Poster C37, Friday, August 17, 10:30 am – 12:15 pm, Room 2000AB

Perilesional white matter microstructure and aphasia recovery

Emilie T. McKinnon1, Jens H. Jensen1, Chris Rorden2, Alexandra Basilakos2, Ezequiel Gleichgerrcht1, Julius Fridriksson2, Joseph A. Helpern1, Leonardo Bonilha1;1Medical University of South Carolina, 2University of South Carolina

Language therapy has shown to be effective in chronic aphasia, although treatment responses are variable, and predicting outcomes remains challenging. Research has previously shown that left hemisphere perilesional areas contribute to therapy-induced neuroplasticity, but little is known about how perilesional white matter affects treatment outcome. The current study attempts to address this gap in knowledge by relating pretreatment residual perilesional white matter integrity to post-treatment changes in confrontational naming. Thirty-one subjects (age=57±11y; 9 females; MRI time post-stroke=36±32m) with chronic post-stroke aphasia underwent the Western Aphasia Battery Revised, multimodal MRI, and 15 sessions (5x/week) of self-administered computerized anomia treatment (45 min). Half of the participants received anodal tDCS (1mA) during the first 20 minutes, and the remaining half received sham stimulation. Aphasia severity ranged from severe to mild (WAB-AQ = [20-92]), with an average (± standard deviation) WAB-AQ of 52 (± 22). Structural (T1-w, T2-w) and diffusional kurtosis images (b=0, 1000, 2000 s/mm2) were acquired at baseline and diffusional kurtosis estimator was used to estimate diffusion and kurtosis tensors. A perilesional white matter mask was created by dilating lesion drawings (4 voxels) including only those voxels categorized as white matter (>90% probability) by enantiomorphic segmentation. Voxels dominated by cerebral spinal fluid (mean diffusivity > 2 μm2/ms) were excluded. Before treatment, subjects elicited an average of 24% (± 25%) correct responses during a confrontational naming test consisting of 80 objects. After treatment, participants demonstrated a significant proportional change in correct responses on trained items with an average improvement of 12% ± 19%. We found that both perilesional axial diffusivity and kurtosis significantly related to changes in correct response (r=0.47, p<0.05; r=-0.49, p<0.05) correcting for lesion size, pre-treatment aphasia severity and tDCS application. Perilesional fractional anisotropy, radial diffusivity and radial kurtosis did not reflect naming improvements (r=0.2, p>0.05; r=0.28, p>0.05; r=-0.33, p>0.05). In addition, axial diffusivity/kurtosis was associated with absolute changes in semantic but not phonemic paraphasias (r=-0.49, p<0.05; r = 0.44; p<0.05) in a subset of temporal lobe white matter voxels. This study shows preliminary evidence on the relationships between baseline perilesional white matter and language recovery, and demonstrates the possible role image analysis can play in predicting recovery potential from baseline neuroimaging data. If pretreatment MRI contains valuable information about treatment response, then imaging has the potential to become a useful tool for guiding clinical management of aphasia. Specifically, these results suggest it may be possible to tap into specific recovery potentials, such as comparing room for semantic and phonemic improvement.

Topic Area: Language Therapy