Sleep changes following intensive cognitive activity

Journal Pre-proof Sleep Changes Following Intensive Cognitive Activity Mariangela Cerasuolo, Francesca Conte, Fiorenza Giganti, Gianluca Ficca PII:

S1389-9457(19)30302-8

DOI:

https://doi.org/10.1016/j.sleep.2019.08.016

Reference:

SLEEP 4171

To appear in:

Sleep Medicine

Received Date: 11 January 2019 Revised Date:

28 August 2019

Accepted Date: 28 August 2019

Please cite this article as: Cerasuolo M, Conte F, Giganti F, Ficca G, Sleep Changes Following Intensive Cognitive Activity, Sleep Medicine, https://doi.org/10.1016/j.sleep.2019.08.016. This is a PDF file of an article that has undergone enhancements after acceptance, such as the addition of a cover page and metadata, and formatting for readability, but it is not yet the definitive version of record. This version will undergo additional copyediting, typesetting and review before it is published in its final form, but we are providing this version to give early visibility of the article. Please note that, during the production process, errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain. © 2019 Published by Elsevier B.V.

SLEEP CHANGES FOLLOWING INTENSIVE COGNITIVE ACTIVITY Mariangela Cerasuoloa, Francesca Contea, Fiorenza Gigantib, Gianluca Ficcaa

a

Department of Psychology, University of Campania “L. Vanvitelli”, Caserta, Italy

b

Department of Neuroscience, Psychology, Drug Research and Child Health, University of Florence, Florence, Italy

Corresponding Author: Gianluca Ficca, MD, PhD, Department of Psychology, University of Campania “L. Vanvitelli”, Viale Ellittico 31, 81100, Caserta, Italy, Tel.: +39.0823.274791, Fax: +39.0823.274792, [email protected]

Declarations of interest: None.

1

Abstract

Studies over the last 40 years have mainly investigated sleep structure changes as a result of wake duration, in the frame of the classical sleep regulation theories. However, wake intervals of the same duration can profoundly differ in their intensity, which actually reflects the degree of cognitive and physical activity. Data on how sleep can be modified by wake intensity changes (initially sparse and of little consistence) have become much more substantial, especially in the frame of the intense research debate on sleep-memory relationships. Our aim is to examine the vast repertoire of sleep modifications that depend on waking cognitive manipulations, highlighting the sleep features that appear most affected. By systematically addressing this issue, we want to set the basis for future research exploring both the specific nature of the mechanisms involved and the applicative psychosocial and clinical fallouts, in terms of possible behavioural interventions for sleep quality improvement.

Keywords: Sleep regulation; wake intensity; cognitive activity; pre-sleep learning; enriched environment; memory consolidation.

2

1. The case for a role of wake intensity in sleep regulation Sleep and wakefulness are behavioural states characterized by a tight interdependence: as the quantity and quality of previous sleep affect subsequent wakefulness, likewise wake characteristics influence the following sleep. Assuming a mainly restorative role for sleep, the influence of sleep characteristics on subsequent wake state has been largely studied in the last century, and it is now generally accepted that a good night’s sleep is a crucial requirement for the effectiveness of many daytime cognitive processes [1]. The inverse research question, regarding the effect of wakefulness on sleep, has been mainly addressed in the frame of the classical “two-process” model of sleep regulation [2]. Based on previous wake duration (in conjunction with circadian factors), the model allows to predict “when” sleep will most likely occur and “how” deep it will be (ie, in essence the amount of Slow Wave Sleep [SWS]). A few years prior to Borbély’s model, it had been suggested that sleep is also modulated by the intensity of waking brain activity [3], initially measured through brain temperature or cerebral metabolic rate [4–6]. Later, the notion of “wake quality” was convincingly discussed by Franken [7], commenting on Huber et al.’s findings [8] of an increase of delta sleep in rats that had been subjected to an acute-dark condition, thus augmenting exploratory behavior at the expense of quiet waking. Pioneering data in rats had also shown that a similar increase of delta activity, higher than could be predicted based on sole time spent awake, was induced by a social stress experience [9]. As stated by Meerlo [9], “sleep intensity may, thus, not only depend on the duration of prior wakefulness but also on the nature of the waking experience”. These contributions highlighted that sleep-wake reciprocal influences cannot be fully enlightened without taking into account wake “content” alongside its duration. In other words, sleep quality and quantity significantly influence the quality and quantity of wakefulness and vice versa. Yet this issue has never been directly addressed, despite the relevant theoretical and applicative implications it could bear. In fact, the understanding of the influence of wake quality on sleep features could contribute refining existing models of sleep regulation as well as allow the design of behavioural protocols aimed to manipulate waking activities in such a way as to obtain desired changes in sleep characteristics. From a methodological point of view, a precise definition of wake intensity has yet to be determined and will need to deal with several issues: (1) what kind of behavioural activities are able to trigger sleep changes? In general terms, wake intensity may be intended as the quantity and nature of physical and cognitive activity carried out during wakefulness, which would affect subsequent sleep via changes of physiological and biological parameters (eg, brain temperature or gene transcription). Further, (2) what are the relevant variables which modulate these effects (eg, duration of the manipulation, effort, specific cognitive processes involved, etc.)? Moreover, (3) what kind of cognitive, physiological and biological mechanisms sustain the induced sleep changes? While these aspects need to be clarified in order to conceive theoretical and applicative models, currently more information is available on the type of sleep features that are modified by variations in wake content. Specifically, research on sleep-memory relationships has produced massive data on the effects of presleep cognitive manipulations on several sleep features other than SWS (eg, sleep spindles [10,11] and REM sleep measures [12,13]). The aim of this reiview is to provide a general overview of the studies observing sleep changes after behavioural manipulations of waking cognitive activity, highlighting which sleep features appear most affected and in which direction. We exclude literature on the effects of physical activity protocols, which have been recently reviewed elsewhere (ie, [14]). Thus, we aim to provide the basis for future

3

research exploration, with specific regards to the nature of the mechanisms involved (at the cognitive, electrophysiological and cellular levels) as well as the applicative fall-outs in terms of possible behavioural interventions for sleep quality improvement. The literature we refer to includes two major groups of studies. First, experiments based on naturalistic paradigms such as the “enriched environment” procedures. Despite the difficulty in disentangling the influence on sleep of the different cognitive functions simultaneously activated by this kind of procedure, data from this line of research have consistently confirmed, both in animals and in humans, that different waking behaviours affect sleep need and sleep features. Second, the vast body of data on post-learning sleep changes coming from literature on sleep-memory relationships. One of the most common paradigms used in this field relies on the comparison of a baseline sleep episode with one preceded by a learning task: the changes observed in post-learning sleep features are believed to reflect the involvement of such features in the overnight memory consolidation process. This paradigm has been often preferred to sleep manipulation paradigms (ie, comparing the effects on memory of early and late sleep episodes or of selective sleep deprivation conditions relative to undisturbed sleep), since it allows to overcome potential biases linked to sleep manipulation [15]. Although initially introduced to investigate the role of sleep in memory consolidation, this research design also inversely provides hints on the role of learning in sleep regulation since it implies experimental manipulations of wake intensity. We would like to specify that, while our main source of data relies on the large body of research on sleep and memory, we do not intend to provide an update on the state of the art in this field, which already appears an overflowing topic in scientific literature. Instead, our interest in these studies is due to the fact that they can indirectly shed light on learning- and experience-dependent sleep changes. In light of our main aim (ie, to highlight which sleep characteristics appear mainly affected by cognitive activity manipulations) we have organized our review according to different groups of sleep variables, both in terms of sleep macro- and microstructure.

2. Method: data sources and study selection An extensive research was carried out on Pubmed to locate studies addressing the effect of cognitive activity on subsequent sleep. The following keywords were used, in all possible combinations: “sleep changes”, “cognitive training”, “learning”, “memory”, “sleep disorders”, “insomnia”, “sleep depth”, “sleep continuity”, “post-training sleep”, “enriched environment”, and “social enrichment”. We considered only articles written in English and published from 1985 until June 2018. In addition, we examined the reference lists of relevant articles and reviews, in order to identify other studies not included in the search results. After a thorough screening of titles and abstracts, we included only those studies investigating sleep modifications after a manipulation of cognitive activity. No restrictions were applied regarding the study design (within- or between-subjects), the samples (animal or human, healthy or clinical populations), the measures assessed (objective or subjective) or the duration of the intervention (single or several days). Cognitive manipulation paradigms based on fear conditioning procedures were excluded because of the possible confounding effects of stress.Database searches produced 7253 records. Twenty-nine additional articles were selected from reference lists. After screening of titles and abstracts, 182 full-text articles were obtained for further analysis and 85 articles met our inclusion criteria. Selected articles were primarly subdivided in animal and human studies, then organized according to the dependent sleep variables assessed.

