You are viewing the SNL 2017 Archive Website. For the latest information, see the Current Website.

Poster C6, Thursday, November 9, 10:00 – 11:15 am, Harborview and Loch Raven Ballrooms

Frontotemporal connectivity during syntactic movement processing

Eduardo Europa1, Darren R Gitelman2,3,4, Swathi Kiran5, Cynthia K Thompson1,2,6;1School of Communication, Northwestern University, 2Feinberg School of Medicine, Northwestern University, 3Advocate Lutheran General Hospital, 4Rosalind Franklin University of Medicine and Science, 5College of Health & Rehabilitation, Boston University, 6Cognitive Neurology and Alzheimer's Disease Center, Northwestern University

Linguistic theory suggests that noncanonical sentences subvert the dominant agent-verb-theme order of arguments in English, such that the theme precedes the verb, via displacement of sentence constituents to argument (NP-movement) or non-argument positions (Wh-movement). The neural model of sentence comprehension by Friederici et al. (2012) proposes that an anteroposterior dorsal pathway, i.e., superior longitudinal fasciculus/arcuate fasciculus, subserves noncanonical sentence processing, with phrase structure building in the left inferior frontal gyrus (IFG) preceding thematic role re-analysis in the left temporo-parietal junction (TPJ) (also see Thompson & Meltzer-Asscher (2014) for a similar model for processing verb argument structure). On the contrary, the sentence comprehension model proposed by Bornkessel-Schlesewsky & Schlesewsky (2013) suggests that all sentences, regardless of complexity, begin with lexical processing in the posterior temporal cortex. Posteroanterior dorsal and ventral projections, subserving syntactic and semantic combinatorial processes, respectively, terminate in the left IFG, where structural and thematic analyses are integrated. These two hypotheses were tested in the present study using Dynamic Causal Modeling (DCM) with 15 cognitively healthy adults during an fMRI auditory sentence-picture verification task using passive and active sentences, contrasted to isolate NP-movement (i.e., Psv>Act), and object and subject cleft sentences, contrasted to isolate Wh-movement (i.e., OC>SC). Noncanonical sentence processing (Psv>Act + OC>SC) elicited peak activation in the left IFG (LIFG), posterior superior temporal gyrus (LSTGp), and the medial superior frontal gyrus (LSFGm), and these regions were used in the DCM analysis. Random-effects family-wise Bayesian Model Selection indicated that models in which neural network activity was driven by the LIFG’s response to sentence stimuli best fit the fMRI data; exceedance probabilities = 0.73 and 0.82 for Wh- and NP-movement, respectively. One-sample t-tests of subject-specific parameters derived from Bayesian Model Averaging suggest that both movement types were associated with a significant increase in connectivity between the LIFG and LSTGp and between the LIFG and LSFGm, p<0.05 corrected for multiple comparisons using false discovery rate. No significant difference in modulation by either movement type was found between the LIFG-LSTGp and LSTGp-LIFG. These findings suggest that Wh- and NP-movement similarly modulate connectivity between regions involved with syntactic movement processing and are in line with previous accounts of Wh-movement modulating fronto-temporal connectivity (den Ouden et al., 2012). Results support the models proposed by Friederici et al. (2012) and Thompson & Meltzer-Asscher (2014), which predict sentence input into the LIFG for phrase structure building, and syntactic movement modulating the LIFG-LSTGp connection for thematic re-analysis in the LSTGp. While the LSTGp-LIFG modulation supports the Bornkessel-Schlesewsky & Schlesewsky (2013) model, given the nature of the task, it more likely reflects response selection in the LIFG (Swick et al., 2008) following thematic re-analysis in the LSTGp. The present study supports the idea that comprehension of noncanonical sentences involves phrase structure building followed by thematic re-analysis, but increased complexity of Wh-structures compared to sentences with NP-movement requires greater engagement of cognitive resources via increased neural activity in LIFG and LSTGp (Europa & Thompson, 2016).

Topic Area: Grammar: Syntax

Back to Poster Schedule