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

 
Poster E18, Friday, November 10, 10:00 – 11:15 am, Harborview and Loch Raven Ballrooms

Agrammatic performance in aphasia: A ventral-stream problem?

Dirk-Bart Den Ouden1, Alexandra Basilakos1, Leo Bonilha2, Ezequiel Gleichgerrcht2, Svetlana Malyutina3, Chris Rorden1, Julius Fridriksson1;1University of South Carolina, 2Medical University of South Carolina, 3National Research University Higher School of Economics, Moscow

Introduction: Agrammatism in aphasia is not a homogeneous syndrome, but rather a characterization of a non-uniform set of language production and comprehension behaviors in which grammatical markers and complex syntactic structures are omitted, simplified or misinterpreted. Likely for this reason, a neural predictor of agrammatism has been elusive, with different neurolinguistic models of syntactic processing predicting damage to inferior frontal cortex, anterior temporal cortex, or posterior superior temporal cortex to result in agrammatic performance. Two recent dual-stream models of language processing emphasize a dorsal pathway between inferior frontal gyrus and posterior superior temporal gyrus or sulcus (pSTG) to support the assignment of grammatical relations or syntactic structuring (Friederici, 2012; Bornkessel-Schlesewsky & Schlesewsky, 2013). Following these models, damage along this pathway and/or its end nodes should result in agrammatism. We performed multimodal lesion-symptom mapping (LSM) to investigate the association between agrammatic performance, as reflected in a variety of syntax-related variables, and structural regional and connectivity damage after stroke. Methods: 75 Stroke survivors participated in this study (all > 6 months post-stroke; mean age 59.5, sd 10.1; 25 females). 53 Participants had aphasia (mean WAB AQ 67.0, sd 21.1), of whom 13 were classified as agrammatic, based on picture-description data, with a focus on omission of grammatical morphemes. Participants performed subtasks of the Northwestern Assessment of Verbs and Sentences to assess production of verb argument structure patterns and different cued sentence types, as well as comprehension of different sentence types. All underwent neuroimaging (Siemens 3T MRI), collecting T1, T2 and Diffusion Tensor Weighted images. Lesions were manually drawn based on visual examination of structural images. After regressing out lesion sizes from the behavioral data, voxel- and connectome-based LSM was performed using linear regression analysis with permutation thresholding and a one-tailed alpha level of .05. Results: After regressing out lesion size, no individual (sub)cortical region was predictive of ‘agrammatism’ as a binary category. Impaired performance on verb argument structure production was predicted by damage to pSTG, as well as to a small section of white matter underlying middle superior temporal gyrus, inferior to the Sylvian fissure and typically part of a ventral pathway between temporal and inferior frontal cortex. Damage to this region was also predictive of impaired sentence production and comprehension, as well as of relatively greater impairment on noncanonical than canonical sentence structures in cued production, which was also predicted by pSTG lesions. Whole-brain connectome effects did not survive the regressing out of lesion size, but we are following up with analyses of specific connections of interest. Conclusion: Results confirm that no single lesion pattern is predictive of the clinical impression of ‘agrammatism’ as a syndrome. However, specific tasks that require different levels and types of syntactic processing do rely on certain regions to be intact, most notably pSTG and parts of a ventral pathway between posterior temporal and inferior frontal cortex. Interestingly, this ventral pathway is not typically considered to support syntactic processing critically, so these clinical data shed new light on the neural architecture underlying syntactic processing.

Topic Area: Language Disorders

Back to Poster Schedule