Memory updating has been established; however, the mechanism supporting this alteration process is subject to disagreement. Reconsolidation theorists argue that memory updating occurs via an old memory becoming reactivated and returned to a state of susceptibility. In this state, newly presented details can become incorporated into the existing memory. As such, memory updating is an effect of old memory reactivation and new information encoding. In contrast, temporal context theory argues that the temporal context in which the old memory was initially formed is reinstated. Newly presented information becomes tagged to the old context. Therefore, at retrieval, when the old context is reinstated again, the initially bound information and the newer information are simultaneously retrieved. Within this theoretical framework, memory modification is the result of retrieval effects. In contrast, this three-paper dissertation provides evidence that reconsolidation is, at least in part, a combined reactivation and encoding effect. In paper 1, I present neural evidence of both old memory reactivation and new encoding, which demonstrates 1) that strength at reactivation predicts the likelihood that a memory will be modified and 2) that greater brain activation during new encoding predicts the extent of accurate recognition. In paper 2, I show that encoding conditions affect the extent to which new information will be misattributed to the old memory. I demonstrate that learners update explicitly encoded memories but not implicitly coded ones. Lastly, in paper 3, I demonstrate that old memories can be reactivated and altered during sleep. When old-memory reactivation is paired with a forget cue, a subsequent degeneration of the memory and its details ensues. In sum, all three papers provide evidence in support of the reconsolidation theory that memory updating occurs during old-memory reactivation and new encoding.

Emotion is thought to modulate the long-term fate of memories. Experiences that elicit an emotional response tend to be better remembered than comparatively unemotional events, while the emotional charge associated with these memories diminishes over time. Sleep – in particular rapid-eye movement (REM) and slow-wave sleep (SWS) - has been implicated in both the selective strengthening and affective uncharging of emotional memories. According to the sleep to forget, sleep to remember (SFSR) hypothesis, both processes occur in parallel during REM sleep. Although evidence strongly supports a role of REM sleep in the selective consolidation of emotional memories, it is far less clear to what extent sleep is involved in the development of emotional charge. While some studies support a primary role of REM in habituation (i.e. the decrease of emotional charge), others suggest a more central role of SWS. Further, existing literature indicates that the physiological (bottom-up) and cognitive (top-down) components of emotional responses may be differentially processed across sleep. Chapter 2 proposes complementary functions of REM and SWS in emotional memory processes based on a combination of evidence from rodent and human research. The experiments presented in this thesis employed polysomnography (PSG), subjective arousal testing, pupillometry, targeted memory reactivation (TMR), and transcranial alternating current stimulation (tACs) to investigate the respective roles of REM and SWS in the overnight development of subjective and autonomic arousal in response to negative emotional and neutral stimuli. In Chapter 3 I assessed how subjective and autonomic responses to neutral and negative stimuli develop across 12 hours containing either nocturnal sleep or daytime wakefulness. I found that autonomic reactivity – indexed by pupil dilation – decreased across sleep but not wake, while subjective arousal did not change across either interval. In a further experiment, I investigated whether the placement of sleep within a 24 hour interval would affect habituation. Once again, autonomic arousal decreased significantly. Subjective arousal towards negative stimuli was found to decrease more if sleep followed rather than preceded daytime wakefulness within the 24 hour interval. In Chapter 4 I explored the role of REM sleep in emotional habituation by applying 5 Hz tACs in an attempt to entrain endogenous cortical theta (4-7 Hz) activity, which has previously been associated with emotional memory consolidation in humans. Surprisingly, I found that stimulation was associated with a reduction in theta power and no change in subjective or autonomic habituation compared to the sham control night. In Chapters 5 and 6, I addressed the contribution of memory reactivations during SWS and REM sleep, respectively, in emotional habituation using TMR. In Chapter 5, I found that TMR was associated with a decrease and simultaneous increase in autonomic habituation towards negative and neutral stimuli, respectively, without affecting overnight changes in subjective arousal. In contrast, in Chapter 6, TMR during REM sleep was associated with an increase in subjective habituation towards both neutral and negative stimuli without affecting autonomic responses. In conclusion, my results provide new insights to the role of sleep in emotional habituation. I have provided evidence that targeted memory reactivation during REM sleep can modulate the development of cognitive evaluations of emotion, while TMR during SWS may interfere with autonomic habituation. This suggests distinct emotional processing during REM and SWS, as well as a dissociation between subjective and autonomic habituation across sleep. These results are discussed in the light of previous research and the model of sleep-dependent emotional memory processing proposed in Chapter 2.

