Neurochemical mechanisms underlying sleep-dependent memory processing

Abstract

Sleep is beneficial for memory consolidation and enhances subsequent learning. It has been hypothesized that the reactivation of memories occurring during sleep, especially slow-wave sleep (SWS), would serve both of these memory processes. Although the benefit of sleep on learning and consolidation has repeatedly been demonstrated, the underlying neurochemical mechanisms supporting these functions of sleep are not yet fully understood. My thesis is centred on manipulating sleep-dependent memory consolidation by using pharmacological agents to unravel the neurochemical mechanisms that convert neural reactivation into plastic changes. My current work is focused on two major memory-related neurotransmitters, i.e., glutamate in the first study, and dopamine in the second study. Glutamate is the main excitatory neurotransmitter in the human brain. This neurotransmitter is involved in different forms of plasticity. In the hippocampus, long-term potentiation (LTP), which is an important factor for synaptic consolidation is mediated, by NMDA receptors, one type of glutamate receptor, containing the NR2A subunits, and long-term depression (LTD) is mediated by NMDA receptors containing NR2B subunits. D-cycloserine (DCS) as NMDA-receptor co-agonist preferentially acts through NR2A containing receptors, which may favour LTP over LTD. Sleep-dependent declarative memory consolidation, has been shown to be facilitated by DCS administration. In the first study, we surveyed whether the administration of DCS during sleep impairs new declarative learning of a similar task after sleep due to the assumed enhanced potentiation of memory traces and corresponding increased proactive interference under DCS during sleep. This potentiation might reduce the capacity for new encoding in the hippocampus. Presumably, this reduction in subsequent new learning under DCS will be enhanced by learning new overlapping information (interference condition). Our results using DCS showed the predicted improvement in new learning after sleep compared to wakefulness. Unexpectedly, however, interference did not impair new learning but rather further improved it. This might be because rather than an interference effect processes of schema generation and knowledge abstraction may have occurred. These effects seem to switch in a time-dependent manner. Furthermore, DCS did not impair, but rather improved new learning, and this was sleep-independent. Therefore, I speculate that forgetting, shown to rely on NMDAR-activation to erase old memories, boosts new learning through glutamatergic processes that occurs independent of sleep. Dopamine, as a major modulatory neurotransmitter, facilitates plasticity for reward associated memories in the hippocampus and other reward-related midbrain structures during encoding. Sleep selectively benefits the strengthening and transformation of highly relevant memories like rewarded memories by selective replay. However, it is not completely clear whether or not these sleep-dependent consolidation processes also engage the dopaminergic circuitry, which facilitated their initial encoding. I conducted a neuropharmacological experiment in humans using sulpiride, a D2-like receptor antagonist, to block the effect of dopaminergic afferents in the hippocampus and thereby reduce plasticity. Our results showed that highly rewarded memories were remembered better than lowly rewarded ones under both sulpiride and placebo conditions. This means blocking dopaminergic neurotransmission did not impact the selectivity of sleep for consolidation. This finding indicates a less important role of dopaminergic pathways for the preferential consolidation of highly-rewarded memories during sleep compared to their role at encoding during wakefulness. We also found that better performance on highly rewarded items is linked to the time spent in sleep stage 4, which lends support to the idea that rewards increase replay activity during sleep to enhance relevant memories selectively. Altogether, these two studies show that pharmacological manipulations can improve our current knowledge about the neurochemical mechanisms, which underlie sleep-dependent memory processes. Importantly, these direct manipulations in humans enabled us to investigate the complexity of human behaviour in response to neurochemical manipulation and allowed us to interpret these findings without translating them from animal models. This may allow developing new therapeutic applications for patients suffering from cognitive disorders with more confidence.