Search Abstracts | Symposia | Slide Sessions | Poster Sessions
Beyond the Perisylvian Cortex: Cerebellar Contributions to Sentence Processing in Post-Stroke Aphasia
Poster Session F, Friday, October 2, 2:45 - 4:45 pm, Wangari Maathai
Joshua Sullivan1, Emerson Kropp1, Swathi Kiran1, Maria Varkanitsa1; 1Boston University
Introduction The right posterior cerebellum, particularly Crus I/II, is increasingly recognized as part of the distributed language network (Stoodley & Schmahmann, 2009), with evidence implicating these regions in sentence-level integration and syntactic processing (Xu et al., 2005; Nakatani et al., 2022). In post-stroke aphasia, cerebro-cerebellar functional connectivity (FC) has been associated with naming (Stilling et al., 2025), suggesting that cerebellar-cortical interactions may contribute to residual language function and recovery. However, whether these interactions support sentence comprehension in aphasia, particularly as syntactic demands vary, remains poorly understood. Here, we test whether resting-state FC between right Crus I/II and cortical language regions predicts performance across grammaticality-judgment and syntactic-comprehension tasks. Methods Forty individuals with chronic left-hemisphere stroke and aphasia (mean age = 57.2; education = 15 years; mean Aphasia Quotient [AQ] = 80.1) completed resting-state fMRI and two English sentence-processing tasks: grammaticality judgment and sentence– picture matching. The comprehension task included simple structures (actives, subject- extracted clefts) and complex structures (object-extracted clefts, object relatives), presented separately in auditory and visual modalities. Behavioral effects of condition and modality were tested using Wilcoxon signed-rank tests. Resting-state FC was computed between right Crus I/II and 38 bilateral cortical language regions, with left cerebellar lobules IV/V included as a control seed. LASSO regression identified FC predictors of task accuracy; retained connections were then entered into covariate-adjusted linear models controlling for age, education, lesion volume, and AQ. Results Participants showed higher accuracy for grammatical vs. ungrammatical sentences (p<0.001) and for simple vs. complex syntactic structures (p<0.001). There were no differences in accuracy between auditory and visual trials, so modalities were averaged for subsequent analyses. For the grammaticality-judgment task, LASSO selected no cerebellar connectivity predictors for grammatical sentences. In contrast, connectivity effects emerged selectively for ungrammatical sentences: Better performance was associated with stronger connectivity between right Crus II and left anterior middle temporal gyrus and right planum temporale, whereas poorer performance was associated with stronger connectivity between right Crus II to left posterior inferior temporal gyrus, left angular gyrus, and right supplementary motor area. For the syntactic-comprehension task, for simple sentences, better accuracy was associated with stronger right Crus II connectivity to the right insula, whereas poorer accuracy was associated with stronger connectivity to left posterior middle temporal gurys, left angular gyrus, right cerebellar lobule 7b, and left supplementary motor area. For complex sentences, better accuracy was associated only with stronger right Crus II connectivity to right planum temporale, while poorer accuracy was associated with stronger connectivity to right inferior frontal gyrus pars opercularis, left anterior middle temporal gyrus, and left Crus II. The left cerebellar control seed showed no predictors for simple sentences, but three negative intracerebellar associations for complex sentences. Conclusions Together, these findings suggest that the cerebellum contributes to sentence processing when the language system is under increased demand. Cerebellar connectivity may therefore support residual language function and contribute to recovery after stroke. However, cerebellar–cortical connectivity was not uniformly beneficial, suggesting that efficient network organization, rather than greater connectivity alone, is critical for language function.
Topic Areas: Disorders: Acquired, Syntax and Combinatorial Semantics