Predicting language recovery in post-stroke aphasia using behavior and functional MRI

Abstract Language outcomes after speech and language therapy in post-stroke aphasia are challenging to predict. This study examines behavioral measures resting state fMRI (rsfMRI) as predictors of treatment outcome. Fifty-seven patients with chronic were recruited treated for one three impairments: anomia, agrammatism, or dysgraphia. Treatment effect was measured by performance on a treatment-specific measure, assessed before months therapy. Each patient also underwent an additional 27 assessments rsfMRI scan at baseline. Patient scans decomposed into 20 components group independent component analysis, the fractional amplitude low-frequency fluctuations (fALFF) calculated each time series. Post-treatment modelled elastic net regression, using pre-treatment either fALFF imaging predictors. Analysis showed strong anomia (R 2 = 0.948, n 28) agrammatism 0.876, 11) dysgraphia 0.822, 18). Models trained features may outperform models some populations. suggests that have prognostic value outcomes.