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

Poster A33, Wednesday, November 8, 10:30 – 11:45 am, Harborview and Loch Raven Ballrooms

White matter matters in the recovery of language in post-stroke aphasia

Erin Meier1, Jeffrey Johnson1, Yansong Geng1, Swathi Kiran1;1Boston University, Sargent College of Health and Rehabilitation Sciences

Research has shown that lesion volume is a significant predictor of language outcomes in persons with aphasia (PWA) (e.g., Lazar et al., 2008). However, recent evidence suggests that lesion site, and in particular, the integrity of white matter (WM) in specific regions of interest (ROIs) is a better indicator of language skills like fluency and semantics (Basilakos et al., 2014; Pani et al., 2016). However, the impact of damage to left hemisphere (LH) WM regions and the potential compensation by right hemisphere (RH) homologues has not been explored in the context of naming abilities in PWA. Therefore, this study addressed two research questions (RQs): (1) To what extent does integrity of LH WM regions differ from RH regions? (2) What is the relationship between the integrity of bilateral WM ROIs and language skills? Fifteen PWA (9M, mean age=61.7 years) underwent a high-resolution, whole-brain DTI scan. Regions implicated in semantic and phonological stages of naming (Indefrey & Levelt, 2004) were selected as ROIs from the Harvard-Oxford atlas, including superior and middle frontal gyri (SFG, MFG); inferior frontal gyrus pars triangularis (IFGtri) and opercularis (IFGop); anterior and posterior middle temporal gyri (aMTG, pMTG), anterior and posterior superior marginal gyri (aSMG, pSMG), and angular gyrus (AG). An FA template thresholded at .20 was superimposed over the aforementioned bilateral cortical ROIs to ensure the extraction of WM only. Fractional anisotropy (FA) in these ROIs was extracted from patient data. To address RQ1, a one-way MANOVA was used to determine if the integrity of LH WM ROIs differed from their RH homologues. For RQ2, linear regression models were conducted, predicting aphasia severity or naming skills from FA, while controlling for lesion volume. Only homologous regions that significantly differed in FA per RQ1 were analyzed in RQ2. For RQ1, the main effect of hemisphere approached significance (F=2.51, p=.056). The univariate results revealed that FA was significantly higher in specific RH versus LH ROIs, including MFG (p=.016), IFGtri (p=.001), IFGop (p<.001), pMTG (p=.005), pSMG (p=.004), and AG (p=.018). For RQ2, when controlling for LH lesion volume, FA in LH regions did not significantly predict aphasia severity (F=2.187, p=.234) or naming skills (F=2.73, p=.174). However, the model predicting aphasia severity from FA in RH ROIs approached significance (F=3.728, p=.065). RH FA values significantly predicted naming skills (F=5.67, p=.025) and lesion volume was a significant factor in this relationship (t=-3.792, p=.009). Univariate results demonstrated different effects by ROI: PWA with greater FA in RIFGtri (t=2.843, p=.029) and RIFGop (t=2.653, p=.038) had better naming abilities while PWA with greater FA in RMFG (t=-2.051, p=.086) and RAG (t=-2.214, p=.069) trended towards poorer naming skills. As expected, FA in the undamaged RH was significantly greater than the LH in several WM ROIs central to language processing in general and naming specifically. Heightened RH FA appeared to be beneficial for naming in certain ROIs but maladaptive in others. These findings expand upon theories regarding potential compensation of specific RH regions in the neural reorganization of language in post-stroke aphasia.

Topic Area: Language Disorders

Back to Poster Schedule