3. Overviewing the effects of cognitive activity on sleep variables

4

Our review has been organized by separately describing all the sleep variables that have shown to be affected by manipulating cognitive activity in previous wakefulness. In each paragraph, starting with data from animal studies and proceeding to human literature, we will list only the findings of significant changes of that given variable in post-training sleep; however, we will present other, negative, findings in a final summarising table. 3.1. SLEEP DURATION A preliminary terminological remark is necessary regarding the sleep variables discussed here. While in humans Total Sleep Time (TST) refers to the duration of the single “anchor” sleep episode which follows training, in animals its definition is quite different, since the majority of species studied (ie, insects, rodents) are characterized by a polyphasic sleep pattern and their ultradian rest-activity cycles are independent of time of day [16]. Therefore, most studies on this kind of sample assess sleep-wake behaviours over long periods (several hours or days) rather than focusing on a specific sleep episode. Within these periods, we consider it more correct to refer to “global sleep duration”, while conversely the duration of time spent awake over the whole recorded time lapse being usually referred to as “wake time”. Modifications of global sleep duration have been documented in animals exposed to enriched environments. Two rat studies assessed sleep behaviour in rats kept for six [17] and seven [18] weeks in enriched cages (ie, cages in which multiple physical structures, believed to stimulate speciesspecific behaviours, are added). Compared to control individuals living in unenriched cages, the experimental group exhibited increased sleep duration [18] and number of sleep bouts [17]. Social enrichment procedures have repeatedly shown, in insect models, that social interactions markedly affect sleep pressure. Increases in global sleep duration and sleep bouts duration were reported after five day expositions to social interactions in fruit flies (Drosophila melanogaster) compared to individually housed siblings [19–21] and these effects were proportional to the size of the social group the flies were exposed to [19]. Furthermore, Chi et al. [22], observed that higher larval population density during early development results in more consolidated sleep (less but longer sleep episodes) in female fruit flies. The effect of social enrichment on sleep has also been investigated in honey bees. Insects experiencing a colony environment for one or two days after birth slept more frequently and spent more time asleep compared with same-age siblings that were caged individually or in small groups outside the colony [23]. Furthermore, bees placed in mesh-enclosures in the colony, that prevented direct contact with nestmates, slept similarly to bees freely moving in the colony, suggesting that also social signals not requiring close distance interactions are sufficient to produce effects on sleep [23]. Notably, the effect on sleep time of social enrichment procedures appears to persist for several days [19,23]. Increases in TST have also been obtained in male flies after a conditioning procedure in which they learned to modify their courtship behaviour based on previous exposure to unreceptive females [19]. Finally, an increase in global sleep duration was reported in rats after administration of the two-way shuttle avoidance task [24]: the increase, relative to baseline, was evident only in rats showing improvements at re-test. The abovementioned results are in agreement with the specular finding of a reduction of wake time in rats following several kinds of cognitive manipulations: after exposure to an enriched environment [25], enhanced exploratory behaviour [8], and a rewarded olfactory discrimination task [26].

5

Despite this massive body of data in animals, only a few human studies have produced evidence of experience-dependent changes in sleep duration. Increases in sleep time have been found in healthy elder individuals after word pairs learning [27], in healthy adults after procedural learning (only in subjects with high baseline performance) [28] and after a complex ecological cognitive task [29], as well as in sleep-disordered adults after an 8-session forest walking program [30]. 3.2. SLEEP ONSET LATENCY In contrast to the assumption that pre-sleep activity hinders sleep propensity (eg, [31]), several studies report sleep latency reductions after enhanced cognitive activity. In a study conducted on Drosophila males, more rapid sleep onset was observed after long-term pair-wise social interaction [21]. In humans, sleep latency was reduced after a behaviourally “active” day [32], after a four-choice visual motor task [33] and in a nap following a complex cognitive task performed on a tablet [29]. Moreover, elderly insomniacs’ average sleep latency was reduced during the week following an eight week computerized cognitive training programme compared to baseline assessment [34]. Only two studies (one on adolescents exposed to an interactive computer game [35], the other on young adults performing a set of diverse cognitive tasks [36]) observed increased post-training sleep latency relative to baseline. For the latter, the increase in objective sleep latency was not paralleled by that subjectively reported [36]. 3.3. STAGE 1 SLEEP No data concerning Stage 1 and Stage 2 are reported in animals, for whom sleep stages are limited to synchronous sleep and paradoxical sleep, respectively corresponding to human SWS and rapid eye movement (REM) sleep states. In humans, results on Stage 1 sleep are sparse. Post-training decreases of Stage 1 proportion have been found in six human studies: two were conducted with a nap paradigm [29,37] while four explored night-sleep episodes [27,28,33,38]. These results have been usually interpreted as reflecting increases of time spent in deeper sleep stages or a decrease in transitions to shallower sleep. 3.4. STAGE 2 SLEEP Several human studies have shown lengthened Stage 2 sleep after cognitive training, both in night sleep episodes [11,28,39,40] and in naps [29,41]. Of note, these changes emerged only after tasks involving procedural skills. In one study [28], the post-learning increases of Stage 2 sleep, compared to baseline, were observed only in subjects showing high performance levels at acquisition. Fogel et al. [40], studied the role of different sleep features over the time course of skill acquisition by recording sleep in four different conditions: (a) control sleep (preceeded by a simple cognitive task), (b) novice sleep (following administration of the Tower of Hanoi task), (c) expert sleep (recorded after a week during which subjects gained proficiency on the task through repeated exercise), and (d) re-test sleep (following re-administration of the task a week after the previous condition). The increase in Stage 2 proportion and total duration appeared only in the re-test condition, supporting Smith et al.’s hypothesis [42] of a role for Stage 2 sleep in the stabilization and maintenance of existing skills. Reductions in Stage 2 proportion compared to baseline sleep were found after a serial reaction time task in adults [33] and a rotory pursuit task in adolescents [43]. In the latter study, the decrease was greater in subjects performing poorly at morning re-test [43]. Furthermore, a significant decrease of Stage 2 proportion emerged after an “active day” in an enriched environment study [32].

3.5. SLOW WAVE SLEEP

6

Animal studies have consistently shown increases in SWS amount after different kinds of cognitive manipulation: associative olfactory learning [26], exposition to enriched environments [25] or novel objects [44] and intensified exploratory activity [8]. Additionally, increases in average duration of SWS episodes were found in rats after two-way shuttle avoidance training [45] and intense exploration [8]. In one study [24], increases in SWS duration and in average duration of SWS episodes (relative to baseline sleep) were found after two-way shuttle avoidance training only in those rats showing no improvement during training but attaining the learning criterion at re-test; whereas no change emerged in untrained rats and “non learning” rats (trained rats showing no improvements in either session). Conversely, Ambrosini et al., found an increase in average duration of SWS episodes after the same task emerged only in unsuccessful rats (showing no improvement at neither acquisition nor re-test) [46]. Alternately, a decrease in SWS emerged in one rat study after avoidance training (only in rats showing improvements at re-test) [47] compared to baseline sleep. In humans, increases in the total amount of SWS were found after exposition to an enriched environment [32], following acquisition of complex motor skills [37,48], sequential finger tapping [49], a serial reaction time task [33], a two week daily program of combined social and physical activity sessions in elderly subjects [50] and a rotary pursuit task in adolescents (here the increase emerged only for subjects who did not show performance improvements at morning re-test) [43]. Moreover, in a study aimed to compare postlearning sleep changes in young and older adults, the administration of the pursuit rotory task was followed by an increase in SWS duration compared to the baseline night only in the older group [38]. In contrast, a decrease in SWS proportion compared to baseline was found in a group of preadolescents exposed to an interactive computer game before bedtime [35]. Numerous studies have also investigated changes in Slow Wave Activity (SWA, 0.5-4 Hz) as a function of previous cognitive activity. Increases of SWA have been reported after intensive exploratory activity in rats [8]. For human subjects, SWA increases have been reported following complex motor skills training (three-ball cascade juggling) [37] and an intensive training at the Tower of Hanoi task [40]. In the latter study, the increase in SWA emerged both in the novice and re-test conditions compared to control, suggesting an involvement of SWA both in initial acquisition of a skill and in its stabilization once expertise is attained [40]. Another set of results on SWA is particularly relevant since it focuses on its local experience-dependent changes (ie, selectively observed over the cortical areas specifically involved in cognitive processing of the task administered before sleep). Taking into account SWA’s purported role in synaptic downscaling [51,52], a growing literature has led to consider SWA as a marker of experience-dependent plasticity. This line of research was initiated by Huber et al. [53], showing selective SWA increases in parietal areas after an implicit learning task (rotation adaptation) in human adults. In subsequent studies, these data have been replicated with different tasks and populations. In rats, it was shown that the administration of the single pellet reaching task produced, during post-training non-rapid eye movement (NREM) sleep, an increase of SWA in the trained (motor) cortex, with smaller or no increase in other cortical areas [54]. In addition, training appeared to enhance the expression of genes for activity-dependent proteins involved in motor learning and this increase was restricted to the same cortical area [54]. Also, a selective decrease in visual cortex SWA was found in mice and cats after a period of sensory deprivation (dark-rearing for the three-four months immediately after birth), with no effect on time spent in SWS, compared to control animals; a gradual recovery of local electroencephalogram (EEG) differences was seen over one-two months upon returning the animals to a normal, lighted environment [55].

