The Time Course of Meaning Construction with Varying Expectations

Abstract

Mechanistic theories of the N400 event-related potential—a neural correlate indexing semantic processing in the brain—implicate the roles of prediction, priming, and bottom-up sensory integration in language comprehension; however, mechanisms explaining volitional aspects of semantic meaning construction are not fully understood. To explore this, participants were visually shown sentences, with words presented one at a time, and evaluated whether the final words of sentences formed sensible (SC) or unconnected completions (UCs). Participant expectancies were modulated using colored boxes that surrounded the words of each sentence cueing the participants to either expect a SC (green) or UC (orange). A neutral cue (purple) did not indicate the completion type and served as a baseline condition. Expectancies were factorially crossed with completion type forming valid, invalid, and neutral conditions. Trial presentations were weighted such that sentences were validly, invalidly, and neutrally cued 60/20/20% of the time, respectively, incentivizing participants to utilize the colored cues. Cues successfully modulated participant expectations such that participants were more accurate when evaluating validly than invalidly cued sentences and selectively faster when solving validly cued sentences that were semantically congruent. The N400, as measured following the presentation of the final word, was modulated by completion type such that UCs elicited more negative deflections than SCs. However, expectations generated via colored cues did not modulate N400 mean amplitudes. These results suggest that volitionally generated expectancies do not dramatically affect neural signatures of semantic access, but ultimately lead to additional processing responsible for resolving discrepancies between semantic congruency and expectancy.

Full text

The Time Course of Meaning Construction with Varying Expectations
• The N400 ERP is a neural correlate indexing semantic processing in the brain [1]
• Mechanistic theories of the N400 implicate the roles of prediction, priming, and bottom-up
sensory integration in language comprehension
• However, mechanisms explaining volitional aspects of semantic meaning construction are
not fully understood
Matthew J. Kmiecik1, Lauren M. Kim1, Mandy J. Maguire1, John Hart Jr.1, & Daniel C. Krawczyk1,2
1The University of Texas at Dallas, 2University of Texas Southwestern Medical Center at Dallas
Introduction
Method
• Participants were presented 400 sentences one word at a time and were instructed to
determine whether the sentence made sense at the final word via button press
• Stimulus creation was based on a normed stimulus set of 200 sentences with the highest
cloze probabilities [2]
• Sentence endings were either sensibly connected (SC) or unconnected (UC) to the
sentence stem and were presented in equal proportions (50/50 split)
• Words were surrounded by three different colored boxes that cued participants’
expectations:
• Trial presentations were weighted such that sentences were validly, invalidly, and neutrally
cued 60/20/20% of the time, respectively, incentivizing participants to utilize the cues
• N400 comparisons were optimized by counterbalancing final words within-subjects, while
sentence stems and button presses were counterbalanced between-subjects
• Behavioral performance and N400 mean amplitudes were analyzed using a model
comparison approach to within-subjects regression (i.e., multilevel modeling) [3]
• Orthogonal contrasts for completion type (SC vs. UC) and condition (Neutral vs. Valid +
Invalid; Valid vs. Invalid) and their interactions were used to calculate within-subjects
dependent variable composites
• Data were collected from 32 right-handed native English speakers (age M = 24.1, SD =
5.17 years; education M = 15.50, SD = 2.23 years; 19 females)
• Colored cues modulated participant expectations by improving performance (better accuracy and faster reaction times) for valid vs. invalid conditions
• Despite the cues’ ability to modulate performance, N400 mean amplitudes were affected only by completion type and not by condition (validity)
• These results suggest that volitionally generated expectancies do not dramatically affect neural signatures of semantic access, but ultimately lead to
additional processing responsible for resolving discrepancies between semantic congruency and expectancy
Discussion
References
[1] Kutas, M., & Federmeier, K. D. (2011). Thirty years and counting: finding meaning in the N400 component
of the event-related brain potential (ERP). Annu Rev Psychol, 62, 621-647.
[2] Block, C. K., & Baldwin, C. L. (2010). Cloze probability and completion norms for 498 sentences: behavioral
and neural validation using event-related potentials. Behav Res Methods, 42(3), 665-670.
[3] Judd, C. M., McClelland, G. H., & Ryan, C. S. (2017). Data Analysis: A Model Comparison Approach to
Regression, ANOVA, and Beyond (Third ed.). New York, NY: Routledge.
Acknowledgements
Many thanks go to Roshni Sharma and Priya Thakur for their assistance in recruiting
and assessing participants. This work was supported by the Distinguished New
Scientist Award (MJK) through the Friends of BrainHealth at The University of Texas
at Dallas Center for BrainHealth®. Figures were prepared in R using ggplot2.
Time
500 500 250 500-1000 ms
+ + [stem]
4-13x
[final].
500
response
1500250
72 72 36 36 72 144 72-144 frames
jittered ITI
[final].
[final].
[final].
cued Sentence Verification Task (cSVT)
Sentence will make sense
Sentence will not make sense
Sentence will either make or not make sense
• SC - Valid - Vic asked her to repeat what she had said.
• UC - Invalid - The baby bird was ready to learn to said.
• SC - Invalid - The kids fed the ducks some stale bread.
• UC - Valid - The athlete enjoyed lifting weights at the bread.
• SC - Neutral - Jesse ran the race at a slower pace.
• UC - Neutral - He had a long day and was in a bad pace.
Sensible Unconnected
Completion
Sensible Unconnected
Completion
SC UC SC UC SC UC
Completion
0.7
0.8
0.9
1.0
Mean Proportion Correct
(SEM Error Bars)
Valid Invalid Invalid Valid
Valid Invalid
Condition
UC
SC
Valid Invalid Invalid Valid
a priori post-hoc Neutral ComparisonsProportion Correct Results
Sensible Unconnected
Completion
0
500
1000
1500
Mean Correct RT (ms)
(SEM Error Bars)
SC UC SC UC SC UC
Completion
Valid Invalid Invalid Valid
200
400
600
800
1000
Invalid
Completion
Valid
UC
SC
Sensible Unconnected
Completion
200
400
600
800
1000
a priori post-hoc Neutral ComparisonsCorrect RT Results
Valid Invalid Invalid Valid
-200 300 500 1000ms
-3
3
6
9
12 µV
SC
UC
Valid
Invalid Valid
Invalid
-3
3
6
9
12 µV
-200 200 400 600 800 1000ms
SC UC
Grandaveraged N400 Results
[final].