Search Abstracts | Symposia | Slide Sessions | Poster Sessions
TMS Efficacy Related to Topological Properties of Right Pars Triangularis in Chronic Aphasia
Poster Session C, Friday, October 25, 4:30 - 6:00 pm, Great Hall 3 and 4
Harrison Stoll1, Apoorva Kelkar1, Denise Harvey2, Olu Faseyitan2, Branch Coslett2, John Medaglia1,2; 1Drexel University, 2Univeristy of Pennslyvania
Recovery from post-stroke aphasia requires the recruitment of new brain regions. Many of these regions are right hemisphere homotopes of left hemisphere language regions with different structural and functional properties in persons with aphasia (PWA) relative to controls (Saur et al., 2006). One region, the right pars triangularis (rPTR), has been associated with language dysfunction in PWA (Postmancheux et al., 2010) and has been a target in inhibitory transcranial magnetic stimulation (TMS) combined with language therapy (Medina et al., 2012). While a promising intervention, inhibiting the rPTR does not improve language outcomes in all PWA. One approach to identify moderating factors of TMS efficacy is to identify individual differences in the rPTR that correspond to treatment effects in language outcomes. The current study examines the structural properties of the rPTR and their relationship with naming performance after TMS. To address this, we examined single-tract connections from the rPTR to language regions in the right and left hemisphere using measures such as network centrality (e.g., degree—number of connections from one region to others) and whole brain density (e.g., fraction of observed connections between regions, relative to all potential connections). Twenty-nine PWA underwent 10 sessions of inhibitory (1 Hz) TMS to the rPTR followed by 1 hour of Constraint Induced Language Therapy. Twenty patients received active TMS and 9 received sham stimulation. Patients also completed high-resolution MRI scanning. The Philadelphia Naming Test (PNT; Roach et al., 1996) was administered at baseline. Using baseline (pre-treatment) MRI scan, we constructed structural connectomes to estimate edge weights between the rPTR and the left pars triangularis (lPTR), left middle temporal gyrus (lMTG), and right middle temporal gyrus (rMTG). For network centrality measures, we examined degree, betweenness centrality, and eigenvector centrality. We also computed measures for the entire brain, including whole-brain density. In the active and sham group, we correlated each measure with the difference between post- and pre-TMS overall accuracy on the PNT and the proportion of error types (i.e., semantic and phonological). For significant correlations, we ran post hoc regression models to if the relationship for the active group was larger than the sham group. We found a correlation between proportion of phonological errors made and single edge weight from the rPTR to the rMTG (r = -.65, p < .005). Specifically, fewer phonological errors were made after TMS when the patient had a weaker connection from the rPTR to the rMTG. The regression model comparing the active to the sham condition was also statistically significant (b = .11, p < .05), with the active group having a stronger slope than the sham group . Our findings provide evidence that the strength of anatomical connections from the rPTR to the rMTG moderate TMS efficacy on naming in PWA. Specifically, individuals with a weaker connection at baseline derived greater benefit from TMS. The data raises the possibility that the strength of structural connection from the rPTr to the rMTG could predict response to TMS to the rPTr in chronic aphasia.
Topic Areas: Speech-Language Treatment,