7

In humans, local SWA increases in post-training sleep episodes were found in three other studies [56– 58] using the same task as in Huber et al., [53]. Notably, in [56], the local SWA enhancement was observed even when the task was performed in the morning rather than at bedtime. Similar SWA increases have been obtained following a finger tapping task [59], a spatial navigation task [60] and after a three week training on a visuospatial N-back task [61]. Moreover, local SWA increases have been reported in children and adolescents [57,61]. In particular, a study comparing adults, adolescents and children [57] detected SWA increases in all groups, with children showing the greatest: the authors propose that brain maturation processes favour experience-dependent plasticity. A recent experiment aimed to investigate whether local post-training changes in SWA are due to overnight consolidation or to metabolic demand [58]. Participants were administered three different tasks on separate days: a single rotation task, similar to that used in [53]; a random rotation task, requiring the same cognitive effort without triggering specific memory processes; and a no-rotation task, requiring minimal effort and attentional resources. Since parietal SWA was increased in both rotation conditions compared to control and overnight gains in the single rotation condition did not correlate with SWA changes, the authors concluded that SWA may be modulated by levels of cognitive effort during prior wakefulness rather than just consolidation [58]. 3.6. REM SLEEP Increases in REM sleep amount after intensive training sessions and exposure to enriched environments have been consistently reported in animals since early studies: those antecedent to 1985 are extensively reviewed in [62]. After, post-training REM sleep increases have been found in rats and mice, using spatial learning [63], novel object exposition [44], environmental [25] and social enrichment [64], as well as active avoidance training (two-way shuttle avoidance task) [47,65–68]. After a complex operant task, one study reported that time in REM sleep was greater in rats able to solve the task compared to those who were not and to the untrained control group [69]. Several studies have reported decreases or increases of REM sleep variables depending on whether the animals were successful in learning the conditioned response [45,46,70]. For example, after avoidance training, a reduced number of REM sleep episodes compared to baseline emerged only in “non learning” rats (showing little improvement during training) [46,70] and in “slow learning” rats (displaying improvement only at re-test) [45], as opposed to “fast learning” rats (attaining the learning criterion already in the training session). In human studies, results on changes in REM sleep amount after cognitive activity appear equally consistent. Increases of time spent in this sleep stage have been found after learning of complex motor skills [48,71], expertise acquisition of complex procedural learning [40], Morse code learning [72], implicit serial reaction time task [33], a six week foreign language course [73], a rotary pursuit task in adolescents [43], and prolonged working-memory training in children and adolescents [61]. It is noteworthy that in De Koninck et al’s study [73], REM sleep was increased only in subjects showing successful language learning at the end of the course. Conversely, Nader et al., found that the rotor pursuit task produced REM increased only in adolescents displaying no performance changes at morning re-test [43]. Only one study [74] detected a decrease in REM sleep proportion after a simple virtual maze task (but not a complex maze) relative to a control group. Finally, task-dependent changes in theta oscillations (4-8 Hz), considered the electrophysiological hallmark of tonic REM sleep, have been investigated in two studies. In one study avoidance training was followed in rats by an increase in REM sleep theta activity, and this increase was greater in rats showing successful avoidance learning at retrieval compared to unsuccessful rats [47]. In humans, neocortical theta activity during REM sleep was found to be enhanced in the central regions after training on word pairs [11] and a decision-making task [75].

8

3.7. NREM-REM CYCLES, SLEEP STATE SEQUENCES A number of animal studies have investigated sleep state sequences and their changes following cognitive tasks. In rats, increases in the number [24,70] and average duration [46,70] of NREM-REM sleep sequences, corresponding to human NREM-REM cycles, and of time spent in these sequences [24,70] were reported following two-way shuttle avoidance training. These changes emerged both for successful and unsuccessful learners (rats attaining or not attaining the learning criterion during the training session) [46,70] and for “slow learners” (rats with no improvement during training but attaining the learning criterion at re-test) [24], relative to the control group. An increase of time spent in NREM-REM sleep sequences was also shown after a spatial habituation task [76] and after exposure to an enriched environment in old aged rats (but not young rats) compared to baseline sleep [25]. The organization of human sleep in NREM-REM cycles also appears to be boosted by pre-sleep learning. An early study by Buchegger and Meier-Koll [48] found an increase in time spent in sleep cycles after an eight week motor learning training. Two recent studies from our group have replicated the finding with declarative learning: increases in the number of complete sleep cycles and in total time spent in cycles (percentage over actual sleep time) emerged in elderly subjects after word pairs learning [27] and in adults after learning a theatrical monologue [77]. 3.8. WAKE AFTER SLEEP ONSET, SLEEP EFFICIENCY Wake after sleep onset (WASO) was not assessed in the animal studies of our selection, and none of the latter found post-training changes in sleep efficiency. As for human research, an increase in sleep efficiency paralleled by a decrease in WASO has been observed in healthy adults after (1) pre-sleep administration of a word game performed on a tablet [29], (2) a modified version of a serial reaction time task, created in order to induce implicit encoding of a hidden sequence [33], and (3) a proselearning task [77]. The same results were obtained in healthy elderly subjects after declarative learning [27] and in elderly insomniacs after an eight week computerized cognitive training program [34]. In a study aimed to investigate the effects of different degrees of mental activity on subsequent sleep [78], it was found that the only sleep variable affected by the heavy mental workload condition (computerized cognitive tasks involving sustained attention, memory, logical thinking, decision making, and calculating) was the percentage of WASO, which was reduced relative to the light mental activity condition (video session). In contrast, a decrease in sleep efficiency was found in a group of pre-adolescents watching a subjectively exciting movie compared to baseline sleep [35]. 3.9. BEHAVIOURAL AWAKENINGS, AROUSALS, STATE TRANSITIONS Only one animal study has assessed sleep continuity after cognitive manipulation: In rats, Huber et al. [8], showed that increased exploratory activity reduces the number of brief awakenings in following sleep. Similar reductions in sleep fragmentation, expressed by decreased frequency of awakenings, were found in humans after the administration of a word pairing task [27], an implicit learning task [33], a theatrical monologue [77], and an ecological cognitive task [29]. Furthermore, sleep was less fragmented in elderly insomniacs after an eight week cognitive training program [34]. Recently, Sergeeva et al. [79], trained subjects with Periodic Limb Movements (PLM) on a procedural (motor sequence) and a declarative (word pairs) task. Post-training sleep, compared to baseline, showed a significant reduction in the number of arousals and awakenings: this improvement was such that subjects’ post-training sleep quality was comparable to that of healthy controls. While no data is

9

available in animals, pre-sleep training appears to affect sleep stability parameters in human subjects, by reducing the frequency of arousals [27,77,79,80] and state transitions [27,29,77]. Contrasting results come from Kirov and coworkers [33], who showed that an implicit learning task induced increased frequency of transitions between sleep states compared to baseline sleep; the transition rate was significantly higher in subjects who developed an explicit knowledge of the task compared to those who did not. 3.10. SUBJECTIVE SLEEP QUALITY Notably, only three studies investigated the impact of cognitive activity on subjective sleep quality. In a sample of healthy elderly individuals, De Almondes et al. [81], compared the effects on cognitive functioning and subjective sleep quality of three interventions (six 90-minute sessions): a cognitive training program, aimed to promote executive functioning skills, a sleep hygiene psychoeducation program, and a cognitive training plus sleep hygiene psychoeducation program. All three interventions resulted in improved subjective sleep quality compared to the no-intervention control group. Specifically, the sleep hygiene program group showed the greatest benefits in subjective sleep quality, followed by the cognitive training group, while the combined sessions did not provide any additional gain. Benloucif et al. [82], also studied a sample of older adults. The intervention consisted in a two week program in which participants were administered daily (either in the morning or in the evening) a 90-minute session of combined physical and social activity. Subjective sleep quality ratings improved both after morning and evening sessions, whereas objective measures (recorded by means of polysomnography and actigraphy) displayed no change. The improvement in subjective sleep quality was limited to a sub-group of bad sleepers, while good sleepers showed no benefit of training. The other study was carried on a group of sleep-disordered adults [30]: after an 8-session forest walking program (administered over four months), subjects reported greater sleep depth and higher sleep quality compared to baseline sleep. 3.11. NREM-RELATED PHENOMENA NREM sleep microstructure has received great attention in the last two decades. Below we will describe evidence on the effects of cognitive tasks on four NREM sleep-specific electrophysiological features: sleep spindles, slow oscillations (SOs), sharp wave ripples (SWRs) and Cyclic Alternating Pattern (CAP). Few studies have addressed the effects of learning on spindles in animals. Relative to baseline sleep, increases in spindle density (number/time unit) were found in rats after an odor-reward association task [83], and in dogs trained to respond to commands in an unfamiliar language [84]. Results on humans consistently point to an enhancing effect of cognitive training on spindle activity. Spindle density increases have been frequently reported in young adults after procedural rotary pursuit [11,28,38,39,85], finger tapping [49,86], visuomotor learning [87] and Tower of Hanoi [40] tasks. Analogue results have been produced following tasks in the declarative and spatial domains: unrelated word pairs [10,88] and virtual maze navigation [74]. Of note, in Schmidt et al. [88], the spindle density increase emerged only in the difficult encoding condition (list containing more abstract words) compared to the easy encoding and to the control condition. Furthermore, Peters et al. [38] studied both a sample of young and of older adults: the spindle density increase was evident only in the young group. A decrease in spindle density was reported only in one study [89], after administration of a declarative learning task: the decrease emerged both after learning neutral and emotional material (independently of emotional valence).