The method targeted memory reactivation (TMR) uses specific stimulation when subjects are completing tasks and during sleep. The TMR process is known to influence the consolidation of declarative memories. The aim of this thesis is to conduct a systematic review on the effects of TMR on declarative memory consolidation during slow-wave sleep (SWS). The research question is to answer what effect TMR during SWS has on the consolidation of declarative memory in healthy humans when presented with associated cues of the targeted learning experiences. Eighteen studies were included in this review. Four studies found a significant effect of TMR on declarative memory consolidation, and 10 found a non-significant effect. In four studies the effect of TMR depended on different inclusions, analyses, and factors, for example between slow oscillation up-and down-states and between participants that vary in pre-sleep performance in the examined task. In contrast to previous findings, this review does not provide evidence for the effect of TMR on declarative memories during SWS. More research analysing different factors, such as different cues, age of participants, duration of SWS, and specific experimental tasks, needs to be done in the fields of TMR and auditory cues.

Why are some memories easy to retrieve, while others are more difficult to access? Here, we tested whether we could bias memory replay, a process whereby newly learned information is reinforced by reinstating the neuronal patterns of activation that were present during learning, towards particular memory traces. The goal of this biasing is to strengthen some memory traces, making them more easily retrieved. To test this, participants were scanned during interleaved periods of encoding and rest. Throughout the encoding runs, participants learned triplets of images that were paired with semantically related sound cues. During two of the three rest periods, novel, irrelevant sounds were played. During one critical rest period, however, the sound cues learned in the preceding encoding period were played in an effort to preferentially increase reactivation of the associated visual images, a manipulation known as targeted memory reactivation. Representational similarity analyses were used to compare multi-voxel patterns of hippocampal activation across encoding and rest periods. Our index of reactivation was selectively enhanced for memory traces that were targeted for preferential reactivation during offline rest, both compared to information that was not targeted for preferential reactivation and compared to a baseline rest period. Importantly, this neural effect of targeted reactivation was related to the difference in delayed order memory for information that was cued versus uncued, suggesting that preferential replay may be a mechanism by which specific memory traces can be selectively strengthened for enhanced subsequent memory retrieval. We also found partial evidence of discrimination of unique temporal sequences within the hippocampus. Over time, multi-voxel patterns associated with a given triplet sequence became more dissimilar to the patterns associated with the other sequences. Furthermore, this neural marker of sequence preservation was correlated with the difference in delayed order memory for cued versus uncued triplets, signifying that the ability to reactivate particular temporal sequences within the hippocampus may be related to enhanced temporal order memory for the cued information. Taken together, these findings support the claim that awake replay can be biased towards preferential reactivation of particular memory traces and also suggest that this preferential reactivation, as well as representations of reactivated temporal sequences, can be detected within patterns of hippocampal activation. / Psychology

Il a été suggéré que lorsqu’une trace de mémoire consolidée est rappelée (réactivée), elle devient instable et sujette aux modifications avant de se stabiliser à nouveau en mémoire à long terme. Nous avons récemment démontré que lorsque la réactivation d’un souvenir négatif est couplée à l’exposition à un stress psychosocial, le souvenir de l’évènement négatif est augmenté de façon durable. En se basant sur ces résultats, le but de cette thèse est de préciser le rôle du stress psychologique et physiologique (hormones de stress) sur la modulation de souvenirs réactivés. Plus précisément, la première étude visait à déterminer si le cortisol, hormone de stress majeure, est un joueur clé dans la modulation des souvenirs réactivés. Pour ce faire, nous avons inhibé pharmacologiquement les niveaux de cortisol au