The Role of Sleep in Episodic Memory Formation

Several lines of evidence indicate, that sleep essentially participates in memory consolidation, a process that converts newly acquired and initially fragile memory traces into a long-lasting and enhanced form. Whereas the enhancing effect of sleep on procedural memory for skills has been well established by a large number of human studies, evidence for a similar improving influence of sleep on hippocampus-dependent episodic memory mainly derives from spatial learning in rodents. Hippocampal neuronal ensembles modified by spatial experience during waking have been shown to be sponaneously reactivated during subsequent sleep. During this `offline` reprocessing the newly encoded memory traces are supposed to be analyzed and gradually translated from short-term hippocampal to long-lasting knowledge structures within the neocortex, thereby becoming independent from the hippocampus. In humans, episodic and spatial learning likewise critically depend on the integrity of the hippocampus and adjacent medial temporal lobe structures. Thus, in the present research project episodic memory will be assessed using a topographical navigation task, that requires subjects to learn the structure of a complex virtual town and to find their way to different target locations. In experiment 1 we will investigate (i) whether sleep improves performance on tests of route retrieval and, (ii) using fMRI, whether sleep-dependent episodic memory consolidation is accompanied by significantly reduced retrieval-related activation in the medial temporal lobe system (with reference to a corresponding interval of wakefulness, respectively). In experiment 2 we will examine whether sleep has a similar selective function in the consolidation of episodic memories as we have most recently found with respect to the procedural motor memory domain. Independent from any differences during the stages of encoding and retrieval testing, sleep after learning two different finger sequences has been shown to selectively improve performance on that one, that directly before sleep had been linked with a monetary reward. Here, we hypothesize, (iii) that sleep selectively enhances memory for target locations that were prospectively linked with a monetary reward (with reference to those that remained neutral) and (iv) that a corresponding interval of wakefulness will exert no such differential motivation-dependent influence. Finally, in experiment 3 we aim to assess the differential contributions of slow wave sleep (SWS) and rapid eye movement (REM-) sleep to the selective enhancement of motivated vs. non-motivated episodic memory formation. We would expect (v) that SWS primarily weakens non-motivated (vs. motivated) spatial memories while (vi) REM-sleep might improve topographical knowledge for rewarded (vs. non-rewarded) target locations.

Several lines of evidence indicate, that sleep essentially participates in memory consolidation, a process that converts newly acquired and initially fragile memory traces into a long-lasting and enhanced form. During this `offline` reprocessing the newly encoded memory traces are supposed to be analyzed and gradually translated from short-term hippocampal to long-lasting knowledge structures within the neocortex, thereby becoming independent from the hippocampus. It is thought that these processes typically last for month and years. The aim of the present research project was to investigate the role of sleep for declarative memory consolidation in healthy human subjects. In experiment 1 we investigated whether an 8-hours retention interval of sleep as compared with a corresponding interval of wakefulness improves declarative memory performance and, using functional magnetic resonance imaging (fMRI), whether sleep-dependent episodic memory consolidation is already accompanied by significantly reduced retrieval-related activation in the hippocampus. For this reason, subjects were trained on a list of word pairs prior to the respective retention intervals. Retrieval testing after the retention interval took place in the fMRI-scanner. The results show that subjects in the post-learning sleep as compared with the post-learning wake condition had less neuronal activation within the medial temporal lobe (including the hippocampus) and more neuronal activation in neocortical brain areas (including the anterior cingulum and superior temporal lobe). In addition, in the post-learning sleep condition, there was a negative correlation between brain activation in the medial temporal lobe and the neocortical areas, i.e. the less retrieval-related activation in the hippocampus, the more retrieval-related activation in the neocortical brain regions. With respect to the behavioural data, post- learning sleep had no beneficial effect as compared with post-learning wakefulness. In supplementary experiments however, sleep`s behavioural significance impressively became obvious when memory retrieval was challenged with an interfering list of word pairs directly before delayed recall testing took place. Thus, sleep appears to participate in the reorganization of declarative memory traces. This reorganization may protect the initially vulnerable memory traces against behavioural interference. In experiment 2 and 3 we aimed to examine whether sleep has a similar selective function in the consolidation of episodic memories as we have most recently found with respect to the procedural motor memory domain. Independent from any differences during the stages of encoding and retrieval testing, sleep after learning two different finger sequences has been shown to selectively improve performance on that one, that directly before sleep had been linked with a monetary reward. Here, we could show that an 8-hours retention interval of sleep selectively improves declarative memories that were associated with anticipated reward. A corresponding interval of wakefulness did not exert such differential motivation-dependent influences. In addition, we found that it is not early sleep, dominated by slow wave sleep (SWS), but late sleep during which rapid eye movement (REM)-sleep prevails that appears to mediate these differential effects.