10

Fewer studies have assessed spindle duration and amplitude rather than density: spindle duration was increased following a finger tapping [49] and a mirror tracing task [90]. In addition, spindle amplitude was enhanced by mirror tracing [90,91]. Based on their frequency, two types of spindles have been identified: slow (<13 Hz) and fast (>13 Hz) spindles. Several studies using procedural tasks suggest that fast spindle density is selectively enhanced by training [40,86,90,91]. Alternately, a selective increase of slow spindle density was found in one study using a declarative task [88], only in the difficult encoding condition versus the easy encoding and control ones. Other studies have assessed changes in the EEG sigma power band during post-training NREM sleep, with the assumption that sigma power reflects spindle activity. Increases in sigma power were found after different kinds of learning sessions, again mostly involving procedural tasks: texture discrimination [92], finger tapping [49,59], Tower of Hanoi [40], rotory pursuit [11], and three ball cascade juggling [37]. As for declarative tasks, two studies using the same task (unrelated word pairs learning) found opposite results: while Schmidt et al. [88], reported a post-training increase of low sigma power (12-14 Hz, corresponding to slow spindle activity), Fogel et al. [11], observed a decrease of the same measure. Finally, a number of studies have observed that the changes in spindle measures were specifically detected over the areas involved in the pre-sleep task [11,49,59,87,91,92], suggesting, as for SWA, that spindle activity may play a major role in the re-processing of prior wake cognitive experiences. In agreement with this idea, an intriguing experiment using a simultaneous EEG- Functional magnetic resonance imaging (fMRI) technique [93] reported that learning face-scene associations triggered an increase in spindle-coupled neocortical activity, despite the absence of significant changes in sleep measures (including spindle variables). In sum, the reactivation during sleep of neocortical and hippocampal regions occurred in temporal synchrony with spindle events and was tuned by ongoing variations in spindle amplitude. These task-dependent changes were topographically specific to the brain areas engaged in pre-sleep learning. SOs and SWRs have lately become the focus of much sleep-memory research due to their purported involvement in long-term plastic changes at the cellular level [94]. Experience-dependent changes in SWR and SO measures have been consistently shown in animal and human studies. In rats, an odorreward association task triggered an increase in the density (number/sec), duration and magnitude (microV/sec) of ripple events in subsequent SWS, compared to baseline sleep, only in rats showing performance improvement during the training session [95]. In addition, ripple density increases have been reported after a 10-day training period at a spatial task [96]. Several human studies showed taskdependent enhancements of slow oscillation parameters. Increases in the 0.5-1 Hz power band have been reported after complex motor skill learning in healthy adults [37] and after a finger tapping task in epileptic patients [97]. Of note, in the latter study, this change selectively appeared after the procedural task, while sleep was unaffected by word pairs learning [97]. Heib et al. [98], showed that word pairs learning produced an increase of down-state amplitude and a trend towards increased upstate amplitude of SOs (only in subjects who showed overnight memory improvements). Some studies [99–101] have specifically addressed the interplay between SOs, SWR complexes and spindle activity, suggesting that pre-sleep training strengthens the top-down control of SOs on spindles and ripple complexes. Mölle et al. [99], compared the effect of learning on sleep in humans and rats by using respectively a word pairs task and an odor-reward association task. In both samples, prior learning triggered increases in the amplitude of the SO up-state and in spindle activity during these upstates; additionally, ripple activity, measured only in the rat sample, displayed an increase that was not synchronized to the depolarized up-state. Another study [100] reported that post-learning sleep, relative to a non-learning condition, increased the occurrence of “trains” of SOs and spindle activity: in the learning condition SOs were preceded by enhanced fast spindle activity and followed by

11

enhanced slow spindle activity. Similarly, Ruch et al. [101], showed that training on a declarative task before a nap resulted in a redistribution of sleep spindles from down- to up-states of Stage 2 SOs; also, spindle density during the up-states was higher in the experimental nap compared to the control nap. Finally, the slow-wave components of CAP (A1 subtypes) displayed a significant increase in the night following a motor learning task [102]. 3.12. REM-RELATED PHENOMENA Recent literature on REM sleep microstructure after cognitive manipulations is sparser compared to that on NREM sleep. In animals, increases in rapid eye movements (REMs) have been frequently reported in early studies, reviewed in Smith [62], while no result is available after 1985. For humans, REM density (number/time unit) increases have been observed following a prolonged (three months) learning period in college students [12], after rotory pursuit [28], mirror tracing [11] and mirror tracing plus Tower of Hanoi tasks [13]. Finally, pontine wave density during REM sleep was found to increase in rats following training on an active avoidance task [66,67].

4. How sleep is modified by previous cognitive activity In Table 1 we have reported for each sleep variable the number of animal (panel 1) and human (panel 2) studies which have or have not found significant sleep modifications after manipulation of wake cognitive activity.

a)

Animal studies

SLEEP VARIABLES Global sleep duration (min) Wake Time (absolute duration, %) Sleep Onset Latency (min) Slow Wave Sleep (absolute duration, mean duration of episodes, %) Slow Wave Activity (power in delta band) REM sleep (absolute duration, mean duration of episodes, %) REM sleep theta power NREM-REM cycles (number, absolute duration, mean duration of episode) Awakenings (%) Sleep spindles (density, mean duration, amplitude) NREM sleep sigma power Slow Oscillations (number, length) Sharp Wave Ripples (density, duration, amplitude) Pontine Waves (density) b) Human studies

INCREASE 7 studies [17–23]

DECREASE 3 studies [8,25,26] 1 study [21]

4 studies [8,25,26,44]

5 studies [54,63,66,67,96]

2 studies [8,54] 7 studies [25,44,63–67]

1 study [54] 2 studies [44,66]

2 studies [25,76] 1 study [8] 3 studies [83,84,99] 1 study [44] 1 study [99] 3 studies [95,96,99] 2 studies [66,67]

SLEEP VARIABLES Sleep duration (total and actual sleep time) (min)

INCREASE 3 studies [27,29,30]

DECREASE

Sleep Onset Latency (min)

2 studies [35,36]

4 studies [29,32–34]

Stage 1 sleep (absolute duration, %)

NO CHANGES 1 study [54] 2 studies [66,67]

6 studies [27–29, 33,37,38]

NO CHANGES 25 studies [11-13,32–34,36– 41,49,50,56,60,61,75,77,78,82,86,88,92, 101] 14 studies [10,27,37,41,49,50,56,59,61,77,78,82,86, 90] 27 studies [10-13,35,36,39–41, 49,50,56,59–61,71,74,75,77,82,86,88, 90–92,97,101]

12

Stage 2 sleep (absolute duration, %)

4 studies [11,29,39,41]

3 studies [32,33,43]

Slow Wave Sleep (absolute duration, %)

6 studies [32,33,37,48,50,97]

1 study [35]

Slow Wave Activity (power in delta band) REM sleep (absolute duration, mean duration of episodes, %) REM sleep theta power NREM-REM cycles (number, duration, %) Wake After Sleep Onset (%)

9 studies [37,40,53,56–61]

Sleep Efficiency (%) Awakenings (number, frequency, mean duration) Arousals (number, frequency) State transitions (number, frequency) Subjective Sleep Quality Sleep spindles (including fast and slow spindles) (number, density, mean duration, amplitude) NREM sleep sigma power

5 studies [33,48,61,71,72]

1 study [74]

31 studies [10–13,27–29,32,35–38, 41,43,49,50,56,59,60,75,77,78,82,86, 88–92,97,101] 2 studies [10,37] 2 studies [29,71]

6 studies [27,29,33,34,77,78]

11 studies [10,13,28,35,38,41,50,59,60,82,92] 12 studies [35,37,38,41,49,56,78,82,86,88,90,92]

2 study [11,75] 3 studies [27,48,77]

5 studies [27,29,33,34,77] 6 studies [27,29,33,34,77,79] 3 studies [27,79,80] 1 study [33]

3 studies [27,29,77]

3 studies [30,81,82] 14 studies [10,11,28,39,40, 49,74,85–88,90,91,99]

1 study [89]

7 studies [11,37,40,49,59,88,92] 3 studies [37,97,99]

29 studies [10,12,13,27,35–38,48–50, 56,59–61,71,74,75,77,82,86,88–93, 97,101] 29 studies [10–13,27–29,39–41, 43,49,56,59–61,71,74,75,77,82,86, 88–93,101] 2 studies [78,101]

1 study [29]

1 study [36] 2 studies [93,101]

2 studies [41,101]

Slow Oscillations (power, 2 studies [93,101] number, length, amplitude) A1 subtypes of the Cyclic 1 study [102] Alternating Pattern (number) Rapid Eye Movements 4 studies [11–13,28] 5 studies [39,40,71,72,89] (number, density) Table 1. Post-cognitive training sleep modifications in animal (panel a) and human studies (panel b). For greater clarity, results appearing only for specific subgroups of subjects (eg, improvers vs. non improvers, different age subgroups) were excluded from the table.