moment de la réactivation d’un souvenir contenant des segments neutres et négatifs. Les résultats démontrent que la réactivation du matériel négatif est amoindrie lorsque les niveaux de cortisol sont inhibés, et cet effet est toujours présent quatre jours plus tard. Étant donné que les stimuli utilisés jusqu’à maintenant ont une faible validité écologique, nous avons voulu déterminer si d’autres types de mémoires pouvaient également être modulables lors de leur réactivation. L’objectif de la deuxième étude était donc de déterminer si les mémoires autobiographiques collectives sont modulables par le stress au moment de leur réactivation. Pour ce faire, nous avons exposé les participants à de vrais extraits de journaux, neutres ou négatifs, afin de réactiver les mémoires collectives associées à ces évènements. Par la suite, tous les participants ont été exposés à un stress psychosocial et leur mémoire des extraits a été évaluée la journée suivante. Les résultats démontrent que les femmes ayant lu les nouvelles négatives avaient une réactivité physiologique accrue face au stresseur et une mémoire augmentée de ces mêmes nouvelles le jour suivant. Ce phénomène n’était cependant pas observable chez les hommes. Le but de la troisième étude était de déterminer si les mémoires autobiographiques personnelles sont modulables par le stress au moment de leur réactivation. Nous avons demandé aux participants de se remémorer deux évènements de leur passé, négatifs ou neutres. Par la suite, ils ont été exposés à un stress psychosocial et leur mémoire pour ces mêmes évènements a été évaluée à nouveau la journée suivante. Les résultats démontrent que les mémoires autobiographiques personnelles réactivées ne semblent pas être modulables par l’exposition à un stresseur. Globalement, les résultats de cette thèse démontrent que le cortisol a la capacité de moduler des souvenirs négatifs réactivés, mais que la nature (extrinsèque vs. intrinsèque) et l'intensité des souvenirs réactivés sont des facteurs déterminants pour que ce phénomène prenne place. / It has been suggested that when a consolidated memory trace is recalled (reactivated), it becomes active and sensitive to modifications before stabilizing again in the long-term memory system. We have recently demonstrated that when the reactivation of a negative memory is followed by exposition to a psychosocial stressor, the memory for the negative material is enhanced in a long-lasting manner. Based on these results, the goal of this thesis is to clarify the role of physiological (stress hormones) and psychological stress on the modulation of reactivated memories. More precisely, the first study aimed to determine whether cortisol, a major stress hormone, is a key player in the modulation of reactivated memories. To do so, we have pharmacologically inhibited cortisol levels at the time of reactivating a memory composed of neutral and negative segments. Results showed that the reactivation of the negative material is decreased when cortisol levels are inhibited and this effect is still present four days later. Given that the stimuli used so far have a weak ecological validity, we wanted to determine whether other types of memories could also be modified upon their reactivation. The goal of the second study was to examine whether collective autobiographical memories were sensitive to the effects of stress at the time of reactivation. To do so, we have exposed participants to real newspaper excerpts, either neutral or negative, in order to reactivate the collective memories associated to these events. We have then exposed all the participants to a psychosocial stressor and their memory for the excerpts was assessed the following day. Results showed that women who have read negative news excerpts had a more pronounced physiological reactivity to the stressor and an increased memory for these news the following day. This phenomenon, however, was not observed in men. The goal of the third study was to determine whether personal autobiographical memories were sensitive to the effects of stress at the time of their reactivation. We have asked participants to recall two personal events, either negative or neutral. We then exposed them to a stressor and their memory for these events was tested the following day. The results showed that reactivated personal autobiographical memories were not sensitive to stress. Globally, the results of this thesis show that cortisol has the capacity to modulate reactivated negative memories, but that the nature (extrinsic vs. intrinsic) and the intensity of the memories are determining factors for this phenomenon to occur.

