Presentation

Xin Ran Li¹, Sara B. Pillay¹, Joe Heffernan¹, Lisa Conant¹, Priyanka Shah-Basak¹, Anna Frieberg¹, Shelley Laitinen¹, Samantha Hudson¹, Jed Mathis¹, Sabine Heuer², Roy Hamilton³, Jeffrey R. Binder¹; ¹Medical College of Wisconsin, ²University of Wisconsin-Milwaukee, ³University of Pennsylvania

Introduction: Damage to the phonological system is common after left hemisphere stroke, resulting in difficulty retrieving correct sounds for speech production. Given prior evidence that left angular gyrus (AG) activation is associated with better oral reading performance in people with this deficit (Pillay et al., J Cogn Neurosci, 2018), we hypothesized that stimulation of the AG with transcranial direct current (tDCS) may promote reorganization of networks relevant to phonological retrieval. We coupled high-definition tDCS (HD-tDCS) with phonological component analysis (PCA) therapy, in which patients explicitly identify phonological features of a target word. We used resting state functional MRI (rs-fMRI) to test whether these treatments induce persisting changes in functional connectivity. Methods: Three patients with chronic left hemisphere stroke and phonological impairment received PCA therapy during either active anodal HD-tDCS to the left AG or sham/control tDCS for 10 days of therapy, then crossed over to the opposite stimulation arm. Patients were trained on reading and picture naming tasks using items they were reliably unable to name at baseline, then tested on trained and untrained items after therapy. RS-fMRI and language assessments were acquired pre-treatment, 10 weeks post-treatment cycle 1, and 10 weeks post-treatment cycle 2. The 20 minutes of rs-fMRI data from each session were denoised using ICA-AROMA. Pairwise connectivity was computed between Brainnetome atlas parcels comprising the left semantic (Lsem) and phonological (Lphon) networks and their right hemisphere homologs (Rsem and Rphon). Results: After AG HD-tDCS, patient 004 showed significant connectivity increases within the Lsem, Lphon, and Rphon networks. Connectivity within the Rsem network did not change, and connections between sem and phon networks became negative (inhibitory) in both hemispheres. Inter-hemisphere connectivity increased between Lsem and Rsem, and between Lphon and Rphon. After sham-tDCS, there were small connectivity increases in Lsem, Rsem, and Rphon networks, while connectivity reverted to baseline in Lphon, and sem-phon connectivity became less inhibitory. Inter-hemisphere connectivity decreased between left and right phon networks. The patient improved approximately 40% on untrained items. After AG HD-tDCS, patient 008 showed significantly decreased connectivity within Lphon and Rphon networks, and between sem and phon networks in both hemispheres. Inter-hemispheric connections, which were strongly inhibitory at baseline, became less inhibitory. After control-tDCS, most intra-hemisphere connectivity continued to decrease, while Lphon connectivity returned to baseline levels, and inter-hemispheric connections became even less inhibitory. The patient showed modest improvements to roughly 20% correct on untrained items after both treatment cycles. After AG HD-tDCS, patient 052 showed significant intra-hemisphere connectivity increases in all networks except Rphon. Inter-hemisphere connectivity increased between Lsem and Rsem. The patient showed modest improvements to roughly 10% correct on untrained items, but dropped out of the study prior to cross-over with control-tDCS. Conclusion: In these preliminary studies, fMRI connectivity showed long-lasting changes after PCA therapy combined with HD-tDCS. Inter-hemisphere connectivity increased in all patients, and all showed improved performance on untrained reading and picture naming items. A larger dataset is needed to identify correlations between behavioral and functional connectivity changes.

Topic Areas: Disorders: Acquired, Language Therapy