This overall view generally supports the hypothesis of a close relationship between wake content and sleep features. The occurrence of particular macro- and microstructural sleep changes consequently to the experimental manipulation of cognitive activity is a frequent finding. However, when looking at the nature of these changes, our synthesis of the available results (Table 1) shows very few constant findings, especially in the human sample. In fact, as shown in panel a of Table 1, while there are few results on sleep microstructure, animal data more consistently shows changes in macro-structural features, and specifically in sleep stage proportions; a longer sleep period after exposition to enriched environment in in fruit flies, a REM rebound and a SWS “local” increase in those brain areas recruited for learning in intensively trained rats. Results from human studies, difficult to compare with animals’ since they are based on different classifications of behavioural states and events, look more complex (Table 1b). First, while in rats more attention has been paid to sleep macro-structure, recent studies in humans focused on more fine-grained analysis of sleep EEG. As shown in panel b of Table 1 a frequent finding is the boosting effect of cognitive activity on specific EEG activities, namely delta and sigma power. The latter is in line with numerous other studies consistently showing increases in sleep spindles after standard procedural (eg, [11,39,49]) and, to a lesser extend, declarative learning tasks (eg, [10,88]). The only study [89] showing decreases in spindle density was based on the

13

administration of a different learning task (ie, an emotional memory task) compared to those usually associated with post-learning spindle enhancement, which probably depends on other sleep variables for successful emotional memory consolidation. Sleep continuity, reflected by decreases of fragmenting events (brief awakenings and arousals), is another parameter which is frequently boosted by pre-sleep cognitive activity, even when classical measures such as sleep efficiency and WASO do not seem modified. It is noteworthy that pre-sleep training improves sleep quality even in populations with impaired sleep maintenance (ie, in elderly individuals) whose sleep is habitually disrupted albeit in good health conditions [27], in subjects affected by PLM [79] and in insomniacs [34]. A possible mechanism proposed by the authors is that the increased sleep continuity may allow sleep-related memory consolidation to proceed with less disruption. This idea is supported by experimental manipulations of sleep continuity in rats that impair sleep-dependent memory consolidation [103,104]. Similarly, the authors [27,29,77,79] found decreased transition rates after training, interpret this result as a “stabilizing” effect on sleep states to task-induced memory demands. An opposite result was found only in one study [33], where the authors interpret the increased post-training frequency of transitions as a sign of high “inter-stage interaction”, which would represent a crucial feature of efficient consolidation processes. However, it must be noted that the different results could be reconducted to the different definitions of stability used in these studies: while Kirov et al.’s data [33] exclude transitions from sleep to wake (ie, awakenings), the other authors include them [27,29,77]. The effects on specific NREM sleep activities (ie, delta and sigma power) and micro-structural events (ie, spindles) seem not surprising since they are predicted by the main hypotheses proposed explaining sleep-dependent consolidation processes. According to the “Synaptic Homeostasis Hypothesis” [51,52], consolidation is a by-product of the downscaling process occurs during sleep. The increase in SWA reflects this process of synaptic depotentiation, necessary to renormalize synaptic strength overactivated during previous wake. In contrast, the “Active System Consolidation Hypothesis” [105,106] proposes that sleep strengthens memory representations tagged as relevant. In this case, SWA and, specifically, slow oscillations, are believed to synchronize the activity of hippocampal sharp wave ripples and talamo-cortical spindles, allowing the transfert of reactivated memory traces from hippocampus to neocortex for long-term storage. Although these two hypothesis pointed different role of NREM sleep features, they both predicted their increase during subsequent sleep. The post-training increase in sleep continuity may favour both processes by creating a favourable physiological frame for synaptic downscaling and sleep-dependent memory consolidation to proceed undisturbed. In the same way, spindle activity may exert its beneficial effect on memory processes both in a direct manner [10,39,88] and through its protective role for sleep maintenance [107,108]. Since sleep depth and sleep continuity are considered crucial landmarks for a good sleep [109], a possible therapeutic impact of pre-sleep training should be carefully evaluated in all kinds of sleep disturbances with impaired sleep maintenance, as recently shown in some studies [27,34,79,80]. It is also noteworthy that most studies assessing sleep latency show that it is not increased after cognitive tasks (sometimes even reduced), challenging the commonplace tenet that cognitive activity hampers falling asleep by raising psychophysiological arousal levels [31,36]. It could be the case that not all cognitive tasks actually increase arousal, and that other environmental [110] and psychological factors (ie, trait predisposition [111]) modulate this effect.

5. Concluding remarks and future directions for research

14

The data described above appear very encouraging in terms of their clinical and psychosocial fall-out. However, there are still some relevant issues in need of further investigation. First, the effects of training protocols on populations with sleep complaints have been so far mainly examined with repeated sessions, administered for many days or even weeks (eg, [30,34]). However, previous studies of our groups have also demonstrated short term effects after a single pre-sleep training session [27,80]. Many previous studies have chosen to approach sleep variables in a “static” manner (ie, only considering their global amount in the sleep episode). An alternative approach could be to take into account, for all sleep components: (a) their time-course over the sleep episode (similarly to what was successfully carried in the study of SWS profile for the understanding of process S [2] as well as their reciprocal interplay across the night sleep, coupling, whenever possible, spectral analysis with the different kinds of methods approaching time dynamics (from basic sleep visual inspection to time domain automatic analyses); and (b) their topographic aspects, namely the way structural features are peculiarly modified in different brain areas. A stunning example comes from the impressive data produced on “local” SWA, whose increase - elicited by pre-sleep motor tasks in those Central Nervous System (CNS) areas responsible for their planning, control and execution - is reputed a solid support to the idea of SWA serving cortical plasticity (eg, [53]). Furthermore, the question should be addressed whether wake content manipulation has a beneficial effect on subjective sleep quality: indeed, although sleep depth and continuity have also been proposed as major determinants of the individuals’ perception of a good night’s sleep [112], data on this issue are still very sparse. We believe that the articulate panel of findings reported above may encourage sleep scientists to further conceive a comprehensive model for experience-dependent sleep changes, able to predict how and to what extent sleep will be modified in response to wake intensity modifications. However, researchers have to deal with several theoretical and methodological issues. The most difficult concerns how to “operationalize” and to “measure” cognitive activity, which is constantly going on during life. Cognitive functions, from the simplest to the most complex, are needed for any sensible representation of the world and for any goal-oriented response. In particular, the continuous process of learning is essential to all species for adaptation to the environment and to its changes [113], and is intrinsic in any experience. There is not one minute of the day that we do not learn. Therefore, what is specifically meant by “intensive cognitive activity”? As we have schematically displayed in Figure 1, we propose that these expressions describe the occurrence, over a background “tonic” cognitive activity level; corresponding to the habitual tasks pursued by the subject, of more or less prolonged “phasic” increases, differentiating from this background activity at a degree which is modulated by the amount and the characteristics (eg, novelty, salience, emotionality, difficulty) of the stimuli to be processed.

15

Figure 1. A schematic model displaying how cognitive activity levels would fluctuate during an average 16-hour waking period. In this model, “baseline” cognitive activity is intended as corresponding to the individual’s habitual activities, while “phasic” increases correspond to more demanding or cognitively engaging tasks, in terms of higher amount and different characteristics of the stimuli to be processed. The factors that affect the “magnitude” of these increases are listed in the left box (continuous border), while those modulating their effects on sleep are reported in the right box (dotted border).

