Right hemisphere sensitivity to word & sentence level context: Evidence From Event-Related Brain Potentials.
Coulson, S. Federmeier, K.D., Van Petten, C., and Kutas, M. (2005)
JEP: Learning, Memory, and Cognition.
Gist
Question: Is right hemisphere message blind? Measurement: ERP Paradigm: associate priming without/within
sentence context Answer: RH is not message blind, but there is
indeed hemispheric asymmetries in the use of word and sentence contexts
Message-blind RH --the hypothesis
LH
has the ability to integrate syntactic, semantic, and pragmatic information to construct a message-level representation of meaning.
RH its language competence extends only to word-level priming mechanisms
Message-blind RH --why people made this claim?
LH priming in RT to words embedded in
normal or scrambled sentences (Faust et al. 1995)
Larger priming effect when the amount of context is increased (Faust et al. 1993)
Longer RT to words in implausible sentence context than plausible context (Faust, 1998)
RH Contexts do not seem to facilitate or
hinder the language process
so…RH seems to be blind to these message level information..
Message-blind RH --however…..
LH RH Should be sensitive to some
message level information, since when RH is damaged
patients cannot understand certain kinds of jokes, metaphoric language, and sarcastic utterances
Some studies did find sentence congruity effect in the RH (Chiarello, Liu, & Faust, 2001; Faust, Bar-lev, & Chiarello, 2003)
Ex1 [word level context effect]Lexical association
Ex2 [sentence level context effect]Lexical associationSentence congruity
Experimental design (EX1)
Associated Unassociated
Congruous
They were truly stuck, since she didn’t have a spare TIRE.
… Ellen leaned over and borrowed my spare PENCIL.
Incongruous
… Ellen leaned over and borrowed my spare TIRE.
They were truly stuck, since she didn’t have a spare PENCIL.
Primes are centrally presented, and
targets are lateralized to either visual field (split visual field display).
SPLIT VISUAL FIELD DISPLAY
SPLIT VISUAL FIELD DISPLAY
PROCEDURE
++++
200ms
200ms
1000~1200ms
300ms
2500ms
spare
tire
?
0ms
Prediction
Ex1[word level context effect]Since both hemispheres are sensitive
to word level info, similar-sized N400 context effects are expected
Experiment 1—results
EX1
LVF/rh
EX1
LVF/rh
N400LPC
EX1
RVF/lh
Experimental design (EX2)
Associated Unassociated
Congruous
They were truly stuck, since she didn’t have a spare TIRE.
… Ellen leaned over and borrowed my spare PENCIL.
Incongruous
… Ellen leaned over and borrowed my spare TIRE.
They were truly stuck, since she didn’t have a spare PENCIL.
The cloze probability were matched between the two types of congruous sentences (associated & unassociated) and also between the two types of incongruous sentences. This was done to ensure that the message level constraints are similar in the associated and unassociated conditions.
Prediction
Ex1[word level context effect] Since both hemispheres are sensitive to
word level info, similar-sized N400 context effects are expected
Ex2[sentence level context effect] LH:
• A large N400 congruity effect• Negligible effects of association
RH: (if the message blind RH model holds)• A large N400 association effect• Negligible effects of context congruity
Experiment 2—results
N400LPC
EX2
LVF/rh
N400LPC
EX2
RVF/lh
Summary
word level: association effects for both LVF/rh and RVF/lh presentation
sentence level: robust congruity effects for both LVF/rh and RVF/lh presentation
the message-blind RH model is not supported
At the sentence level, the congruity effect lead to a dramatic attenuation of the association effect.Lexical context is less important in sentence
contexts
5 µV
Association effect
Association effect
Congruity effect
unassociated
associated
unassociated
associated
incongruous
congruous
Summary
Although both hemispheres make use of word level as well as sentence level contexts, they seem to use them in different ways.
LH seems to make use of lexical association only when the sentence context is incongruous.
RH shows a smaller lexical association effect at the word level, which suggests that RH might be weaker to use this source of semantic context.
RH shows the association effect in congruous sentences there might be a greater reliance on word level relationships in the lexical integration processes in understanding sentences
Questions
Is the LPC the same thing as the P600? Do blinks also produce surges that might overshadow brain activity
on the EEG? How common is subject attrition due to excessive artifacts? Onset or peak? Which point is more interesting? In what situations
would one or the other be the focus? What if they ran their experiments on brain damaged patients ? Why was a naming paradigm chosen for the target as opposed
lexical decision? Great Britain norms vs. US participants ….isn’t it problematic? Is the hemispheric asymmetry in reliance on lexical relationships
has to do with other abilities commonly associated with RH function (e.g., spatial abilities).
Other questions?
ELECTRODE POSITIONS