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

Poster A14, Wednesday, November 8, 10:30 – 11:45 am, Harborview and Loch Raven Ballrooms

An ALE-based meta-analysis of neuroimaging studies of sentence comprehension

Matthew Walenski1, Eduardo Europa1, David Caplan4, Cynthia K. Thompson1,2,3;1Department of Communication Sciences and Disorders, Northwestern University, Evanston, IL, USA, 2Cognitive Neurology and Alzheimer’s Disease Center, Northwestern University, Evanston, IL, USA, 3Department of Neurology, Northwestern University, Evanston, IL, USA, 4Massachusetts General Hospital, Department of Neurology, Harvard Medical School, Boston, MA, USA

Introduction. Comprehending a sentence is a complex endeavor that requires the coordinated activity of multiple brain regions. Our goal in this project was to merge findings from imaging studies of sentence comprehension to summarize regions shown to be engaged in language comprehension in healthy individuals. We expect that these would largely involve perisylvian regions of the left hemisphere. We also examined the contrast between non-canonical and canonical sentence structures. Non-canonically ordered sentences subvert the dominant agent-verb-theme order of arguments in English, such that the theme precedes the verb. For example, in an object relative clause (The boy that the girl saw [gap] is …), the theme argument (boy) is extracted from the [gap] position, preceding its verb (saw) and the agent argument (girl). In a subject relative clause (The girl that [gap] saw the boy is …), the dominant order is preserved. For non-canonical sentences, we anticipated involvement of left inferior frontal regions that putatively subserve the complex syntactic processing required for these structures. Method. Literature searches revealed 443 imaging studies (FMRI, PET) of sentence comprehension published between 1992 and 2016. Of these, studies were considered eligible for inclusion in our analyses if they reported whole-brain 3-D coordinates for two or more activation foci in a standard space (Talairach, MNI) from healthy, right-handed, monolingual adults. Of the eligible studies, we first examined those that reported contrasts for sentence comprehension (auditory or visual) against a perceptual baseline (e.g., backwards speech, nonsense fonts). Additionally, we identified studies that reported a contrast of sentences with non-canonical order against those with canonically ordered arguments. Studies with these contrasts were included only if the participants were required to perform a task while comprehending the sentences. We performed an automated estimation likelihood analysis on the published coordinates using the GingerAle software package, with a cluster level inference thresholding method and false discovery rate correction for multiple comparisons. Results. In comparison against the perceptual baseline, 36 contrasts met all criteria for inclusion. The results indicated significant clusters with extrema in left perisylvian regions (temporal pole, anterior and posterior middle temporal gyrus, angular gyrus, inferior frontal gyrus) as well as in the right temporal lobe (temporal pole, anterior and posterior middle temporal gyrus). A follow-up conjunction analysis comparing auditory and visual modalities found areas of common activation in left middle temporal gyrus and left inferior frontal cortex. Thirty-four contrasts between non-canonical and canonical sentences met criteria for inclusion. The results showed significant clusters in left inferior frontal, middle frontal, paracingulate, middle temporal, and angular gyri. No significant clusters were found in the right hemisphere. Follow-up analyses of auditory vs. visual modality differences reveal substantial overlap in left inferior frontal gyrus. Conclusions. The results revealed a set of bilateral regions activated in response to sentence comprehension relative to perceptual baseline conditions, consistent with our expectations. A more restricted subset of left-hemisphere regions was implicated in the comprehension of non-canonical sentence structures, indicating specialized syntactic processing for these structures.

Topic Area: Grammar: Syntax

Back to Poster Schedule