Hemispheric asymmetries and joke comprehension

1
Hemispheric asymmetries and joke comprehension

• Coulson, S., Williams, R. F. (2005)
• Neuropsychologia, 43, 128-141.

2
Background

• Joke comprehension seems to include 2 parts
  surprise and updating (frame-shifting)
• I asked the bartender for something cold and
  full of rum, and he recommended
• a daiquiri expected fits context of bar and
  drinks
• his wife unexpected requires reevaluation of
  sentence context (cold rum bartenders wife)

3
ERPS Joke Comprehenison Coulson Kutas,
2000

• Are there 2 separate stages?
• Recorded ERPs while participants read RSVP
  sentences that ended with either a joke or
  non-joke word (wife/daiquiri)
• Why ERPs?
• temporally precise can examine timecourse of
  joke comprehension, specifically when surprise
  and updating stages occur, or if they are even
  separate

4
ERPS Joke Comprehenison Coulson Kutas,
2000

• Stimuli
• Jokes and non-joke controls were identical until
  final word
• Joke and non-joke endings were matched on cloze
  probability (3)
• They were further divided by the constraint of
  the sentence, or cloze probability of the most
  popular response
• High constraint (gt 40 cloze prob.)
• She read so much about the bad effects of
  smoking that she decided shed have to give up
  reading/the habit.
• Low constraint (lt 40 cloze. Prob.)
• Statistics indicate that Americans spend 80
  million a year on games of chance, mostly
  weddings/dice.

5
ERPS Joke Comprehenison Coulson Kutas,
2000

• Procedure
• 200ms per word 500ms for last word
• Sentence followed with true/false comprehension
  question, to ensure correct interpretation of
  joke
• e.g., The bartender suggested a frozen drink
• The bartenders wife was a frigid lush
• Responses were used to divide good comprehenders
  (avg. 83 correct) and bad comprehenders (avg.
  64 correct)
• All but one subject scored above 83 on
  non-jokes, so comprehension problems were only
  with jokes

6
ERPS Joke Comprehenison Coulson Kutas,
2000

• Enhanced N400 for joke endings
• Negative-going wave, peaking 400ms
• Distribution broad, slightly right-lateralized,
  stronger over centro-parietal areas
• Enhanced N400s typically associated with
  difficulty of lexical processing/ integration
• LAN (left anterior negativity / sustained
  negativity)
• Associated with WM demands and updating sentences

7
ERPS Joke Comprehenison Coulson Kutas,
2000

• Effects varied with comprehension and constraint
  groupings
• N400
• Broader N400 distribution for poor comprehenders
  (distribution for good comprehenders more
  canonical)
• N400 effects only significant for high constraint
  sentences
• LAN
• Only present for good comprehenders, regardless
  of constraint

8
ERPS Joke Comprehenison Coulson Kutas,
2000

• Conclusions
• Joke endings produced an N400 effect only in
  constraining sentence contexts
• This seems to represent the surprise element of
  the joke ending
• But the distribution varied, and the effect was
  stronger for poor comprehenders
• Good comprehenders also showed a LAN
• Seems to represent updating the sentence context
• However, authors conclude that these 2 components
  do not support the 2-stage theory, because of
  temporal overlap

9
Evidence for hemispheric asymmetries?

• Studies on unilateral brain damage have shown
  different deficits in joke comprehension
• LHD (left hemisphere brain damage) gt problems
  with recognizing the initial disruption
• RHD (right hemisphere brain damage) gt problems
  with updating the sentence context appropriately
• Neuroimaging studies (fMRI) have shown increased
  activation in the RH (right hemisphere) during
  joke comprehension

10
Visual Half-field Paradigm
Stimuli presented in one hemifield are received
by contralateral hemisphere RVF (right visual
field) LH (left hemisphere) LVF (left visual
field) RH (right hemisphere)
Subjects fixate centrally stimuli are then
presented to the left (left visual field LVF) or
to the right (right visual field RVF) of fixation
11
Coulson Williams, 2005 Design

• Stimuli
• Same as 2000 study, but now sentence-final joke
  word is presented to left or right of fixation
• Also included expected filler items for
  comparison
• Procedure
• Read sentence word-by-word
• When prompted, name final word (only correctly
  named words were included in analysis)
• Answer comprehension question (gt90)

12
Coulson Williams, 2005 Results

• Predictability effects
• Enhanced N400s for joke and non-joke endings,
  relative to fillers
• Joke effects
• For rvf/LH presentation, larger N400 for jokes
  than non-jokes
• For lvf/RH presentation, no N400 differences

13
Coulson Williams, 2005 Results

• Jokes elicited a LAN both VFs
• Only present at one channel (F7)
• Jokes elicited frontal positivity both VFs

14
Coulson Williams, 2005 Conclusions

• Clear differences in how each hemisphere
  processes jokes
• LH showed N400 differences, reflecting difficulty
  integrating joke ending
• RH showed no difference, suggesting that it has
  no problem understanding jokes
• Consistent with
• Coarse-coding
• RH activates a broader array of semantic
  associates more prepared for deviation from
  expected
• LH has more fine-grained activation more
  disrupted by deviation
• Prediction / integration
• RH builds a flexible context able to accept
  unexpected items if they still make sense
• LH rapidly builds context and tries to predict
  upcoming items thrown off by contrasting items

15
DISCUSSION

• Why are they not getting P600 effects?
• Process of joke comprehension seems similar to
  garden-path processing why are there
  differences
• If N400 effect reflects recognition that the joke
  is inconsistent with current interpretation, why
  does it not show up for the RH at all?
• If the LAN represents reevaluation, this must
  have been signaled by inconsistency

16
DISCUSSION

• Given that level of constraint and comprehenders
  status was so important in previous study, why
  not divide into these groups again?