For experimental purposes, the question of measuring cognitive load is complex. Learning has been often conceptualized in terms of its outcome (the behavioural changes that follow), which basically implies measuring how the subject’s performance has improved after the learning session. However, as we have extensively discussed in a previous review of ours [114], the result of a memory process, corresponding to what we assess at recall, is notably influenced by a number of modulating factors intervening during acquisition and retention (eg, task difficulty, depth of encoding, basic individual abilities) as well as degree of motivation, intentionality, and awareness. Therefore, it is very likely that in different individuals and/or situations the same behavioral changes could come as the result of very different levels of cognitive engagement (for a paradigmatic example see [88], where learningdependent spindle changes are found only after practising at the task with the maximal encoding difficulty). Therefore, concieving a way to reliably measure cognitive activity is probably a crucial requirement to eventually manipulate it before sleep, in order to set up adequate “dose-response” designs, greatly needed to formulate predictive models. Moreover, the effects of intensive

cognitive training on sleep could be substantially different depending on whether the task administered is more or less “ecological”. On one hand, classical neuropsychological tasks are able to selectively prompt the response of one specific cognitive function but fail in satisfactorily replicating real-life situations and mechanisms. Alternately, the “enriched environment” paradigm and naturalistic tasks are certainly more promising in terms of

16

ecological validity, but the feasibility of their adoption and the reliability of their results are often limited by methodological factors (eg, the impossibility to assess the differential contribution of many cognitive functions simultaneously called into action). In conclusion, we are confident that our synthesis of the available data on sleep changes following cognitive activity may foster sleep scientists to further address the relationship between waking cognition and subsequent sleep from many different perspectives. Future research could contribute to better understand physiological and psychological processes that might modulate sleep and its characteristics and to eventually update sleep regulation models. Following Webb’s “three-factor” model of sleep regulation [115], pre-sleep cognitive activity may represent one of the behaviours that modulates sleep onset, termination and structure, together with circadian and homeostatic processes [2]. A psychosocial approach would also be welcome, given that wake and sleep quality are crucial factors for human well-being and for performance and organization in many applied contexts (eg, schools, sports, shift work). Finally, clinicians will be in charge of further evaluating the beneficial effects of pre-sleep training for all kinds of sleep disturbances with impaired sleep maintenance, and of exploring planned training sessions as an alternative treatment or complement strategy to be introduced in standard behavioural therapies for sleep disordered populations.

REFERENCES [1]

Diekelmann S. Sleep for cognitive enhancement. Front Syst Neurosci 2014;8:1–12. doi:10.3389/fnsys.2014.00046.

[2]

Borbély AA. A two process model of sleep regulation. Hum Neurobiol 1982;1:195–204.

[3]

Feinberg I. Changes in sleep cycle patterns with age. J Psychiatr Res 1974;10:283–306. doi:10.1016/0022-3956(74)90011-9.

[4]

Feinberg I, Campbell IG. Ketamine administration during waking increases delta eeg intensity in rat sleep. Neuropsychopharmacology 1993;9:41–8. doi:10.1038/npp.1993.41.

[5]

Campbell IG, Feinberg I. Noncompetitive NMDA channel blockade during waking intensely stimulates NREM delta. J Pharmacol Exp Ther 1996;276:737–42.

[6]

Campbell IG, Feinberg I. NREM delta stimulation following MK-801 is a response of sleep systems. J Neurophysiol 1996;76:3714–20. doi:10.1152/jn.1996.76.6.3714.

[7]

Franken P. The Quality of Waking and Process S. Sleep 2007;30:126–7.

[8]

Huber R, Tononi G, Cirelli C. Exploratory behavior, cortical BDNF expression, and sleep homeostasis. Sleep 2007;30:129–39. doi:10.1093/sleep/30.2.129.

[9]

Meerlo P, Pragt BJ, Daan S. Social stress induces high intensity sleep in rats. Neurosci Lett 1997;225:41–4. doi:10.1016/S0304-3940(97)00180-8.

[10]

Gais S, Mölle M, Helms K, et al. Learning-dependent increases in sleep spindle density. J Neurosci 2002;22:6830–4. doi:20026697.

17

[11]

Fogel SM, Smith CT, Cote KA. Dissociable learning-dependent changes in REM and nonREM sleep in declarative and procedural memory systems. Behav Brain Res 2007;180:48–61. doi:10.1016/j.bbr.2007.02.037.

[12]

Smith C, Lapp L. Increases in number of REMS and REM density in humans following an intensive learning period. Sleep 1991;14:325–30. doi:10.1093/sleep/14.4.325.

[13]

Smith CT, Nixon MR, Nader RS. Posttraining increases in REM sleep intensity implicate REM sleep in memory processing and provide a biological marker of learning potential. Learn Mem 2004;11:714–9. doi:10.1101/lm.74904.

[14]

Kredlow MA, Capozzoli MC, Hearon BA, et al. The effects of physical activity on sleep: a meta-analytic review. J Behav Med 2015;38:427–49. doi:10.1007/s10865-015-9617-6.

[15]

Peigneux P, Laureys S, Delbeuck X, et al. Sleeping brain, learning brain. The role of sleep for memory systems. Neuroreport 2001;12:A111-24.

[16]

Stephenson R, Lim J, Famina S, et al. Sleep-wake behavior in the rat: ultradian rhythms in a light-dark cycle and continuous bright light. J Biol Rhythms 2012;27:490–501. doi:10.1177/0748730412461247.

[17]

Abou-Ismail UA, Burman OHP, Nicol CJ, et al. The effects of enhancing cage complexity on the behaviour and welfare of laboratory rats. Behav Processes 2010;85:172–80. doi:10.1016/j.beproc.2010.07.002.

[18]

Abou-Ismail UA, Mahboub HD. The effects of enriching laboratory cages using various physical structures on multiple measures of welfare in singly-housed rats. Lab Anim 2011;45:145–53. doi:10.1258/la.2011.010149.

[19]

Ganguly-Fitzgerald I, Donlea J, Shaw PJ. Waking experience affects sleep need in Drosophila. Science 2006;313:1775–81. doi:10.1126/science.1130408.

[20]

Donlea JM, Ramanan N, Silverman N, et al. Genetic Rescue of Functional Senescence in Synaptic and Behavioral Plasticity. Sleep 2014;37:1427–37. doi:10.5665/sleep.3988.

[21]

Lone SR, Potdar S, Srivastava M, et al. Social experience is sufficient to modulate sleep need of drosophila without increasing wakefulness. PLoS One 2016;11. doi:10.1371/journal.pone.0150596.

[22]

Chi MW, Griffith LC, Vecsey CG. Larval population density alters adult sleep in wild-type Drosophila melanogaster but not in Amnesiac mutant flies. Brain Sci 2014;4:453–70. doi:10.3390/brainsci4030453.

[23]

Eban-Rothschild A, Bloch G. The colony environment modulates sleep in honey bee workers. J Exp Biol 2015;218:404–11. doi:10.1242/jeb.110619.

[24]

Ambrosini MV, Mariucci G, Bruschelli G, et al. Sequential hypothesis of sleep function. V. Lengthening of post-trial SS episodes in reminiscent rats. Physiol Behav 1995;58:1043–9.

[25]

van Gool WA, Mirmiran M. Effects of aging and housing in an enriched environment on sleepwake patterns in rats. Sleep 1986;9:335–47.

[26]

Magloire V, Cattarelli M. Delayed changes of sleep duration after rewarded olfactory discrimination learning in the rat. Behav Brain Res 2009;205:568–71. doi:10.1016/j.bbr.2009.08.039.

[27]

Conte F, Carobbi G, Errico BM, et al. The effects of pre-sleep learning on sleep continuity, stability, and organization in elderly individuals. Front Neurol 2012:1–9.

18

doi:10.3389/fneur.2012.00109. [28]

Peters KR, Smith V, Smith CT. Changes in sleep architecture following motor learning depend on initial skill level. J Cogn Neurosci 2007;19:817–29.

[29]

Arzilli C, Cerasuolo M, Conte F, et al. The effect of cognitive activity on sleep maintenance in a subsequent daytime nap. Behav Sleep Med 2018:1–9. doi:10.1080/15402002.2018.1425870.

[30]

Morita E, Imai M, Okawa M, et al. A before and after comparison of the effects of forest walking on the sleep of a community-based sample of people with sleep complaints. Biopsychosoc Med 2011;5:13. doi:10.1186/1751-0759-5-13.

[31]

Higuchi S, Motohashi Y, Liu Y, et al. Effects of playing a computer game using a bright display on presleep physiological variables, sleep latency, slow wave sleep and REM sleep. J Sleep Res 2005;14:267–73. doi:10.1111/j.1365-2869.2005.00463.x.

[32]

Horne JA, Minard A. Sleep and sleepiness following a behaviourally ‘active’ day. Ergonomics 1985;28:567–75. doi:10.1080/00140138508963171.

[33]

Kirov R, Kolev V, Verleger R, et al. Labile sleep promotes awareness of abstract knowledge in a serial reaction time task. Front Psychol 2015;6:1–10. doi:10.3389/fpsyg.2015.01354.

[34]

Haimov I, Shatil E. Cognitive Training Improves Sleep Quality and Cognitive Function among Older Adults with Insomnia. PLoS One 2013;8. doi:10.1371/journal.pone.0061390.

