Modulating Sleep Oscillations Through Targeted Memory Reactivation: Linking Functional Dynamics to Brain Microstructure

Background: Sleep-dependent memory consolidation reputedly rely on the coordinated activity of sleep oscillations, which may facilitate the reactivation and redistribution of newly acquired memories. Targeted memory reactivation (TMR) during sleep can experimentally modulate these processes, although effects remain variable. Moreover, how learning-related microstructural plasticity relates to sleep oscillatory activity and spindle-related network dynamics remains poorly understood. We investigated whether auditory TMR modulates sleep spindles, slow oscillations (SOs) and their co-occurrence, whether these changes relate to motor memory consolidation and diffusion MRI (dMRI) measures of brain microstructure, and whether learning and TMR alter the evolution of cortico-cortical connectivity during spindle development. Methods: Sixty healthy young adults underwent five dMRI sessions before and after motor sequence learning (Day 1), after a post-training night of regular sleep (RS) or TMR (Day 2), and before and after relearning (Day 5). Baseline and experimental nights were recorded with 256-channel high-density EEG. Oscillation densities were analyzed using mixed-design ANOVAs and related to behavioral and dMRI changes using Pearson or Spearman correlations. Intra-spindle connectivity between 148 source-reconstructed cortical regions was compared between spindle onset and end, across nights and between groups using Network-Based Statistics. Results: Spindle density, SO density and SO-spindle co-occurrence remained stable across nights in the RS group but decreased after TMR. Higher spindle density and co-occurrence were associated with greater overnight performance gains, and all oscillatory measures were associated with learning-related dMRI changes in subcortical regions, notably striatum and thalamus. Connectivity increased from the beginning to the end of the spindle duration during the baseline night in both groups, but during the experimental night only in the RS group. Conclusions: Motor learning is associated with coordinated changes in sleep oscillatory activity, within-spindle cortical connectivity and brain microstructure, supporting a dynamic interaction between learning-induced plasticity and sleep-dependent network reorganization. Under the present stimulation conditions, TMR did not enhance these processes but instead reduced sleep oscillatory activity and attenuated spindle-related connectivity dynamics, underscoring the importance of precisely targeting endogenous sleep rhythms when using TMR.