Post on 06-Sep-2018
Rapid lexical processing revealed by the time-course of ERP responsesEmma Folk (efolk@villanova.edu)
Villanova UniversityJoe Toscano (joseph.toscano@villanova.edu)
Villanova University
wraplab.co
WRAP Labhttp://wraplab.co/
Method References & Acknowledgments
Introduction DiscussionResultsExperiment 1: Stimulus Norming
Experiment 2: Lexical Status
Experiment 3: Syntactic Class Experiment 4: Semantics (Animacy)
Fz F3
F7
FC5 FC1
C3 T7
CP5 CP1
Pz P3
P7
O1 Oz O2
P4
P8
CP6 CP2
Cz C4 T8
FC6 FC2
F4
F8
-8
+4
-200
Time (ms)
800
Fz F3
F7
FC5 FC1
C3 T7
CP5 CP1
Pz P3
P7
O1 Oz O2
P4
P8
CP6 CP2
Cz C4 T8
FC6 FC2
F4
F8
-8
+4
-200
Time (ms)
800
Fz F3
F7
FC5 FC1
C3 T7
CP5 CP1
Pz P3
P7
O1 Oz O2
P4
P8
CP6 CP2
Cz C4 T8
FC6 FC2
F4
F8
-8
+4
-200
Time (ms)
800
Fz F3
F7
FC5 FC1
C3 T7
CP5 CP1
Pz P3
P7
O1 Oz O2
P4
P8
CP6 CP2
Cz C4 T8
FC6 FC2
F4
F8
-8
+4
-200
Time (ms)
800
Fz F3
F7
FC5 FC1
C3 T7
CP5 CP1
Pz P3
P7
O1 Oz O2
P4
P8
CP6 CP2
Cz C4 T8
FC6 FC2
F4
F8
-8
+4
-200
Time (ms)
800
Fz F3
F7
FC5 FC1
C3 T7
CP5 CP1
Pz P3
P7
O1 Oz O2
P4
P8
CP6 CP2
Cz C4 T8
FC6 FC2
F4
F8
-8
+4
-200
Time (ms)
800
Fz F3
F7
FC5 FC1
C3 T7
CP5 CP1
Pz P3
P7
O1 Oz O2
P4
P8
CP6 CP2
Cz C4 T8
FC6 FC2
F4
F8
-8
+4
-200
Time (ms)
800
Fz F3
F7
FC5 FC1
C3 T7
CP5 CP1
Pz P3
P7
O1 Oz O2
P4
P8
CP6 CP2
Cz C4 T8
FC6 FC2
F4
F8
-8
+4
-200
Time (ms)
800
Fz F3
F7
FC5 FC1
C3 T7
CP5 CP1
Pz P3
P7
O1 Oz O2
P4
P8
CP6 CP2
Cz C4 T8
FC6 FC2
F4
F8
-8
+4
-200
Time (ms)
800
Fz F3
F7
FC5 FC1
C3 T7
CP5 CP1
Pz P3
P7
O1 Oz O2
P4
P8
CP6 CP2
Cz C4 T8
FC6 FC2
F4
F8
-8
+4
-200
Time (ms)
800
Fz F3
F7
FC5 FC1
C3 T7
CP5 CP1
Pz P3
P7
O1 Oz O2
P4
P8
CP6 CP2
Cz C4 T8
FC6 FC2
F4
F8
-8
+4
-200
Time (ms)
800
Fz F3
F7
FC5 FC1
C3 T7
CP5 CP1
Pz P3
P7
O1 Oz O2
P4
P8
CP6 CP2
Cz C4 T8
FC6 FC2
F4
F8
-8
+4
-200
Time (ms)
800
Fz F3
F7
FC5 FC1
C3 T7
CP5 CP1