[35]

Dworak M, Schierl T, Bruns T, et al. Impact of singular excessive computer game and television exposure on sleep patterns and memory performance of school-aged children. Pediatrics 2007;120:978–85. doi:10.1542/peds.2007-0476.

[36]

Wuyts J, De Valck E, Vandekerckhove M, et al. The influence of pre-sleep cognitive arousal on sleep onset processes. Int J Psychophysiol 2012;83:8–15. doi:10.1016/j.ijpsycho.2011.09.016.

[37]

Morita Y, Ogawa K, Uchida S. The effect of a daytime 2-hour nap on complex motor skill learning. Sleep Biol Rhythms 2012;10:302–9. doi:10.1111/j.1479-8425.2012.00576.x.

[38]

Peters KR, Ray L, Smith V, et al. Changes in the density of stage 2 sleep spindles following motor learning in young and older adults. J Sleep Res 2008;17:23–33. doi:10.1111/j.13652869.2008.00634.x.

[39]

Fogel SM, Smith CT. Learning-dependent changes in sleep spindles and Stage 2 sleep. J Sleep Res 2006;15:250–5. doi:10.1111/j.1365-2869.2006.00522.x.

[40]

Fogel SM, Ray LB, Binnie L, et al. How to become an expert: A new perspective on the role of sleep in the mastery of procedural skills. Neurobiol Learn Mem 2015;125:236–48. doi:10.1016/j.nlm.2015.10.004.

[41]

Hoedlmoser K, Birklbauer J, Schabus M, et al. The impact of diurnal sleep on the consolidation of a complex gross motor adaptation task. J Sleep Res 2015;24:100–9. doi:10.1111/jsr.12207.

[42]

Smith C, B. Aubrey J, Peters K. Different roles for REM and Stage 2 sleep in motor learning: a proposed model. Psychol Bel 2004;44:79–102.

[43]

Nader RS, Murkar AL, Smith CT. Sleep changes in adolescents following procedural task training. Front Psychol 2016;7:1555. doi:10.3389/fpsyg.2016.01555.

[44]

Schiffelholz T, Aldenhoff JB. Novel object presentation affects sleep-wake behavior in rats. Neurosci Lett 2002;328:41–4.

19

[45]

Mandile P, Vescia S, Montagnese P, et al. Post-trial sleep sequences including transition sleep are involved in avoidance learning of adult rats. Behav Brain Res 2000;112:23–31. doi:10.1016/S0166-4328(00)00158-3.

[46]

Ambrosini MV, Sadile AG, Gironi Carnevale UA, et al. The sequential hypothesis on sleep function. I. Evidence that the structure of sleep depends on the nature of the previous waking experience. Physiol Behav 1988;43:325–37.

[47]

Fogel SM, Smith CT, Beninger RJ. Evidence for 2-stage models of sleep and memory: learning-dependent changes in spindles and theta in rats. Brain Res Bull 2009;79:445–51. doi:10.1016/j.brainresbull.2009.03.002.

[48]

Buchegger J, Meier-Koll A. Motor learning and ultradian sleep cycle: an electroencephalographic study of trampoliners. Percept Mot Skills 1988;67:635–45. doi:10.2466/pms.1988.67.2.635.

[49]

Morin A, Doyon J, Dostie V, et al. Motor sequence learning increases sleep spindles and fast frequencies in post-training sleep. Sleep 2008;31:1149–56. doi:10.1016/j.anifeedsci.2016.03.025.

[50]

Naylor E, Penev PD, Orbeta L, et al. Daily social and physical activity increases slow-wave sleep and daytime neuropsychological performance in the elderly. Sleep 2000;23:87–95.

[51]

Tononi G, Cirelli C. Sleep and synaptic homeostasis: a hypothesis. Brain Res Bull 2003;62:143–50.

[52]

Tononi G, Cirelli C. Sleep function and synaptic homeostasis. Sleep Med Rev 2006;10:49–62. doi:10.1016/j.smrv.2005.05.002.

[53]

Huber R, Ghilardi MF, Massimini M, et al. Local sleep and learning. Nature 2004;430:78–81. doi:10.1038/nature02663.

[54]

Hanlon EC, Faraguna U, Vyazovskiy VV, et al. Effects of skilled training on sleep slow wave activity and cortical gene expression in the rat. Sleep 2009;32:719–29.

[55]

Miyamoto H, Katagiri H, Hensch T. Experience-dependent slow-wave sleep development. Nat Neurosci 2003;6:553–4. doi:10.1038/nn1064.

[56]

Määttä S, Landsness E, Sarasso S, et al. The effects of morning training on night sleep: A behavioral and EEG study. Brain Res Bull 2010;82:118–23. doi:10.1016/j.brainresbull.2010.01.006.

[57]

Wilhelm I, Kurth S, Ringli M, et al. Sleep Slow-Wave Activity reveals developmental changes in experience-dependent plasticity. J Neurosci 2014;34:12568–75. doi:10.1523/JNEUROSCI.0962-14.2014.

[58]

Li Z, Sheth AB, Sheth BR. What drives slow wave activity during early non-REM sleep: Learning during prior wake or effort? PLoS One 2017;12:1–18. doi:10.1371/journal.pone.0185681.

[59]

Tamaki M, Huang T-R, Yotsumoto Y, et al. Enhanced spontaneous oscillations in the supplementary motor area are associated with sleep-dependent offline learning of fingertapping motor-sequence task. J Neurosci 2013;33:13894–902. doi:10.1523/JNEUROSCI.119813.2013.

[60]

Moroni F, Nobili L, Iaria G, et al. Hippocampal slow EEG frequencies during NREM sleep are involved in spatial memory consolidation in humans. Hippocampus 2014;24:1157–68. doi:10.1002/hipo.22299.

20

[61]

Pugin F, Metz AJ, Wolf M, et al. Local Increase of sleep Slow Wave Activity after three weeks of working memory training in children and adolescents. Sleep 2015;38:607–14. doi:10.5665/sleep.4580.

[62]

Smith C. Sleep states, memory processes and synaptic plasticity. Behav Brain Res 1996;78:49– 56.

[63]

Smith C, Rose GM. Posttraining paradoxical sleep in rats is increased after spatial learning in the Morris water maze. Behav Neurosci 1997;111:1197–204. doi:10.1037/07357044.111.6.1197.

[64]

Febinger HY, George A, Priestley J, et al. Effects of housing condition and cage change on characteristics of sleep in mice. J Am Assoc Lab Anim Sci 2014;53:29–37.

[65]

Bramham CR, Maho C, Laroche S. Suppression of long-term potentiation induction during alert wakefulness but not during “enhanced” REM sleep after avoidance learning. Neuroscience 1994;59:501–9. doi:10.1016/0306-4522(94)90172-4.

[66]

Datta S, Bloom F, Sweigart J, et al. Avoidance task training potentiates phasic pontine-wave density in the rat: A mechanism for sleep-dependent plasticity. J Neurosci 2000;20:8607–13.

[67]

Ulloor J, Datta S. Spatio-temporal activation of cyclic AMP response element-binding protein, activity-regulated cytoskeletal-associated protein and brain-derived nerve growth factor: a mechanism for pontine-wave generator activation-dependent two-way active-avoidance memor. J Neurochem 2005;95:418–28. doi:10.1111/j.1471-4159.2005.03378.x.

[68]

Sanford LD, Yang L, Wellman LL, et al. Differential effects of controllable and uncontrollable footshock stress on sleep in mice. Sleep 2010;33:621–30.

[69]

Smith C, Wong PT. Paradoxical sleep increases predict successful learning in a complex operant task. Behav Neurosci 1991;105:282–8.

[70]

Ambrosini MV, Langella M, Gironi Carnevale UA, et al. The sequential hypothesis of sleep function. III. The structure of postacquisition sleep in learning and nonlearning rats. Physiol Behav 1992;51:217–26.

[71]

Buchegger J, Fritsch R, Meier-Koll A, et al. Does trampolining and anaerobic physical fitness affect sleep? Percept Mot Skills 1991;73:243–52. doi:10.2466/pms.1991.73.1.243.

[72]

Mandai O, Guerrien A, Sockeel P, et al. REM sleep modifications following a Morse code learning session in humans. Physiol Behav 1989;46:639–42.

[73]

De Koninck J, Lorrain D, Christ G, et al. Intensive language learning and increases in rapid eye movement sleep: evidence of a performance factor. Int J Psychophysiol 1989;8:43–7.

[74]

Meier-Koll A, Bussmann B, Schmidt C, et al. Walking through a maze alters the architecture of sleep. Percept Mot Skills 1999;88:1141–59. doi:10.2466/pms.1999.88.3c.1141.

[75]

Seeley CJ, Smith CT, MacDonald KJ, et al. Ventromedial prefrontal theta activity during rapid eye movement sleep is associated with improved decision-making on the Iowa Gambling Task. Behav Neurosci 2016;130:271–80. doi:10.1037/bne0000123.

