Search Abstracts | Symposia | Slide Sessions | Poster Sessions
Therapy-Induced Functional Brain Network Reorganization in Post-Stroke Aphasia
Poster Session E, Sunday, September 14, 11:00 am - 12:30 pm, Field House
louisa Suting1, Nabin Koirala2,3,4, Jennifer Mozeiko1; 1Dept. of Speech, Language, and Hearing Sciences, University of Connecticut, 2Child Study Center, Yale School of Medicine, 3Brain Imaging Research Core, University of Connecticut, 4Nathan Kline Institute for Psychiatric Research, New York
INTRODUCTION: Graph theoretical analysis of rsfMRI helps examine brain network organization and its post-stroke alterations. In aphasia, pre-treatment network efficiency and segregation gains predict language recovery, but treatment-induced longitudinal changes in both integration and segregation remain unexamined. We address this gap by investigating how Constraint Induced Aphasia Therapy (CIAT) reshapes network measures supporting language recovery by evaluating network properties pre-treatment, immediate post-treatment, and at eight weeks follow-up. METHODS: Nine adults (>6 months post-stroke) completed 15 hours/week of CIAT over two weeks. Behavioral assessments and 30 minute rsfMRI scan were collected at three-time points. rsfMRI data were preprocessed using SPM12, including coregistration, enantiomorphic normalization, segmentation, smoothing, nuisance regression, and bandpass filtering. Functional images were parcellated into 384 homotopic regions (AICHA atlas), and connectivity matrices were computed across multiple network densities (5–50%). Graph theoretical measures were calculated, including global/local efficiency, clustering coefficient, modularity, degree and betweenness centrality. Density-specific permutation tests (5,000 iterations) identified significant effects in network measures, and partial Spearman correlations- controlling for demographics, lesion volume, and chronicity- linked these changes to behavioral gains. RESULTS: Network Reconfiguration— CIAT led to significant increases in network segregation (clustering, modularity) and hub metrics from baseline to post-treatment. Clustering coefficient increased from 0.293 ± 0.045 to 0.343 ± 0.053 (significant across densities 0.10–0.50; p < .05), and modularity increased from 0.425 ± 0.068 to 0.469 ± 0.050 (densities 0.15–0.50; p < .05) indicating higher network segregation post-treatment. Betweenness centrality, a measure of nodal influence, also increased, from 555.9 ± 47.1 to 594.7 ± 65.5 (densities 0.10–0.30; p < .05). At eight weeks of follow-up, gains in segregation and hub influence (increases in degree centrality) persisted. Clustering remained elevated at 0.344 ± 0.076 (densities 0.05 and 0.15–0.50; p < .05), and modularity held steady at 0.470 ± 0.067 (densities 0.10–0.40; p < .05). Degree centrality also increased compared to the baseline, from 64.1 ± 4.9 to 66.9 ± 6.5 (densities 0.05–0.25; p < .05). Integration measures including, local efficiency climbed significantly to 0.406 ± 0.056 (densities 0.15–0.50; p < .05), while global efficiency showed a modest but significant shift at density 0.45 (0.246 ± 0.024 → 0.241 ± 0.024; p < .05). Association to Behavioral Improvement— Post-treatment increases in modularity strongly predicted language recovery. Change in modularity correlated with verb naming gains (ρ = .95, q < .001) and noun naming gains (ρ = .80, q < .05). Within executive functions, faster completion of Trail Making B was linked to increases in clustering (ρ = –.98, q < .001) and betweenness centrality (ρ = –.97, q < .001). At follow-up, the association between change in clustering and verb naming remained robust (ρ = .97, q < .001), and gains in local efficiency continued to predict faster Trail Making A performance (ρ = –.87, q < .05). CONCLUSION: CIAT drives immediate and sustained changes in network segregation and hub connectivity, which strongly predict language and executive gains, while network integration properties evolve more gradually. rsfMRI graph metrics could provide objective biomarkers to guide personalized aphasia rehabilitation.
Topic Areas: Disorders: Acquired, Speech-Language Treatment