Il a été suggéré que lorsqu’une trace de mémoire consolidée est rappelée (réactivée), elle devient instable et sujette aux modifications avant de se stabiliser à nouveau en mémoire à long terme. Nous avons récemment démontré que lorsque la réactivation d’un souvenir négatif est couplée à l’exposition à un stress psychosocial, le souvenir de l’évènement négatif est augmenté de façon durable. En se basant sur ces résultats, le but de cette thèse est de préciser le rôle du stress psychologique et physiologique (hormones de stress) sur la modulation de souvenirs réactivés. Plus précisément, la première étude visait à déterminer si le cortisol, hormone de stress majeure, est un joueur clé dans la modulation des souvenirs réactivés. Pour ce faire, nous avons inhibé pharmacologiquement les niveaux de cortisol au moment de la réactivation d’un souvenir contenant des segments neutres et négatifs. Les résultats démontrent que la réactivation du matériel négatif est amoindrie lorsque les niveaux de cortisol sont inhibés, et cet effet est toujours présent quatre jours plus tard. Étant donné que les stimuli utilisés jusqu’à maintenant ont une faible validité écologique, nous avons voulu déterminer si d’autres types de mémoires pouvaient également être modulables lors de leur réactivation. L’objectif de la deuxième étude était donc de déterminer si les mémoires autobiographiques collectives sont modulables par le stress au moment de leur réactivation. Pour ce faire, nous avons exposé les participants à de vrais extraits de journaux, neutres ou négatifs, afin de réactiver les mémoires collectives associées à ces évènements. Par la suite, tous les participants ont été exposés à un stress psychosocial et leur mémoire des extraits a été évaluée la journée suivante. Les résultats démontrent que les femmes ayant lu les nouvelles négatives avaient une réactivité physiologique accrue face au stresseur et une mémoire augmentée de ces mêmes nouvelles le jour suivant. Ce phénomène n’était cependant pas observable chez les hommes. Le but de la troisième étude était de déterminer si les mémoires autobiographiques personnelles sont modulables par le stress au moment de leur réactivation. Nous avons demandé aux participants de se remémorer deux évènements de leur passé, négatifs ou neutres. Par la suite, ils ont été exposés à un stress psychosocial et leur mémoire pour ces mêmes évènements a été évaluée à nouveau la journée suivante. Les résultats démontrent que les mémoires autobiographiques personnelles réactivées ne semblent pas être modulables par l’exposition à un stresseur. Globalement, les résultats de cette thèse démontrent que le cortisol a la capacité de moduler des souvenirs négatifs réactivés, mais que la nature (extrinsèque vs. intrinsèque) et l'intensité des souvenirs réactivés sont des facteurs déterminants pour que ce phénomène prenne place. / It has been suggested that when a consolidated memory trace is recalled (reactivated), it becomes active and sensitive to modifications before stabilizing again in the long-term memory system. We have recently demonstrated that when the reactivation of a negative memory is followed by exposition to a psychosocial stressor, the memory for the negative material is enhanced in a long-lasting manner. Based on these results, the goal of this thesis is to clarify the role of physiological (stress hormones) and psychological stress on the modulation of reactivated memories. More precisely, the first study aimed to determine whether cortisol, a major stress hormone, is a key player in the modulation of reactivated memories. To do so, we have pharmacologically inhibited cortisol levels at the time of reactivating a memory composed of neutral and negative segments. Results showed that the reactivation of the negative material is decreased when cortisol levels are inhibited and this effect is still present four days later. Given that the stimuli used so far have a weak ecological validity, we wanted to determine whether other types of memories could also be modified upon their reactivation. The goal of the second study was to examine whether collective autobiographical memories were sensitive to the effects of stress at the time of reactivation. To do so, we have exposed participants to real newspaper excerpts, either neutral or negative, in order to reactivate the collective memories associated to these events. We have then exposed all the participants to a psychosocial stressor and their memory for the excerpts was assessed the following day. Results showed that women who have read negative news excerpts had a more pronounced physiological reactivity to the stressor and an increased memory for these news the following day. This phenomenon, however, was not observed in men. The goal of the third study was to determine whether personal autobiographical memories were sensitive to the effects of stress at the time of their reactivation. We have asked participants to recall two personal events, either negative or neutral. We then exposed them to a stressor and their memory for these events was tested the following day. The results showed that reactivated personal autobiographical memories were not sensitive to stress. Globally, the results of this thesis show that cortisol has the capacity to modulate reactivated negative memories, but that the nature (extrinsic vs. intrinsic) and the intensity of the memories are determining factors for this phenomenon to occur.

