The reorganization mechanism of brain functional networks by long term actual exercise and exercise imagination training: spectrum and network topology analysis based on EEG

Background: Motor learning optimizes motor skills and promotes neuroplasticity through motor execution (ME) and motor imagery (MI). However, the mechanisms underlying long-term training-induced reorganization of functional brain networks remain unclear. Existing studies predominantly focus on short-term effects, and their analytical methods are limited to spectral properties, with insufficient exploration of complex network interactions. Elucidating differential regulation of functional brain networks by different training modes is important for understanding the characteristics of motor learning stages and optimizing rehabilitation strategies. Methods: This randomized controlled trial compares seven-day ME and MI training effects on cortical reorganization and behavioral performance. Forty-eight healthy right-handed participants were randomly assigned to four groups: ball rotation task (BRT), video guide imagery (VGI), tactile stimulation (Tactile), and control group (Control). All groups underwent a one-week training intervention. Behavioral performance was quantified by ball-rotation scores, while brain function was evaluated using resting-state electroencephalography (EEG) pre- and post-training. Power spectral density (PSD) in theta (4–8 Hz), alpha (8–13 Hz), and beta (13–30 Hz) bands, phase-lag index (PLI) connectivity, and graph-theoretical metrics (clustering coefficient, node degree) were analyzed. Statistical comparisons used paired t-tests, ANOVA, Kruskal–Wallis tests and post hoc test (α = 0.05). Results: All groups showed significant motor improvement (p < 0.001), with BRT demonstrating superior gains versus others (F = 32.279, p < 0.001, η² = 0.6876). Frontal β-band power increased significantly in BRT, suggesting enhanced skill output via neuronal recruitment, whereas VGI exhibited attenuated β-power, indicating optimized neural resource integration. θ-band frontal-parietal functional connectivity was enhanced in both the BRT and the VGI groups, supporting motor-cognitive integration. Parietal β-band connectivity weakened, implying enhanced neural efficiency and reduced visual dependence. α-band clustering coefficients decreased in BRT, indicating network reorganization toward distributed processing, while θ-band frontoparietal node degree increased in VGI, reflecting elevated cognitive control demands. Conclusion: ME and MI enhance motor skills through distinct network reorganization: ME sustains neuronal recruitment via beta-enhancement and local network dispersion, whereas MI optimizes efficiency through beta-suppression and frontoparietal integration. These frequency-specific mechanisms provide preliminary neurophysiological insights that may inform future development of personalized rehabilitation approaches for motor disorders. Future investigations should validate these findings in clinical populations with motor impairments.