Search Abstracts | Symposia | Slide Sessions | Poster Sessions
Longitudinal functional neuroimaging of recovery from aphasia in the first year after stroke
Poster Session B, Friday, October 25, 10:00 - 11:30 am, Great Hall 3 and 4
Anna V. Kasdan1, Lily Walljasper1, Marianne Casilio1, Sarah M. Schneck1,2, Deborah F. Levy1,3, Jillian L. Entrup1, Sydney Franklin1, Michael de Riesthal1, Stephen M. Wilson1,4; 1Vanderbilt University Medical Center, 2Shirley Ryan Ability Lab, 3University of California San Francisco, 4University of Queensland
Most individuals with aphasia due to stroke recover over time to varying extents, with lesion location being the most important predictor of recovery[1]. This recovery process shows a decelerating timecourse, with the greatest gains made in the first month, but with continued improvements throughout the first year and into the chronic phase. Recovery is thought to reflect neuroplasticity, or functional reorganization of surviving brain regions, but the nature of this putative process is largely unknown. To investigate the neural correlates of language recovery after stroke, we are carrying out a longitudinal project in which individuals who have experienced an acute, left-hemisphere supratentorial stroke are studied at 4 timepoints: 2-5 days (T1), 1 month (T2), 3 months (T3), and 1 year (T4) post-stroke. Speech and language abilities are evaluated at each timepoint using the Quick Aphasia Battery (QAB)[2], and fMRI is completed at T2-T4 wherever possible. Participants perform an adaptive language matching fMRI paradigm: the comparison between the language (semantic matching) and perceptual (symbol matching) conditions is a reliable and valid contrast for use in people with aphasia[3]. Here, we report data from a total of 61 patients with longitudinal fMRI data (either 2 or 3 timepoints from T2, T3, T4). Data were processed with standard procedures using AFNI, FSL, and FMRISTAST. Group analyses were thresholded at p<.01 and then corrected for multiple comparisons with permutation testing (n=1000) within bilateral perisylvian regions of interest[4]. Behaviorally, patients improved on the QAB overall measure from 1 to 3 months (T2: 7.2+2.6; T3: 7.8+2.5; ΔQAB = 0.6 + 0.8, p<.001) and from 3 to 12 months (T4: 8.1+2.2; ΔQAB = 0.3 + 0.5, p=.008). First, we explored group-level maps at each timepoint. Language maps appeared quite similar at 1 month (n=51), 3 months (n=56), and 1 year (n=39), all revealing a core left-hemisphere fronto-temporal language network, not dissimilar from language maps in neurotypical controls. Next, we looked at changes in activation across time. From 1 month to 3 months (n=46), there was increased functional activation in the left inferior frontal gyrus pars orbitalis (extent=7,472 mm3, p=.003) and some subthreshold activation along the posterior superior temporal sulcus (extent=1,944 mm3, p=.08). There were no significant changes in functional activation between 3 months and 1 year (n=34), despite continued behavioral gains. Overall, activation changes were modest and did not fully account for the substantial behavioral recovery that was observed. Future work will explore whether there are changes in the capacity of spared language regions to process language, which may not be reflected in changes in activation magnitude or extent across recovery. [1] Wilson et al., Brain. 2023; 146(3): 1021-39. [2] Wilson et al., Plos One. 2018; 13(2): e0192773. [3] Wilson et al., Hum Brain Mapp. 2018; 39(8): 3285-307. [4] Casilio et al., Brain. 2024; in press.
Topic Areas: Disorders: Acquired,