Logo Logo
Switch language to English
Schönauer, Monika (2014): Sleep-dependent consolidation in multiple memory systems. Dissertation, LMU München: Graduate School of Systemic Neurosciences (GSN)



Before newly formed memories can last for the long-term, they must undergo a period of consolidation. It has been shown that sleep facilitates this process. One hypothesis about how this may occur is that learning-related neuronal activity is replayed during following sleep periods. Such a reactivation of neural activity patterns has been repeatedly shown in the hippocampal formation in animals. Hippocampally-induced reactivation can also be observed in other brain areas like the neocortex and basal ganglia. On the behavioral level, sleep has been found to benefit performance on a broad range of memory tasks that rely on different neural systems. Up to now, however, it is unclear whether the same mechanisms mediate effects of sleep on consolidation in different memory systems. In this thesis, we investigated both the effects and the mechanisms of sleep-dependent consolidation in multiple memory systems. We find that sleep benefits performance on a broad range of procedural and declarative memory tasks (studies 1 and 2). These beneficial effects of sleep go beyond a reduction of retroactive interference as effected by quiet wakeful meditation (study 1). In study 2, we demonstrate that the processes underlying these beneficial effects of sleep are different for different memory systems. We assessed performance on typical declarative and procedural memory tasks during one week after participants slept or were sleep deprived for one night after learning. Sleep-dependent consolidation of hippocampal and non-hippocampal memory follows different time-courses. Hippocampal memory shows a benefit of sleep only one day after learning. Performance after sleep deprivation recovers following the next night of sleep, so that no enduring effect of sleep can be observed. However, sleep deprivation before recall does not impair performance. For non-hippocampal memory, on the other hand, long-term benefits of sleep after learning can be observed even after four days. Here, delayed sleep cannot rescue performance. This indicates a dissociation between two sleep-related consolidation mechanisms, which rely on distinct neuronal processes. We studied the neuronal processes underlying sleep effects on declarative memory in study 3, where we investigate learning-related electrophysiological activity in the sleeping brain. With the help of multivariate pattern classification algorithms, we show that brain activity during sleep contains information about the kind of visual stimuli that were learned earlier. We thus find that learned material is actively reprocessed during sleep. In a next step, we examined whether procedural memory can also benefit from reactivation during sleep. We find that a procedural memory task that has been found to activate the hippocampus can be strengthened by externally cueing the reactivation process during sleep. Similar to study 2, this indicates that it is not the traditional distinction between declarative and procedural memory that determines how memories are consolidated during sleep. Rather, memory systems, and in particular hippocampal contribution, decide the sleep-dependent consolidation process. In the first four studies, we examined how sleep affects memory in different memory systems. In our last study, we went one step further and investigated whether multiple memory systems can also interact during consolidation in sleep. We devised a task during which both implicit and explicit memory develop during learning. Results show that sleep not only strengthens implicit and explicit memory individually, it also integrates these formerly separate representations of the learning task. Implicit and explicit memory are negatively correlated immediately after training. Sleep renders this association positive and allows cooperation between the two memory traces. We observe this change both in behavior, using structural equation modeling, and on the level of brain activity, measured by fMRI. After sleep, the hippocampus is more strongly activated during recall of implicit memory, whereas the caudate nucleus shows stronger activity during explicit memory recall. Moreover, both regions show correlated stimulus-induced responses in a task that allows memory systems cooperation. These results provide conclusive evidence that sleep not only strengthens memory, but also reorganizes the contributing neural circuits. In this way, sleep actually changes the quality of the memory representation.