On the dynamics of structural connectivity of the motor network during interactive rehabilitation in the early recovery period after stroke

Stroke leads not only to local damage but also to large-scale reorganization of brain networks, including motor regions and interhemispheric interactions. Modern diffusion MRI methods allow for the study of microstructural dynamics in white and gray matter; however, the classical tensor model has limitations in describing complex fiber architectures. The introduction of kurtosis-based approaches increases sensitivity to tissue heterogeneity, making it a promising tool for analyzing neuroplasticity in the context of rehabilitation. The objective of the research is to provide a quantitative description of the dynamics of diffusion-based metrics (both tensor and kurtosis) in the motor network under functional MRIand EEG-guided neurofeedback and to identify prognostic markers of therapeutic efficacy of interactive neurorehabilitation. Material and methods . Fourteen patients (10 men, 4 women; mean age 58 years) who had experienced ischemic stroke within the previous 6 months were enrolled. Patients were randomized into the main group (n = 7), which underwent an fMRI-EEG neurofeedback course, and the control group (n = 7), which received only standard rehabilitation. Each participant underwent three consecutive stages of clinical and neuroimaging assessment (before therapy, immediately after the course, and at 6-month follow-up). Diffusion tensor imaging (DTI) was used for tract reconstruction, and diffusion kurtosis imaging (DKI) – for analyzing kurtosis parameters in gray and white matter, including the premotor and supplementary motor cortices as well as the corpus callosum. Results. In gray matter, late-stage follow-up revealed a decrease in kurtosis fractional anisotropy (kFA), more pronounced in the control group, interpreted as a decline in microstructural organization. In white matter, signs of axonal loss and demyelination were observed in the lesion area, accompanied by fiber remodeling detected by tractography. The main group demonstrated a trend toward the formation of new interhemispheric connections via the II–III segments of the corpus callosum, while the control group showed predominantly densification of pre-existing pathways. Clinical outcomes improved significantly more in the main group (modified Rankin scale, Fugl–Meyer scale, handgrip strength, Box-and-Blocks test, CES-D questionnaire (Center for Epidemiologic Studies Depression Scale)). Discussion. The findings confirm the complementarity of DTI and DKI: the former is more sensitive to myelination and axonal integrity, while kurtosis enables detection of cortical microstructural changes. Correlations between diffusion metrics and functional outcomes highlight their prognostic value in assessing rehabilitation efficacy. Conclusions. Interactive neurorehabilitation promotes structural reorganization of the motor network and leads to significant clinical improvement. Diffusion metrics, including fractionation anisotropy, radial diffusion, and kFA, may serve as objective biomarkers of neuroplasticity and therapeutic effectiveness in the early recovery period after stroke.