This doctoral thesis aimed at better understanding the contribution of sleep, Targeted Memory Reactivation and transcranial Direct Current Stimulation (tDCS) on the consolidation of neutral and emotional memories. In the first part of this work, we found that REM-enriched naps and more specifically rapid eye movement density is associated with the consolidation of sad stories, suggesting a possible implication of Ponto-Geniculo-Occipital (PGO) waves in the consolidation of sad information. In addition, we observed an increase in emotional reactivity during re-exposure to the sad story following a REM-enriched nap. We postulate that REM sleep favored the consolidation of the emotionalsalience of the sad memories, leading to exacerbated emotional reactivity during re-exposure. We also investigated the impact of TMR during NREM sleep on the consolidation of neutral and negative word pairs leanred with a specific sound. We found an equal benefit of the TMR procedure on neutral and emotional material, suggesting that emotion does not modulate the selective enhancing effect of TMR during NREM sleep. In an additional study, we tested the impact of verbatim presentation of the pairs of words during NREM sleep and did not find the memory benefits of TMR. We ascribed the absence of TMR memory benefit to the detrimental effect of the auditory presntation of the second word which impaired the memory reactivation processes initiated by the presentation of the first word. Together, theseresults indicate the crucial role of a sensitive plastic time window necessary for the successful processes of memory reactivation during sleep. Finally, we evaluated how the combination of tDCS and TMR procedure during a wakeful rest consolidation interval benefits memory consolidation. We found that TMR alone led to selective memory benefits for cued word pairs. When the TMR procedure was combined with either right-anodal or left anodal tDCS, we observed a significant improved global learning, suggesting that tDCS does not potentiate but overshadows the TMRprocedure. Altogether, these studies offer new perspectives in the field of memory consolidation. More specifically, the application of an alternating current during post-learning sleep concomitantly to a TMR procedure might favor the specific brain oscillations involved in successful memory reactivation, and might enhance the associated memory gains. / Doctorat en Sciences psychologiques et de l'éducation / info:eu-repo/semantics/nonPublished

Le sommeil joue un rôle important dans la consolidation de la mémoire. Les expériences nouvellement acquises à l’éveil sont réactivées spontanément durant le sommeil, un processus qui aiderait à consolider et intégrer la mémoire à plus long terme. Il a été suggéré que ces réactivations de mémoire se reflétaient, du moins partiellement, dans le contenu des rêves et que les rêves pouvaient jouer un rôle actif dans la consolidation de la mémoire. L'objectif général de cette thèse de doctorat est ainsi d'évaluer si et comment le rêve est impliqué dans le traitement de la mémoire épisodique et procédurale en utilisant de nouvelles technologies et approches expérimentales pour étudier ces relations. La première étude de cette thèse visait à influencer les réactivations de mémoire durant le sommeil afin de clarifier leurs liens avec les rêves et la mémoire procédurale. Nous avons stimulé ces réactivations de mémoire en réexposant des participants, durant leur sommeil, à un stimulus sonore préalablement associé à un apprentissage moteur, une approche nommée « targeted memory reactivation » (TMR). Nous montrons que la TMR, lorsqu’appliquée en stade de sommeil paradoxal, permet d’améliorer l’apprentissage d’une tâche motrice complexe, soit apprendre à « voler » en réalité virtuelle. De plus, le fait de rêver à des éléments kinesthésiques de la tâche motrice en sommeil paradoxal, mais pas en sommeil lent léger, est associé à une meilleure amélioration à cette tâche (article I). Ces résultats appuient les modèles suggérant que le sommeil paradoxal joue un rôle important dans la consolidation de la mémoire procédurale complexe et suggèrent en outre que la simulation de sensations sensorimotrices dans les rêves pourrait contribuer à ce rôle. Bien que la TMR n’ait pas eu d’impact direct sur les rêves, nous montrons qu’elle peut influencer le décours temporel des incorporations de mémoire dans les rêves sur plusieurs jours. La TMR a amplifié les incorporations tardives de la tâche, soit 1-2 jours plus tard lorsqu’appliquée en sommeil paradoxal, et 5-6 jours plus tard lorsqu’appliquée en sommeil lent profond (article II). Nous suggérons que ces effets à plus long terme pourraient être dus à un mécanisme de marquage (tagging) des traces de mémoire, initiant ou amplifiant leur traitement subséquent au cours de plusieurs nuits de sommeil. De plus, nous montrons que notre expérience immersive en réalité virtuelle a augmenté l’incorporation de sensations de vol dans les