Consolidation effect of repeated processing of declarative knowledge in mental experiences during human sleep.

Sleep may positively influence declarative memory through the processing, which transforms items of declarative knowledge into contents of mental sleep experience (MSE). A prediction from this general hypothesis is that the consolidation level should be higher for the output of items repeatedly processed and transformed into identical or very similar (so-called interrelated) contents of distinct MSEs of the same night rather than for the output of items presumably processed once (that is, all other, non-interrelated contents). Two experiments examined whether and how far the frequency and long-term retention of interrelated contents depend on the repeated processing of given items rather than on the experimental procedure applied for detection of interrelated contents. This procedure entails both multiple awakenings and a verbal report of MSE after awakening. Multiple awakenings could facilitate the re-access and elaboration of some contents into the subsequent (i.e. contiguous) MSE rather than non-contiguous MSEs; verbal reports could enhance the delayed recall of interrelated contents in as much as repeatedly encoded. The first experiment showed that interrelated contents were more frequent and better retained than both non-interrelated and pseudo-interrelated (i.e. by-chance similar or identical) contents, and even more in pairs of contiguous than non-contiguous MSEs collected from the first four periods of REM sleep on each experimental night. The second experiment showed that the frequency and retention rate of interrelated contents, while higher than those of non-interrelated and pseudo-interrelated contents, were not significantly different in pairs of MSEs which were verbally or mentally recalled after awakening provoked during the first four periods of REM sleep in each experimental night. Taken together, these findings indicate that the advantage provided by repeated processing during REM sleep for the consolidation of the output of items of declarative knowledge is conspicuous and largely independent from the experimental procedure, as this only slightly enhances the frequency and retention rate of interrelated contents.