[76]

Montagnese P, Mandile P, Vescia S, et al. Long-term habituation to spatial novelty modifies posttrial synchronized sleep in rats. Brain Res Bull 1993;32:503–8. doi:10.1016/03619230(93)90298-P.

[77]

Mango D, Albinni B, Conte F, et al. The effects of learning monologues on subsequent sleep in actors and non-actors. J Sleep Res 2016;25:116–7.

21

[78]

De Bruin EA, Beersma DGM, Daan S. Sustained mental workload does not affect subsequent sleep intensity. J Sleep Res 2002;11:113–21. doi:10.1046/j.1365-2869.2002.00290.x.

[79]

Sergeeva V, Viczko J, Ray LB, et al. Sleep-dependent motor sequence memory consolidation in individuals with periodic limb movements. Sleep Med 2017;40:23–32. doi:10.1016/j.sleep.2017.09.005.

[80]

Conte F, Carobbi G, Errico BM, et al. Pre-sleep learning improves sleep continuity in young individuals with irregular sleep patterns. J Sleep Res 2010;14:91–2.

[81]

de Almondes KM, Leonardo MEM, Moreira AMS. Effects of a cognitive training program and sleep hygiene for executive functions and sleep quality in healthy elderly. Dement Neuropsychol 2017;11:69–78. doi:10.1590/1980-57642016dn11-010011.

[82]

Benloucif S, Orbeta L, Ortiz R, et al. Morning or evening activity improves neuropsychological performance and subjective sleep quality in older adults. Sleep 2004;27:1542–51.

[83]

Eschenko O, Molle M, Born J, et al. Elevated sleep spindle density after learning or after retrieval in rats. J Neurosci 2006;26:12914–20. doi:10.1523/JNEUROSCI.3175-06.2006.

[84]

Iotchev IB, Kis A, Bódizs R, et al. Author Correction: EEG Transients in the Sigma Range During non-REM Sleep Predict Learning in Dogs. Sci Rep 2018;8:6138. doi:10.1038/s41598018-24096-6.

[85]

Fogel S, Jacob JS. The role of sleep spindles in simple motor procedural learning. Sleep 2002;25:A279–80.

[86]

Barakat M, Doyon J, Debas K, et al. Fast and slow spindle involvement in the consolidation of a new motor sequence. Behav Brain Res 2011;217:117–21. doi:10.1016/j.bbr.2010.10.019.

[87]

Johnson LA, Blakely T, Hermes D, et al. Sleep spindles are locally modulated by training on a brain-computer interface. Proc Natl Acad Sci 2012;109:18583–8. doi:10.1073/pnas.1207532109.

[88]

Schmidt C. Encoding difficulty promotes postlearning changes in sleep spindle activity during napping. J Neurosci 2006;26:8976–82. doi:10.1523/JNEUROSCI.2464-06.2006.

[89]

Ward MP, Peters KR, Smith CT. Effect of emotional and neutral declarative memory consolidation on sleep architecture. Exp Brain Res 2014;232:1525–34. doi:10.1007/s00221013-3781-0.

[90]

Tamaki M, Matsuoka T, Nittono H, et al. Fast sleep spindle (13-15 Hz) activity correlates with sleep-dependent improvement in visuomotor performance. Sleep 2008;31:204–11. doi:10.1093/sleep/31.2.204.

[91]

Tamaki M, Matsuoka T, Nittono H, et al. Activation of fast sleep spindles at the premotor cortex and parietal areas contributes to motor learning: a study using sLORETA. Clin Neurophysiol 2009;120:878–86. doi:10.1016/j.clinph.2009.03.006.

[92]

Bang JW, Khalilzadeh O, Hämäläinen M, et al. Location specific sleep spindle activity in the early visual areas and perceptual learning. Vision Res 2014;99:162–71. doi:10.1016/j.visres.2013.12.014.

[93]

Bergmann TO, Mölle M, Diedrichs J, et al. Sleep spindle-related reactivation of categoryspecific cortical regions after learning face-scene associations. Neuroimage 2012;59:2733–42. doi:10.1016/j.neuroimage.2011.10.036.

[94]

Diekelmann S, Born J. The memory function of sleep. Nat Rev Neurosci 2010;11:114–26.

22

doi:10.1038/nrn2762. [95]

Eschenko O, Ramadan W, Molle M, et al. Sustained increase in hippocampal sharp-wave ripple activity during slow-wave sleep after learning. Learn Mem 2008;15:222–8. doi:10.1101/lm.726008.

[96]

Ramadan W, Eschenko O, Sara SJ. Hippocampal sharp wave/ripples during sleep for consolidation of associative memory. PLoS One 2009;4:e6697. doi:10.1371/journal.pone.0006697.

[97]

Moroni F, Nobili L, Curcio G, et al. Procedural learning and sleep hippocampal low frequencies in humans. Neuroimage 2008;42:911–8. doi:10.1016/j.neuroimage.2008.05.027.

[98]

Heib DPJ, Hoedlmoser K, Anderer P, et al. Slow oscillation amplitudes and up-state lengths relate to memory improvement. PLoS One 2013;8:e82049. doi:10.1371/journal.pone.0082049.

[99]

Mölle M, Eschenko O, Gais S, et al. The influence of learning on sleep slow oscillations and associated spindles and ripples in humans and rats. Eur J Neurosci 2009;29:1071–81. doi:10.1111/j.1460-9568.2009.06654.x.

[100] Mölle M, Bergmann TO, Marshall L, et al. Fast and slow spindles during the sleep slow oscillation: disparate coalescence and engagement in memory processing. Sleep 2011;34:1411– 21. doi:10.5665/SLEEP.1290. [101] Ruch S, Markes O, Duss SB, et al. Sleep stage II contributes to the consolidation of declarative memories. Neuropsychologia 2012;50:2389–96. doi:10.1016/j.neuropsychologia.2012.06.008. [102] Ferri R, Huber R, Aricò D, et al. The slow-wave components of the cyclic alternating pattern (CAP) have a role in sleep-related learning processes. Neurosci Lett 2008;432:228–31. doi:10.1016/j.neulet.2007.12.025. [103] Tartar JL, Ward CP, McKenna JT, et al. Hippocampal synaptic plasticity and spatial learning are impaired in a rat model of sleep fragmentation. Eur J Neurosci 2006;23:2739–48. doi:10.1111/j.1460-9568.2006.04808.x. [104] Rolls A, Colas D, Adamantidis A, et al. Optogenetic disruption of sleep continuity impairs memory consolidation. Proc Natl Acad Sci U S A 2011;108:13305–10. doi:10.1073/pnas.1015633108. [105] McClelland JL, McNaughton BL, O’Reilly RC. Why there are complementary learning systems in the hippocampus and neocortex: insights from the successes and failures of connectionist models of learning and memory. Psychol Rev 1995;102:419–57. [106] Marshall L, Born J. The contribution of sleep to hippocampus-dependent memory consolidation. Trends Cogn Sci 2007;11:442–50. doi:10.1016/j.tics.2007.09.001. [107] Dang-Vu TT, McKinney SM, Buxton OM, et al. Spontaneous brain rhythms predict sleep stability in the face of noise. Curr Biol 2010;20:R626-7. doi:10.1016/j.cub.2010.06.032. [108] Dang-Vu TT, Bonjean M, Schabus M, et al. Interplay between spontaneous and induced brain activity during human non-rapid eye movement sleep. Proc Natl Acad Sci U S A 2011;108:15438–43. doi:10.1073/pnas.1112503108. [109] Bonnet MH, Arand DL. Clinical effects of sleep fragmentation versus sleep deprivation. Sleep Med Rev 2003;7:297–310. [110] Buechner VL, Maier MA. Not always a matter of context: direct effects of red on arousal but context-dependent moderations on valence. Peer J 2016;4:e2515. doi:10.7717/peerj.2515.

23

[111] Palagini L, Ong JC, Riemann D. The mediating role of sleep-related metacognitive processes in trait and pre-sleep state hyperarousal in insomnia disorder. J Psychosom Res 2017;99:59–65. doi:10.1016/j.jpsychores.2017.03.001. [112] Keklund G, Akerstedt T. Objective components of individual differences in subjective sleep quality. J Sleep Res 1997;6:217–20. [113] Mery F. Natural variation in learning and memory. Curr Opin Neurobiol 2013;23:52–6. doi:10.1016/j.conb.2012.09.001. [114] Conte F, Ficca G. Caveats on psychological models of sleep and memory: a compass in an overgrown scenario. Sleep Med Rev 2013;17:105–21. doi:10.1016/j.smrv.2012.04.001. [115] Webb WB. An objective behavioral model of sleep. Sleep 1988;11:488–96.

24

HIGHLIGHTS - Wake affects subsequent sleep not only as a function of its duration, but also of its intensity. - Beyond Slow Wave Activity, other important sleep variables appear to go through experience-dependent changes. - Further research is needed to adequately operationalize and measure cognitive activity.