rêves, particulièrement la nuit suivant l’exposition à celle-ci (article III). Nous identifions certains facteurs individuels qui sont associés à une plus grande incorporation de la tâche dans les rêves, tels que le fait d’avoir déjà eu des rêves de vols ou des rêves lucides. Un examen plus approfondi des rêves pendant 10 jours suivants l'expérience de réalité virtuelle montre que les sensations de vol sont progressivement décontextualisées du contexte de vol original au fil du temps. Les rêves de vol après l’expérience en réalité virtuelle avaient également des niveaux de contrôle et d'intensité émotionnelle plus élevés que ceux ayant eu lieu avant l’expérience. L'induction de rêves de vol nous a permis de faire une analyse qualitative approfondie sur ceux-ci, menant à une nouvelle proposition de la manière dont les sensations de vol, ou les sensations de mouvement de manière plus générale, peuvent survenir dans les rêves. Ces résultats pourraient à leur tour faciliter le développement de technologies pour mieux influencer et étudier les rêves. Une deuxième étude visait à évaluer quand et comment une source de mémoire épisodique commune à tous les participant – visiter le laboratoire de sommeil – est incorporé dans les rêves. Les résultats montrent que près du tiers des rêves incorporent des éléments du laboratoire, et ce, particulièrement dans les rêves en sommeil paradoxal lors d’une sieste matinale (article IV). Une étude qualitative de ces rêves met de l’avant les manières typiques par lesquelles des éléments du laboratoire réapparaissent dans les narratifs de rêve. Nous proposons l’existence de différentes « pressions intégratives » qui structurent les fragments de mémoire au sein de scénarios de rêves. Souvent, ces scénarios impliquent une certaine pression de performance, sont de nature sociale, projettent le rêveur dans le temps et dans l’espace, ou incorporent des sensations réelles du corps et de l’environnement de sommeil. Étudier le phénomène de « rêver au laboratoire » aide ainsi à clarifier comment les rêves sont façonnés autour de fragments de mémoire et souligne à la fois les avantages et les limites méthodologiques de l’étude des rêves et de la mémoire en laboratoire. Finalement, une troisième étude visait à suivre le décours temporel des sources de mémoire autobiographique des rêves au cours d’une nuit de sommeil entière en utilisant un protocole de réveils en série. Nos résultats montrent que les rêves peuvent combiner plusieurs sources de mémoire, en particulier lorsqu'ils se produisent en stade N1 ou en sommeil paradoxal, ce qui pourrait refléter une plus grande richesse ou capacité intégrative des rêves ayant lieu en ces stades (article V). Nous reproduisons le résultat voulant que les souvenirs récents sont préférentiellement réactivés tôt dans la nuit, tandis que les souvenirs plus lointains sont relativement plus représentés dans les rêves de fin de nuit – nous précisons que cet effet est indépendant des stades de sommeil. La co-activation de plusieurs sources de mémoire dans un même récit de rêve appuie la suggestion que l’une des fonctions du sommeil est d'intégrer les nouvelles connaissances à nos réseaux de mémoire préexistants. Nos résultats suggèrent que cette fonction pourrait être un processus cumulatif au cours d'une nuit de sommeil. Nous montrons entre autres qu'une seule source de mémoire peut être réactivée à plusieurs reprises dans plusieurs rêves, et en différents stades de sommeil, ce qui pourrait permettre un traitement continu ou séquentiel de souvenirs épisodiques avec d'autres souvenirs tout au long de la nuit. Dans l’ensemble, nos études quantitatives et qualitatives des incorporations de mémoire dans les rêves permettent d’éclairer les mécanismes fondamentaux de la formation des rêves ainsi que leurs associations avec le traitement et la consolidation des mémoires épisodique et procédurale. Nos résultats suggèrent qu’un rôle potentiel du rêve dans la mémoire irait au-delà de sa simple réactivation, soutenant des processus de transformation et d’intégration de la mémoire. La création de scénarios multisensoriels et immersifs basés sur des fragments mémoire est possiblement centrale à ces processus et permettrait d’optimiser l’utilisation de la mémoire pour le futur. / Sleep plays an important role in memory consolidation. New experiences acquired while awake are reactivated spontaneously during sleep, a process that is thought to facilitate their consolidation and integration into longer-term memory. It has been suggested that these memory reactivations are, at least partially, reflected in dream content and that dreams play an active role in memory consolidation. The general objective of this doctoral thesis is to assess these claims; I examine whether and how dreams are involved in episodic and procedural memory processes by using new technologies and experimental approaches to study relationships between memory and dreaming. Our first study aimed to influence memory