Pz P3
P7
O1 Oz O2
P4
P8
CP6 CP2
Cz C4 T8
FC6 FC2
F4
F8
-8
+4
-200
Time (ms)
800
Fz F3
F7
FC5 FC1
C3 T7
CP5 CP1
Pz P3
P7
O1 Oz O2
P4
P8
CP6 CP2
Cz C4 T8
FC6 FC2
F4
F8
-8
+4
-200
Time (ms)
800
Fz F3
F7
FC5 FC1
C3 T7
CP5 CP1
Pz P3
P7
O1 Oz O2
P4
P8
CP6 CP2
Cz C4 T8
FC6 FC2
F4
F8
-8
+4
-200
Time (ms)
800
Fz F3
F7
FC5 FC1
C3 T7
CP5 CP1
Pz P3
P7
O1 Oz O2
P4
P8
CP6 CP2
Cz C4 T8
FC6 FC2
F4
F8
-8
+4
-200
Time (ms)
800
Fz F3
F7
FC5 FC1
C3 T7
CP5 CP1
Pz P3
P7
O1 Oz O2
P4
P8
CP6 CP2
Cz C4 T8
FC6 FC2
F4
F8
-8
+4
-200
Time (ms)
800
Fz F3
F7
FC5 FC1
C3 T7
CP5 CP1
Pz P3
P7
O1 Oz O2
P4
P8
CP6 CP2
Cz C4 T8
FC6 FC2
F4
F8
-8
+4
-200
Time (ms)
800
Time (ms)
Volta
ge (µ
V)
-8
+4
-200 800
Fz F3
F7
FC5 FC1
C3 T7
CP5 CP1
Pz P3
P7
O1 Oz O2
P4
P8
CP6 CP2
Cz C4 T8
FC6 FC2
F4
F8
-8
+4
-200
Time (ms)
800
Fz F3
F7
FC5 FC1
C3 T7
CP5 CP1
Pz P3
P7
O1 Oz O2
P4
P8
CP6 CP2
Cz C4 T8
FC6 FC2
F4
F8
-8
+4
-200
Time (ms)
800
NounVerb
Fz F3
F7
FC5 FC1
C3 T7
CP5 CP1
Pz P3
P7
O1 Oz O2
P4
P8
CP6 CP2
Cz C4 T8
FC6 FC2
F4
F8
-1
+2
-200
Time (ms)
800
Fz F3
F7
FC5 FC1
C3 T7
CP5 CP1
Pz P3
P7
O1 Oz O2
P4
P8
CP6 CP2
Cz C4 T8
FC6 FC2
F4
F8
-1
+2
-200
Time (ms)
800
Fz F3
F7
FC5 FC1
C3 T7
CP5 CP1
Pz P3
P7
O1 Oz O2
P4
P8
CP6 CP2
Cz C4 T8
FC6 FC2
F4
F8
-1
+2
-200
Time (ms)
800
Fz F3
F7
FC5 FC1
C3 T7
CP5 CP1
Pz P3
P7
O1 Oz O2
P4
P8
CP6 CP2
Cz C4 T8
FC6 FC2
F4
F8
-1
+2
-200
Time (ms)
800
Phonetic Effects Cross-splicing Procedure
Major debate in language processing concerns whether information is processed sequentially (Trueswell et al., 1995)
Sequential processing predicts late effects for higher-level information; parallel processing predicts simultaneous effects
Questions:
What is the time-course of lexical, semantic, and syntactic processing?
Is information processed in parallel across different levels of organization?
