Abstract

Introduction: Successful storage of memory can be divided into three fundamental processes: encoding, consolidation and retrieval. During encoding, information is acquired e.g. in a learning session of an experiment. New mnemonic traces are formed in the brain. When the information needs to be remembered e.g. at the retrieval session of an experiment, memory needs to be recovered. Since encoding does not lead to instantaneous permanent storage of the learned material, a form of memory stabilization is necessary. A person’s freshly acquired memory is initially fragile until the memory trace is reinforced through a process of consolidation. While learning and retrieval must occur during wakefulness, memory consolidation can occur during sleep. One question that is still under debate in sleep literature is whether a period of sleep, in comparison to wakefulness, significantly and persistently benefits the consolidation of recently and explicitly acquired declarative information (such that memory retrieval after a period of sleep is significantly better than retrieval after a period of wakefulness). A further problem regarding the benefit of sleep for memory is the discrepancy between functional and behavioral findings: Sleep associated changes are possibly a covert process and changes on the anatomical level are not necessarily congruent with behavioral results. Another question concerns memory consolidation in the long run. In humans, the medial temporal lobe, especially the hippocampus, is an important brain structure involved in declarative memory retrieval. Through the process of consolidation, declarative memory has been found to become independent of the hippocampus over time. Yet, human imaging studies investigating memory retrieval for a longer period of time (several months) are scarce. Another gap of knowledge lies in the role of the hippocampus. Several different hypotheses about its role exist: The multiple trace theory, established by Nadel and Moscovitch (1997), states that personally experienced episodes stay hippocampus dependent, whereas semanticized memories become independent over time. O’Keefe et al. (1978) proposed that the hippocampus is permanently accessed for spatial memory retrieval. According to Eichenbaum (2000), the hippocampus binds new information coupled with an episode into a network of existing memory traces. This thesis focuses on long-term memory. The major focus lies on declarative memory, whereas the minor focus lies on non-declarative memory. All five studies of this thesis investigate declarative memory and the last study (study 5) additionally investigates non-declarative memory. Study 1: Objective: To investigate the relation between episodic (declarative) memory and sleep versus sleep deprivation on the functional and behavioral level. The aim is to do the investigation on a time scale of 2 ½ months. Methods: The analysis was based on a between-group (factor: sleep / wake), within-subject (factor: autobiographical task / spatial task) design. Each subject learned two episodic memory tasks (word associations): an autobiographical task and a spatial task. Brain activity (using a 3T MRT) and behavioral performances were measured at 3 times: 1) Immediately after learning; 2) after a night of sleep/wake and two recovery nights of sleep; 3) 2 ½ months after learning. Results: No sleep related changes in hippocampal activation could be concluded from the neuroimaging results. Supporting this, behavioral results (free recall) showed no difference between sleep and sleep deprivation groups. Recall results showed no difference between the sleep group and the sleep deprivation group. Study 2: Presuming that sleep supports hippocampus dependent declarative memory, but given the results of study 1, it was important to investigate the role of the hippocampus. Objective: This study focused on the role of the hippocampus in declarative memory retrieval, given the different hypotheses (mentioned above) about its role. Methods: Using a between-group design, hippocampal involvement during free recall at an early stage after encoding was compared between sequential, spatial and autobiographical learning strategies. (Study 2 was not a sleep-study). Free recall performance of concrete nouns was measured on the functional as well as behavioral level. Results: Not all episodic memory traces depended equally on the hippocampus when information was retrieved in free recall: Whereas recall of autobiographical memory relied on the hippocampus after consolidation, recall of spatially and sequentially associated information did not. Functional conjunction analyses showed that brain areas mutually involved in all tasks tested, were: the precuneus (medial parietal cortex), medial occipital gyrus and superior parietal lobe (SPL). Studies 3 – 5: The specific mechanisms underlying the process of memory consolidation are still not clarified. It has been suggested that a positive effect of sleep on memory occurs when a sensitive set of requirements is met, although to date, pinpointing the exact requirements has not been possible from sleep literature. Study 3: Objective: The question to be answered was: Is the type of retrieval, that is, cued recall or recognition, crucial for an effect of sleep on declarative memory? Methods: The following parameters were applied: i) Cued recall and recognition as the type of retrieval test; ii) Circadian rhythm: Learning either in the morning or in the evening; iii) The retention period between learning and the post-conditional test was kept constant at 12 hours; iv) Interference learning was used; v) The learning material was restricted to non-sense syllables. Results: A beneficial effect of sleep on memory retrieval 12 hours after learning non-sense syllables occurred only when syllables were tested via cued recall. However, results were influenced by circadian rhythm effects with better test scores in the morning than in the evening. Study 4: Objective: Same as in study 3, but controlling for the circadian rhythm effects by using nap sleep instead of nocturnal sleep. Methods: Circadian rhythm effects were controlled by choosing a 60 minute nap sleep paradigm, in which encoding and retrieval both took place at the same time of day (in the afternoon), for both the sleep and wake conditions. The two types of retrieval in relation to nap sleep and wakefulness were examined: cued recall and recognition. The following parameters were applied: i) Cued recall and recognition for the type of retrieval test; ii) Circadian rhythm: Learning in the afternoon; iii) The retention period between learning and the post-conditional test was kept constant at three hours (including a 60 minute nap or time spent awake); iv) Interference learning was used; v) The learning material consisted of concrete German nouns. Results: subjects did not perform significantly better after a period of napping compared to a period of wakefulness, neither for words tested via cued recall nor words tested via recognition. A sleep benefit on the behavioral level did not show to be specific to the type of retrieval test. Study 5: Objective: To examine whether a sleep benefit occurs between a critical period of 12 to 144 hours post learning. In addition to declarative memory, the relation between sleep and procedural memory is tested, using a motor sequence (finger tapping) task. Methods: Subjects learned a procedural and a declarative task. The following parameters were applied: i) Free recall for the declarative and procedural retrieval tests; ii) In contrast to the other studies, total sleep deprivation and daytime wakefulness were used as wake condition iii) The retention period between learning and testing was 12, 72 or 144 hours (3 groups); iv) Interference learning was not used for the declarative task (a main and new motor sequence task were learned); v) The learning material was restricted to non-sense syllables. Results: No beneficial post-learning effect of sleep could be detected in the declarative and procedural tasks over the retention interval of up to six days. Results of study 5 demonstrated that sleep after learning did not lead to better performance of motor skills than wakefulness after learning. Conclusion: From the results of the five studies of this thesis, it can be concluded that declarative and procedural memories are consolidated equally well over a period of wakefulness compared to a period of sleep. The type of retrieval, circadian rhythm, retention period, interference, and the type of material might all contribute to a set of variables influencing the benefit of sleep on memory. It can also be assumed that the human brain is capable of compensating a night of sleep deprivation without significant behavioral deficits during retrieval of verbal declarative and motor skill tasks, whether memory is tested shortly after encoding (a few hours), after days or after months.