Introduction: Recovery in the subacute phase of stroke involves rapid network reorganization, but detecting these changes is hindered by imaging noise. Multi-echo resting-state fMRI (ME-rsfMRI), which models T2* decay across 3 echoes, combined with independent component analysis (ICA), enables detection of echo-specific signatures that reflect patient recovery post-stroke. This approach may reveal restorative processes in the lesioned hemisphere and compensatory mechanisms contralaterally. This study examined whether ME-rsfMRI with ICA can capture connectivity improvements of unique signatures during subacute stroke recovery. Methods: Nine male participants (mean age 68 ± 8.5) with aphasia and left hemisphere subacute stroke lesions underwent ME-rsfMRI during early (2–10 weeks) and late (10–16 weeks) subacute phases. Independent components were denoised, ranked on a scale, assigned to brain regions, and matched to participants with overlapping lesions. Longitudinal changes in rankings were qualitatively assessed to characterize recovery-related network patterns. Results: Connectivity improvements in memory- and salience-related systems were detected in Echo 1 (7/9 participants), involving deep limbic and paralimbic hubs such as the insula, thalamus, and medial temporal regions. Echo 2 (5/9 participants) captured higher-order language and executive control networks spanning frontoparietal and temporal cortices, reflecting domains critical for communication and cognitive regulation. Echo 3 (4/9 participants) revealed specific engagement of perilesional language and associative hubs, including inferior frontal, temporal, and parietal regions, consistent with compensatory mechanisms. Discussion: These findings highlight the clinical utility of ME-rsfMRI with ICA in tracking specific network reorganization during stroke recovery. By capturing distinct but complementary networks, ranging from memory and salience to language, executive, and associative systems, the multiple echoes provide a layered view of neural reorganization, underscoring the value of ME-rsfMRI in tracking recovery trajectories. Future work will quantify lesion–component overlap to improve translational relevance. Ultimately, this framework could bridge advanced imaging with individualized rehabilitation planning in stroke care.
Ferro et al. (Thu,) studied this question.