reactivations during sleep in order to clarify their relationships with dreams and procedural learning. We experimentally stimulated memory reactivations by re-exposing participants during their sleep to an auditory stimulus previously associated with motor learning, a process called targeted memory reactivation (TMR). We show that TMR, when applied during rapid eye movement (REM) sleep, improves performance of a complex motor task, i.e., learning how to "fly" in a virtual reality (VR) setting. Moreover, dreaming about kinesthetic elements of the motor task in REM sleep, but not in stage 2 sleep, is associated with better improvement on this task (article I). These results support previous models suggesting that REM sleep plays an important role in the consolidation of complex procedural memory and further suggest that the simulation of sensorimotor sensations in dreams contribute to this role. Although TMR did not directly impact dreams, we show that it can influence the time course of memory incorporations in dreams over multiple days. It amplified delayed incorporations of the task 1-2 days later when applied in REM sleep, and 5-6 days later when applied in slow-wave sleep (article II). We suggest that these longer-term effects could be due to a “tagging” mechanism of memory traces, which primes or amplifies their subsequent processing over several nights of sleep. Furthermore, we show that our immersive VR task increased the incorporation of flying sensations in dreams, especially the night after exposure to it (article III). We identify individual factors that are associated with the incorporation of the flying task in dreams, such as previous experience with both flying and lucid dreams. A closer look at dreams over 10 days following the VR experience shows that flying sensations become progressively decontextualized from the original flying context over time. Flying dreams after VR exposure also had higher levels of control and emotional intensity compared to baseline flying dreams. The successful induction of flying dreams allowed us to do an in-depth qualitative analysis of them, based on which we propose a new mechanistic explanation of how flying sensations or movements may arise in dreams. These results could facilitate the development of technologies to better influence and study dreaming. Our second study aimed to assess when and how an episodic memory source shared by all participants – visiting the sleep laboratory – is incorporated into dreams. The results show that almost a third of dreams incorporate elements of the laboratory, particularly REM dreams from a morning nap (article IV). A qualitative study of these dreams highlights the typical ways in which elements of the laboratory reappear in dream narratives. We suggest the existence of different “integrative pressures” that structure memory fragments into these dream scenarios. These are often performative or social in nature, project the dreamer in time and space, or incorporate real sensations from the sleeping body or the sleep environment. Studying the phenomenon of dreaming about the laboratory helps clarify how dreams are shaped from memory fragments, and highlights the advantages and methodological limits of laboratory dream and memory studies. Finally, our third study evaluated the time course of autobiographical dream memory sources during an entire night of sleep using a serial awakenings protocol. Our results show that dreams can combine multiple memory sources at once, especially when they occur at sleep onset or in REM sleep, which may reflect a greater dream richness or a more widespread associative memory activation in those stages (article V). We replicate the finding that recent memories are preferentially reactivated early in the night, while more distant memories are relatively more represented in late night dreams – we here clarify that this effect is independent of sleep stages. The coactivation of multiple memory sources in a dream narrative aligns with the suggestion that a function of sleep is to integrate new knowledge with existing knowledge. Our results further suggest that the latter may be a cumulative function of a night of sleep. We show that a single memory source can be repeatedly reactivated in multiple dreams in different sleep stages, which could allow a continuous or sequential processing of episodic memories with other memories across the night. Overall, our quantitative and qualitative studies of memory incorporations in dreams shed light on fundamental mechanisms of dream formation and on their association with episodic and procedural memory processing and consolidation. Our results suggest that a potential role of dreams in memory goes beyond simple reactivation, supporting long-term processes of memory transformation and integration. The creation of immersive and multisensory dream scenarios built upon memory fragments may be key to these processes and to optimizing the use of these memories for the future.