes 55 yau worry rest 35
yáu paint 33 yau thin snail 21 yàuh
oil swim 13 yáuh have friend 22 yauh
again right (hand)
F0 most important cue
From Fangxin Chen, Chinese Research Lab.
23
TPW 55 Tone
CW 33 Tone
yau (33) seng (55)
han (33) aa (33)