Recent work suggests lexical processing begins by 140-200 ms (Baart & Samuel, 2015; Toscano et al., in prep)
Certain ERP components may reflect these processes (N400; P600), but there is debate over what these components index (Kutas & Federmeier, 2011; Tanner & Van Hell, 2014)
Alternative approach is to use component-independent design to examine time-course (Thorpe et al., 1996)
Word-initial voicing minimal pair stimuli presented over Etymotic ER-3A insert earphones at MCL
Cross-spliced to control for acoustic differences
Each stimulus presented 40 times in Exp. 2 and 20 times in Exp. 3-4, in random order
2AFC task in which stimuli were categorized based on lexical status (Experiment 2), syntactic class (noun vs. verb; Experiment 3), or lexical semantic information (animacy; Experiment 4)
EEG responses recorded with a Brain Vision actiCHamp system with 32 active electrodes (FP1, Fz, F3, F7, FT9, FC5, FC1, C3, T7, A1, CP5, CP1, Pz, P3, P7, O1, Oz, O2, P4, P8, A2, CP6, CP2, Cz, C4, T8, FT10, FC6, FC2, F4, F8, FP2), at 500 Hz, with a high-pass filter at 0.01 Hz. Referenced online to A1 and re-referenced offline to average mastoid
Pair Prop. Word (Word, NW)
Freq. (Word)
babe/pabe 1, 0 8
baint/paint 1, 0 38
badge/padge 1, 0 6
bouch/pouch 1, 0 2
desk/tesk 1, 0 65
dooth/tooth 1, 0 25
Sets of minimal pair stimuli presented to identify pairs that served as clear exemplars of the relevant category
Pair Prop. Word (Word, NW)
Freq. (Word)
dice/tice 1, 0 15
doan/tone 1, 0 78
gang/kang 1, 0 22
gake/cake 1, 0 13
gas/kas 1, 0 100
gast/cast 1, 0 48
Pair Prop. Noun (Noun, Verb)
Freq.(Noun, Verb)
bull/pull 1, 0 19, 60
buy/pie 0.88, 0.12 14, 70
dame/tame 0.89, 0 8, 5
deem/team 0.94, 0 86, 1
grave/crave 0.94, 0 33, 2
grab/crab 1, 0 2, 16
Pair Prop. Animate (Animate/Non)
Freq. (Animate/Non)
Ben/pen 0.76, 0 21, 18
big/pig 1, 0 8, 360
Dan/tan 0.94, 0 27, 9
dim/Tim 0.82, 0 0, 19
granny/cranny 0.94, 0 6, 2
grab/crab 0.88, 0 2, 16
Word Non-word
badge padge
paint baint
N=15 subjects; no N1 differences (i.e., cross-splicing was successful)
Lexicality effect observed at several scalp locations, particularly at central-parietal channels (Cz, Pz, CP1, CP2)
Effect onset occurs 62 ms after POD (coarticulatory info for final phoneme; 326 ms)
N=9 subjects
No N1 differences (as expected)
Syntactic class (noun vs. verb) difference observed for Pz 92 ms after POD (288 ms)
Potentially earlier effects but insufficient statistical power
Fz F3
F7
FC5 FC1
C3 T7
CP5 CP1
Pz P3
P7
O1 Oz O2
P4
P8
CP6 CP2
Cz C4 T8
FC6 FC2
F4
F8
-8
+4
-200
Time (ms)
800
Fz F3
F7
FC5 FC1
C3 T7
CP5 CP1
Pz P3
P7
O1 Oz O2
P4
P8
CP6 CP2
Cz C4 T8
FC6 FC2
F4
F8
-8
+4
-200
Time (ms)
800
AnimateNon-animate
Cross-splicing technique successfully controlled for acoustic differences in order to isolate lexical status, syntactic class, and semantic effects
Time-course of processing reveals extremely rapid lexical activation; similar to other recent work suggesting early effects (60 ms, Baart & Samuel, 2015; 144 ms, Toscano et al., in prep)
Similar rapid onsets for syntactic class (92 ms) and animacy (46 ms) effects
Time-course of each effect overlaps with that of phonetic processing (approx. first 200 ms after onset of acoustic difference), suggesting parallel activation of phonetic and lexical representations
Lexicality effect may correspond to leading edge of N400 onset (Kutas & Federmeier, 2011), though component-independent design allows us to examine time-course independently of specific ERP components
Early syntactic class effect fits with recent work arguing against traditional N400/P600 distinction (Tanner & Van Hell, 2014)
Overall, results suggest spoken language comprehension relies on parallel activation across multiple levels of representation
We would like to thank Steph Vicari, Nicole Johnson, Beach Brooks, Lexie Tabachnick, Chris Burley, Olivia Pereira, and Tifani Biro for assistance with data collection. This work was supported by a Villanova Graduate Summer Fellowship to EF.
References
1. Baart, M., & Samuel, A.G. (2015). Early processing of auditory lexical predictions reveled by ERPs. Neuroscience Letters, 585, 98-102.
2. Kutas, M., & Federmeier, K.D. (2011). Thirty years and counting: Finding meaning in the N400 component of the event related brain potential (ERP). Annual Review of Psychology, 62, 621-647.
3. Tanner, D., & Van Hell, J.G. (2014). ERPs reveal individual differences in morphosyntactic processing. Neuropsychologia, 56, 289-301.
