Slide Slam D4
Hub functional connectivity differences in the dual-stream language network in stroke survivors with aphasia: a resting-state fMRI study
Haoze Zhu1, Megan C. Fitzhugh2, Lynsey M. Keator3, Lisa Johnson3, Arianna N. LaCroix4, Julius Fridriksson3, Corianne Rogalsky1; 1Arizona State University, Tempe AZ, 2University of Southern California, Los Angeles, CA, 3University of South Carolina, Columbia, SC, 44Midwestern University, Glendale AZ
The dual-stream model of speech processing has become a prevailing framework to explore the critical neural substrates for speech and language recovery in stroke survivors with aphasia. Yet, very little is known regarding the network-level organization of the dual-stream model’s regions, and how this organization may be disrupted in aphasia. Previous work indicates that resting-state functional connectivity is a promising tool to understand the network-level properties of the dual-stream regions, yet our previous work indicates that functional connectivity across the entire dual-stream network does not on average differ in stroke survivors with aphasia versus controls, or as a function of severity of language impairments. But, graph-theoretical approaches may be a more sensitive method to better understand how dual-stream network organization is disrupted in individuals with aphasia. Thus, in the current study, 28 neurotypical adults (20-79 years, native English-speaking, right-handed) and 28 left hemisphere stroke survivors (35-78 years, native English-speaking, right-handed), underwent resting-state functional MRI and structural MRI. Speech and language abilities of the stroke survivors were assessed using the Western Aphasia Battery. SPM12 and the Brain Connectivity Toolbox were used to calculate mean functional connectivity measures for 14 dorsal stream and 18 ventral stream nodes as identified by previous task-based functional MRI studies (Labache et al., 2019). In-house Matlab and R scripts then were used to identify hub nodes within the dual-stream network in the control subjects, by determining the shortest average path lengths using the small-world network assumption. This procedure identified the left precentral gyrus, left pars triangularis, bilateral posterior superior temporal sulcus, and bilateral middle temporal gyrus as the hub regions of the dual-stream network. The average functional connectivities of these hub nodes then were compared between the control and stroke survivor groups, and multiple regression models were used to predict each stroke survivor's performance on naming, auditory comprehension, spontaneous speech rate and repetition measures. The hub connections within the left dorsal stream and the connections between the hubs of the left dorsal and ventral streams were found to be significantly lower in the stroke survivor group compared to the controls. The regression models, after controlling for age, gender, education, and lesion size, found the following: functional connectivity of left dorsal-ventral hubs was a significant predictor for spontaneous speech rate, left dorsal and bilateral ventral hub connectivities were the significant predictors of repetition, and bilateral ventral and right dorsal-ventral were the significant predictors of naming. Overall, our results suggest that graph theory, and particularly examining the integrity and functionality of hub regions, may be particularly valuable in characterizing the critical disruptions to the dual-stream network in left-hemisphere stroke survivors with aphasia.