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

Poster B31, Wednesday, November 8, 3:00 – 4:15 pm, Harborview and Loch Raven Ballrooms

Lesion predictors of response to semantically-based naming treatment in chronic aphasia

Michelle Gravier1, Michael Dickey1,2, William Hula1,2, Patrick Doyle1,2;1VA Pittsburgh Healthcare System, 2University of Pittsburgh

INTRODUCTION Several recent studies have implicated lesion site in naming treatment response in chronic aphasia, with better outcomes being variously related to more intact frontal (Abel et al., 2015; Marcotte et al., 2015), temporo-parietal (Bonilha et al., 2016; Fridriksson, 2010), or subcortical (Parkinson et al., 2009; Meinzer et al., 2010) brain regions. Independently, Dell and colleagues (2013) found that lesion in distinct regions predicted parameters related to semantic and phonological processing (s- and p-weight) in the interactive two-step model of naming (Schwartz et al., 2006). These brain areas partially but not completely aligned with the frontal/temporo-parietal/subcortical distinction. If lesions to s-/p-weight volumes of interest (VOIs) underlie naming impairments and help determine naming-treatment response, this might help explain the inconsistent findings above. Therefore, the current study investigated whether Dell et al’s parameter VOIs or frontal/temporal-parietal/subcortical VOIs predicted response to semantically-focused naming treatment. METHOD Participants: Fourteen adults with chronic aphasia following single left-hemisphere stroke received intensive semantic feature analysis (SFA; Boyle & Coelho, 1995) treatment as part of an ongoing clinical trial. Lesion Characterization: Prior to treatment, all participants completed a T1-weighted structural MRI scan. Lesions were manually segmented using ITK-SNAP (Yushkevich et al., 2006) and normalized to the MNI template using the SPM12 clinical toolbox (Rorden et al., 2012).VOIs corresponding to s-weight and p-weight voxel-lesion parameter maps from Dell et al (2013) and frontal/temporo-parietal/subcortical VOIs based on Crosson et al. (2009) were defined via the Automated Anatomical Label Atlas (AAL, Desikan et al., 2006). Overlap between lesions and VOIs (lesion proportion) was calculated using MarsBaR (Brett et al., 2002). Analysis: The outcome measure was naming-probe accuracy for treated and semantically-related untreated items at treatment entry, exit, and one-month follow-up. Item-level data were analyzed using multilevel generalized linear regression with a logistic link function in R. Each lesion VOI was tested in a separate model including main effects and interactions of lesion proportion, probe time (entry/exit or exit/follow-up), and item type (treated/untreated). Models also included total lesion size as a covariate and random intercepts for items and participants. RESULTS Entry/exit: The main effect of lesion proportion in s-weight VOI approached significance when controlling for total lesion size (p = 0.075), but no two- or three-way interactions were significant. No other VOIs had significant effects. Exit/follow-up: The main effect of s-weight lesion proportion remained nearly significant (p=0.08). Two-way interactions between subcortical VOI lesion and probe time (p=0.04) and probe type (p=0.02) were significant. DISCUSSION Larger lesions in areas linked to s-weight (the parameter representing efficiency of connections between semantic and lexical representations) predicted worse naming overall, but did not predict treatment-related changes. This contrasts with recent findings that s-weight positively predicted SFA-related improvement for treated and untreated items (Dickey et al., in prep). Interestingly, greater lesion in subcortical VOI predicted better maintenance for treated items but worse maintenance of generalization effects. This suggests that subcortical structures may play an important and sometimes surprising role in longer-term maintenance of treatment effects (Meinzer, et al., 2010; Parkinson et al., 2009).

Topic Area: Language Disorders

Back to Poster Schedule