Investigating acoustic startle priming-induced cortical network reorganization in post-stroke dysarthria: a fNIRS...

Post-stroke dysarthria (PSD) involves impaired speech motor initiation and disrupted large-scale cortical coordination, but the neural effects of acoustic startle priming (ASP) during rehabilitation remain unclear. We conducted a randomized, double-blind, sham-controlled trial in 28 individuals with subacute PSD who received 2 weeks of ASP-based or sham speech training combined with conventional rehabilitation. Multimodal evaluation included Frenchay dysarthria assessment (FDA), acoustic analysis, and 53-channel functional near-infrared spectroscopy (fNIRS) during task and resting states; 26 participants were retained for final imaging analyses. After adjustment for baseline FDA scores, the ASP group showed lower post-intervention FDA scores than controls, indicating reduced dysarthria severity (F(1,24) = 7.49, p = 0.012, partial η² = 0.238). Acoustic analysis showed limited main effects overall, whereas significant Time × stimulation condition and Vowel × stimulation condition interactions were observed for selected parameters (both p < 0.01). Task-based fNIRS revealed increased oxygenated hemoglobin responses after ASP, with a greater change in channel 51 in the contralesional language network relative to controls (t = 2.25, p = 0.037). In resting-state analyses, within-group increases in connectivity between the ipsilesional sensorimotor and contralesional language networks and between the contralesional default mode and sensorimotor networks survived false discovery rate correction (both q = 0.017), whereas between-group connectivity changes did not. Clinical improvement correlated positively with ipsilesional sensorimotor-language connectivity (r = 0.612, p = 0.038) and negatively with ipsilesional default mode oxygenation (r = -0.759, p = 0.009). These findings suggest that ASP may enhance speech rehabilitation in PSD and that multimodal fNIRS can capture associated changes in language-motor network organization, while network-level between-group effects remain preliminary.