In our view, spindle activity enables local, undisturbed cortical reprocessing of previously reactivated memories, while the reactivation itself is associated with theta activity. Thus, a successful reinstatement of memory representation (as indicated by transient increases in theta activity) together with subsequent occurrence of oscillatory spindle activity might be necessary to induce plastic changes underlying the stabilization of memories upon their reactivation
during sleep. Disturbing these combined oscillatory signals (e.g., by presenting auditory stimuli immediately after
successful reactivation) disrupts the reinstatement (theta) and stabilization (spindle) process and blocks the beneficial effect of cueing during sleep (see Fig. 1E–G). Our working model would also predict that successful memory reactivation during REM sleep could result in increases in theta activity (i.e. successful memory reinstatement), but does not lead to a stabilization of the memory trace due to missing spindle oscillations. Conversely (if at all possible),
inducing spindle activation immediately after reactivation during REM sleep should result in memory benefits of cueing.