Presentation
Search Abstracts | Symposia | Slide Sessions | Poster Sessions | Lightning Talks
Brain regions associated with rule-based and observational learning abilities in post-stroke aphasia
Poster E57 in Poster Session E, Thursday, October 26, 10:15 am - 12:00 pm CEST, Espace Vieux-Port
Sofia Vallila-Rohter1; 1MGH Institute of Health Professions
Individuals with stroke-aphasia present with reduced comprehension or expression of language. While language is the primary factor that determines the targets and goals of therapy for people with aphasia (PWA), research is increasingly drawing attention to co-occurring cognitive deficits. Importantly, cognition might play a role in response-to-intervention (Gilmore et al., 2019). We examine rule-based and observational (implicit) learning ability in PWA to expand upon work that suggests that method of instruction may play a role in learning success for PWA (Middleton et al., 2015, 2016; Vallila-Rohter et al., 2013) as it does in other clinical populations such as those with Parkinson’s Disease or Amnesia (Knowlton et al., 1994; Filoteo et al., 2005). We apply voxel-based lesion symptom mapping (VLSM; Bates et al., 2003) to better understand how damage to specific brain regions might contribute to reduced rule-based and/or observational learning. Enrolled participants completed two computer-based artificial grammar learning (AGL) tasks: one learned through observation (AGLObs) and another learned via rule instruction (AGLrb). For each AGL task, four unique geometric shapes (circle, pentagon, diamond, triangle) were arranged into three-to-six shape sequences according to AGL rules (Knowlton et al., 1992). For AGLObs training, participants saw 23 unique sequences, each presented twelve times in a passive match, non-match decision task. For the AGLrb, participants learned five rules of the AGL dependency structure through computerized training. Rules were taught one at a time with visual and audio supports and feedback-based practice. Each rule practice was followed by a criterion test. Participants had to achieve 65% accuracy to advance to the next rule. For both tasks, a testing phase immediately followed training in which participants judged the grammaticality of 30 sequences each presented twice. Eligible participants completed an MRI scan on a 3-T Siemens scanner with 32-channel head coil to obtain MPRAGE (176 interleaved slices, 1mm isotropic voxels, TR=2530ms, T1=1200ms) and FLAIR sequences. Lesion masks were manually drawn on axial slices using FSLView and confirmed by a neurologist. Images were skull stripped and spatially normalized to the Montreal Neurological Institute (MNI) template. Fourteen PWA have been enrolled. Eleven have completed all behavioral tasks, six have additionally completed an MRI scan. Average AGLObs and AGLrb learning scores did not differ (MObs= 60.76, SD =14; Mrb= 68.72, SD = 17.98). At the individual level, seven PWA scored above chance on AGLObs. Nine scored above chance on AGLrb, with four learning in one condition but not the other. Whole-brain percent spared tissue has been examined as well as percent spared tissue in regions of interest that include the caudate nucleus, posterior cingulate cortex, inferior frontal gyrus, precuneus and insula. Percent spared tissue in each of these regions currently ranges from 47.2 to 100. Learning ability does not correlate with overall lesion size. Data collection and analyses are ongoing. Results expand on research suggesting that variability in learning arises in PWA. Language ability does not predict learning ability. A better understanding of cognitive, linguistic and neural bases of learning in aphasia may help align language intervention with learning profile.
Topic Areas: Disorders: Acquired, Speech-Language Treatment