4. Thorpe, S., Fize, D., Marlot, C. (1996). Speed of processing in the human visual system. Nature, 381, 520-22.
5. Toscano, J.C., Anderson, N.D., Garnsey, S.M., Fabiani, M., & Gratton, G. (in prep). Non-invasive measurement of cortical responses to speech using fast diffuse optical imaging.
6. Trueswell, J.T., Tanenhaus, M.K., & Garnsey, S.M. (1994). Semantic influences on parsing: Use of thematic role information in syntactic ambiguity resolution. Journal of Memory and Language, 33, 285-318.
N=10 subjects
No N1 differences (as expected)
Animacy effect at Cz 46 ms after POD (288 ms)
Earlier effect at Pz (-142 ms before POD); may be due to coarticulatory information in vocoid from /n/-final stimuli
-1
+2
-8
+4
-1
+2
-8
+4
Fz F3
F7
FC5 FC1
C3 T7
CP5 CP1
Pz P3
P7
O1 Oz O2
P4
P8
CP6 CP2
Cz C4 T8
FC6 FC2
F4
F8
-8
+4
-200
Time (ms)
800
Fz F3
F7
FC5 FC1
C3 T7
CP5 CP1
Pz P3
P7
O1 Oz O2
P4
P8
CP6 CP2
Cz C4 T8
FC6 FC2
F4
F8
-8
+4
-200
Time (ms)
800
Fz F3
F7
FC5 FC1
C3 T7
CP5 CP1
Pz P3
P7
O1 Oz O2
P4
P8
CP6 CP2
Cz C4 T8
FC6 FC2
F4
F8
-8
+4
-200
Time (ms)
800
Fz F3
F7
FC5 FC1
C3 T7
CP5 CP1
Pz P3
P7
O1 Oz O2
P4
P8
CP6 CP2
Cz C4 T8
FC6 FC2
F4
F8
-8
+4
-200
Time (ms)
800
Fz F3
F7
FC5 FC1
C3 T7
CP5 CP1
Pz P3
P7
O1 Oz O2
P4
P8
CP6 CP2
Cz C4 T8
FC6 FC2
F4
F8
-1
+2
-200
Time (ms)
800
Fz F3
F7
FC5 FC1
C3 T7
CP5 CP1
Pz P3
P7
O1 Oz O2
P4
P8
CP6 CP2
Cz C4 T8
FC6 FC2
F4
F8
-1
+2
-200
Time (ms)
800
Fz F3
F7
FC5 FC1
C3 T7
CP5 CP1
Pz P3
P7
O1 Oz O2
P4
P8
CP6 CP2
Cz C4 T8
FC6 FC2
F4
F8
-1
+2
-200
Time (ms)
800
Fz F3
F7
FC5 FC1
C3 T7
CP5 CP1
Pz P3
P7
O1 Oz O2
P4
P8
CP6 CP2
Cz C4 T8
FC6 FC2
F4
F8
-1
+2
-200
Time (ms)
800
WordNon-word
-1
+2
-8
+4
Point of disambiguationSignificant difference from zero (yellow: p<0.05)
Onset of effect
Noun Verb
bull pull
pie buy
Larger N1 for short VOTs
Fz F3
F7
FC5 FC1
C3 T7
CP5 CP1
Pz P3
P7
O1 Oz O2
P4
P8
CP6 CP2
Cz C4 T8
FC6 FC2
F4
F8
-10
+4
-200
Time (ms)
400
Fz F3
F7
FC5 FC1
C3 T7
CP5 CP1
Pz P3
P7
O1 Oz O2
P4
P8
CP6 CP2
Cz C4 T8
FC6 FC2
F4
F8
-10
+4
-200
Time (ms)
400
Fz F3
F7
FC5 FC1
C3 T7
CP5 CP1
Pz P3
P7
O1 Oz O2
P4
P8
CP6 CP2
Cz C4 T8
FC6 FC2
F4
F8
-10
+4
-200
Time (ms)
400
VoicedVoiceless
Experiment 2 Experiment 3 Experiment 4
badge padge baint paint
Onset and coda of each stimulus cross-spliced to control for acoustic differences; e.g., Experiment 2: