Effects of a Daytime Nap on Primed and Repeated Remote Associates Tests and Relations with Divergent Creativity

Abstract

The effect of napping versus wakefulness was studied on primed and repeated Remote Associate Tests (RATs) and on divergent creativity tests. The participants were 42 students from the USA, studying international courses at a Swedish university. The hypotheses for the RATs were (1), when the correct answers were primed before the nap, the RAT should be solved better for those who entered REM sleep, compared to those with no REM sleep or a resting condition; and (2), when retested the RAT should be solved better after a nap than after rest. For the creativity tests, hypothesis (3) was that creativity should be higher after the nap than after rest. Hypothesis 1 and hypothesis 3 were not supported. Hypothesis 2 was supported in an ANOVA. The REM group improved more than the rest group on the repeated RAT. Also, the No-REM and rest groups differed, strengthening the importance of both REM and No-REM sleep for creative problem-solving.

The importance of sleep for higher mental functions gathers increasing scientific support. Many studies indicate the importance of sleep for memory processing and plasticity (as reviewed by Rasch & Born, 2013). Specifically, regarding problem-solving, studies have found that sleep facilitated insight (Wagner, Gais, Haider, Verleger, & Born, 2004; Yordanova et al., 2008). Walker, Liston, Hobson, and Stickgold (2002) found that people woken up from REM were better at solving insight problems than people woken up from non-REM sleep. Sleep has also been shown to promote structural generalization between tasks with low surface similarity (Monaghan et al., 2015).

According to Hobson and Pace-Schott (2002), different forms of learning may benefit from different stages of sleep. Creative cognition has long been associated with sleep and particularly dreaming but it is very probable that the relations are complex, and research so far is sparse. Also, theories diverge concerning the functional significance of REM sleep; for instance, Pagel (2008) suggested that one function of REM sleep is to bring about a high level of associative brain activity, less easily accomplished when awake, while Kramer (2007) viewed the REM stage to function as an emotion regulator.

An alternative to studies of nightly sleep, that has proven useful for certain research questions, is to use naps during the day. For instance, the restorative function of naps has been investigated (Mednick et al., 2002). Carr and Nielsen (2015), using naps, showed an advantage of REM over No-REM on a word association test for words that were primed before sleep, but not for unprimed words.

In a pioneer study, Cai, Mednick, Harrison, Kanady, and Mednick (2009) investigated the effects of a nap on implicit priming for creative problem-solving in the Remote Associates Test (Mednick, 1962). In this study, the participants were first primed in the morning, followed by either a nap or rest p.m. and a subsequent test period after the nap or rest. The primes consisted of the correct solutions (words) on certain items in an analogies task. These words were also the correct solutions on a RAT taken after the nap/rest (hereafter referred to as primed RAT).

Cai et al. (2009) found that the primed RAT was solved better by those who entered REM sleep during their nap, than either those who did not get into REM sleep or those who had been awake and resting during an equivalent time. Moreover, in another sub-study, Cai et al. investigated if a nap would improve results on a RAT that was taken first in the morning and then once again after the nap. In that study, with no implicit primes, the retested RAT did not differ between nap or rest. Thus, the nap did not improve the general ability to solve the RAT.

Cai et al. (2009) controlled for memory and concluded that although all groups had similar memory for the implicit primes, only REM sleep promoted the integration of unassociated information. They interpreted the results such that REM sleep provides an opportunity for the brain to process new information (i.e. the implicit primes), which in turn would make it easier to consciously activate that information afterwards.

Whitehurst, Cellini, McDevitt, Duggan, and Mednick (2016) registered polysomnography and also a measure of heart rate variability, and used the same nap paradigm as in Cai et al. (2009). They found, similarly to Cai et al., that those who entered REM sleep improved on a primed RAT. Moreover, they found that when analyzing the heart rate variability measure together with the polysomnography, this improved the prediction for the primed RAT. Interestingly, they also found that the nap together with heart rate variability predicted performance also on a repeated RAT. It seems that individual differences occur during REM sleep, suggesting that increased vagal activity may contribute to “an optimized internal environment, specific to REM sleep, that promotes creative associative memory processing” (p. 7276).

Sio, Monaghan, and Ormerod (2013) used a different remote associate task, developed by Bowden and Jung-Beeman (2003) and analyzed the task split on easy and difficult items. They found that a whole night of sleep facilitated the solving of difficult items, but no difference between conditions sleep or rest for easy items, possibly due to the beneficial effects of sleep on activation spreading.

Landmann et al. (2016) used a similar task as the RAT (the Compound remote associate task) in a design with either an entire night of sleep, an entire night of sleep deprivation, or daytime rest. They found that the percentage of resolved items was higher after sleep than after sleep deprivation, but not higher than after daytime rest. They also found that the time it took to solve previously solved items was shorter in the sleep group than the rest groups. The study did not find any difference in the number of newly solved (i.e., previously unsolved) items, when the task was repeated.

In the present study daytime, napping or resting wakefulness was used to investigate effects on the Remote Associate Test. Further, it was studied if a nap would be beneficial for divergent creativity which was not controlled for in Cai et al. (2009), nor in Whitehurst et al. (2016). The RAT has been considered to measure so-called convergent creativity, because the task is to find a correct answer. It has been shown to depend also on verbal and reasoning skills (Mendelsohn, 1976). Therefore, tests of divergent creativity were included in the present study.

Considering creativity, an important ability is to make new associations between previously unrelated concepts. Among these new associations, ideas emerge that are creative, i.e. they are both original and adaptive (Martindale, 1989). To be able to be both original and adaptive implies that the creative individual has flexible access to different experiential levels, including both subjective mental representations (for example imagery) and outer-directed reality-oriented perception (Smith & Carlsson, 1990a). Such an open communication channel between subjective and objective experiential levels has been reflected in research showing relationships between creativity and the experience of dreaming (Carlsson & Neuman, 2008; Smith & Carlsson, 1990a; see also a review by Ruby, 2011). Naturally, backsides exist with openness, such as possible sleep disturbances (Healey & Runco, 2006) and nearness to psychic distress and disorders (Carson, 2011).

Relations between creativity and the sleep stages were studied by Drago et al. (2011), by collecting polysomnography data for eight participants during whole nights and tests of creativity on two mornings after sleep. They found positive relations between fluency and flexibility and the early sleep stage 1, as well as positive relations between originality and figural creativity and Stage 4 sleep. Especially, they found that a synchronized frontal alpha pattern (CAPs, type 1) in Stage 4 sleep was associated with originality. Interestingly, they found a significant negative correlation between REM sleep and originality. The authors concluded that deep sleep is associated with low cortical arousal, and that low cortical arousal may enhance the ability to associate in divergent and innovative ways. Also, as related by Drago et al. (2011), the synchronized EEG pattern CAP Type 1 has earlier been related to frontal activity, which in turn is important for divergent thinking. The authors cautioned that their study needs replication with more participants and better statistical controls. They also suggested that both deep sleep and REM sleep could be involved in the full creative process.

Based on the above-reviewed research, three hypotheses were formulated.

Hypothesis 1.

It was expected that improvement on a primed RAT (compared to a baseline RAT serving as a control) should be higher for those who entered REM sleep, than for those who did not get into REM sleep, or for a group that stayed awake and resting. This would imply a specific role for REM sleep regarding processing of primed information, in line with Cai et al. (2009).

Hypothesis 2.

It was expected that improvement on a retested RAT (the difference between a baseline RAT taken in the morning and retested after nap or rest should be improved after a nap compared to rest, considering the results of Whitehurst et al. (2016).

Hypothesis 3.

It was expected that participants should show higher divergent creativity after a nap compared to after rest, in line with the assumption by Drago et al. (2011), that sleep as a whole would be beneficial for creativity.

Method

Participants

Participants were recruited from international master courses at Lund University. Since the same RATs were used as in Cai et al., only participants with English as their first language were recruited. On beforehand everyone got information that the project studied relations between sleep and various cognitive and personality variables. They were required not to drink any alcohol or caffeine during 24 h before testing. They declared they did not take anti-depressives, sleep-, or high blood pressure medicine, and that they had slept no less than 6 h on average per night during the preceding week. Participation was anonymous and voluntary with written informed consent. Debriefing took place immediately after the last test. They received lunch and 500 Swedish Crowns (about 55 Euros at that time). The study was approved by the regional ethics committee (2011/98).

The total sample’s mean age was 23.9 years (SD = 3.78). They came both from the USA and from other English-speaking countries. In preliminary analyses, age and gender were found unimportant for the present hypotheses and not further considered. Considering that the RAT was constructed in the USA, a comparison was made between the US participants (n = 42) and those from other English-speaking countries (mainly Canada, the U.K., Ireland, Australia, and South Africa; n = 23). This control showed that the people from the USA did significantly better on the primed RAT (p < .001), as preliminary reported in Carlsson et al. (2014). The decision was made to proceed solely with the US group for the final analyses.

Instruments

The analogies test with implicit primes

The Analogies test encompassed 60 items. It had no time limit. The items were the same as in Cai et al. (2009). One example: “Fast is to slow as hard is to …” The correct answers on 15 of these items served as implicit primes for the primed RAT (i.e., these 15 words were also the correct answers on the primed RAT. Because a few participants had not solved one or two primes correctly, information about the correct words was given matter-of-factly during the lunch break.

The remote associates test

Each RAT item consists of three given words, and the task is to find a correct fourth word connected to these (Mednick, 1962). The three words are unrelated to one another, thus the test taps the ability to associate in different directions and to find a common nominator. Two examples are “Worm, Red, Ticker (…)”, and “Bug, Drill, Hell (…)”. Two parallel RATs were used. The time limit was 15 min on each. Each form comprised 15 items (following Cai et al., 2009). The baseline RAT was tested in the morning and retested after nap or rest. The primed RAT was tested after nap or rest.

Dependent variable for hypothesis 1

To control for individual differences on the RAT, the dependent variable for hypothesis 1 was the raw score difference between the baseline RAT and the primed RAT.

Dependent variable for hypothesis 2

The raw score difference between the baseline RAT and when it was retested after nap or rest was used for testing hypothesis 2.

Divergent creativity tests

Two tests of divergent creativity were used, one with verbal stimuli and written output and the other with visual stimuli and oral output.

The alternate uses test (AUT)

The AUT taps divergent thinking ability, an essential aspect of creativity (Guilford, 1968). The participant was equipped with paper and pencil. The instruction was to write as many uses that they could think of, for two objects separately (a brick and an empty beer-can), and that there was a time limit; the limit was not specified to the participant. They had 10 min before they were stopped. Fluency (total number of uses) and flexibility (number of categories) were scored. Because they were significantly correlated (Rho = .25, p = .04), flexibility was chosen as the more pertinent variable.

AUT flexibility

Flexibility in the AUT was operationalized as the number of different categories that the answers can be sorted into. Two independent judges (the first and the second author) did the categorizing and scored the protocols. Both were blind for condition and all results. Some differences were discussed regarding which categories to use. The protocols were then rescored independently until consensus was reached. The raw scores for both objects were summarized.

The creative functioning test (CFT)

The CFT measures creativity during a perceptual process when a digitalized picture is repeatedly shown (Smith & Carlsson, 1990b). Many instances get registered during this process, as the participant after each exposure reports what he or she sees on a TV screen. Starting with very short exposure time, when nothing can be perceived, the exposure times get prolonged up to stimulus adequate reporting and then get shortened, back to the briefest exposure. Creative functioning is operationalized as a proneness to shift cognitive mindset during the part when the exposures get shortened. Here, the creative person reports subjective, stimulus-distant, perceptions.

Stimulus and instruction

The picture is an artist’s black and white still-life of two common objects. The instruction is that pictures will be shown very quickly and that the participant describes what they think they see after each exposure.

Test procedure

The exposure time starts at .017 milliseconds and is prolonged in an outdrawn fashion, enabling visual subjective representations. Two examples are: “now it looked like a tree and some bushes in the background”, and “now I saw a large face on the screen”. The prolongation continues up to the level when the reporting gets stimulus-near, as detailed in the manual (Smith & Carlsson, 1990b). The exposures then continue, but times are shortened in the same way as when prolonged.

CFT creative functioning

The stimulus-near perception on long exposure times is exerting a considerable influence. Therefore, a crucial question is if the person behaves in a stimulus-bound way. This individual would (implicitly) regard earlier subjective perceptions as mistaken, or “wrong answers”, and would during shortened exposures not change mindset, but stay with the objective reality of “facts”. A highly creative person, on the other hand, is able to report in a stimulus-near way at sufficiently long exposures. But when the ambiguity increases again as times get shorter, he or she is apt to diverge from the veridical, reality-oriented perception. This individual seems open for and to assign higher priority to the subjective imaginary world (or, differently formulated, does not inhibit recent subjective interpretations from conscious awareness).

Rating uses an ordinal scale with six steps. Very high (step 6) implies reporting altogether different than the picture (ex., “now it was only that large tree and bushes again, as I saw in the beginning”). Very low (step 1) implies that there is no change at all in the reports (“It is the same bottle and bowl all the time but most of the picture has disappeared now”). Intermediate steps imply more or less stimulus-near descriptions.

An experienced test leader (the first author) did the CFT testing. Two independent judges (the first and the second author), scored the protocols. The inexperienced judge first trained by scoring several unrelated protocols. Both judges were blind to condition and all other results. The judges had slight differences in certain protocols. These protocols were re-scored independently and then discussed again until consensus was reached.

Previous studies have shown high correlations with the richness of ideas, expressiveness, originality, creative interests, and predictions of creative achievement, all judged by external experts, in studies of researchers, artists, children, and architects (Schoon, 1992; Smith & Carlsson, 1990a). There were negative correlations with depressive illness and grave anxiety (Smith & Carlsson, 1988) and positive relations to the experience of aging (Smith & van der Meer, 1990), to a moderate anxiety level (Carlsson, 2002), to androgyny (Jönsson & Carlsson, 2000), to (a lack of) alexithymia (Smith & van der Meer, 1994), and to balanced hemispheric activity in the brain (Carlsson, Wendt, & Risberg, 2000). The CFT median varies between groups and was particularly low for children in their first school year (Smith & Carlsson, 1983).

Polysomnography

Sleep was recorded with a portable EMBLA apparatus and analyzed with the Somnologica software. For the sleep EEG, two electrodes (Blue sensor) were each put on C3 and C4, each referenced to the contralateral mastoid; as well as a ground (on the middle of the forehead). The eye movements (EOG) were measured by four electrodes placed on EOG1 (right eye laterally), EOG2 (above left eyebrow laterally), EOG3 (below left eye laterally), and EOG4 (above right eyebrow laterally).

Experienced staff at a university hospital sleep laboratory analyzed the polysomnography data. The authors did not check this until all tests were scored.

Procedure

Morning session

The participant got information on arrival if they would belong to the napping or the resting group. The testing began at 10.30 with a sleepiness scale (Hoddes, Zarcone, Smythe, Phillips, & Dement, 1973). This scale did not differ between groups or affect any outcome variables and was not considered further. Then followed the baseline RAT and the analogies test. The participant then had lunch in a quiet room for ½ hour, followed by nap or rest.

Napping and resting sessions

Napping group

The polysomnography apparatus was attached by sleep laboratory staff from the university hospital. The period of sleep could last up to 120 min. The room was an ordinary office room in a silent environment and equipped as a bedroom. The participant was required to stay in the bed at least 90 min, even if waking up earlier. They were checked on after 90 min.

Resting group

The rest lasted 90 min. The participant sat in an armchair, listened to instrumental music, and was checked on for staying awake a few times by the test leader.

Afternoon session

After a break, the participant was first re-tested on the baseline RAT. Initially, a short break (15 min.) was used in order to exclude wake time for the nap group, when they could possibly reflect on the baseline RAT. After a first data collection, with null-results, more participants were added who got a long break (2 h) as in Cai et al. (2009). It was reasoned that a longer break for some reason might be important. As reported in Carlsson et al. (2014), ANOVAs (REM, No-REM, and Rest) showed no differences between a long or short brake however, neither on the primed, nor on the retested RAT.

Next followed the sleepiness scale, the primed RAT, and anxiety and depression questionnaires (State and Trait Anxiety, Forms Y, Spielberger, 1983, and the Center for Epidemiologic Studies Depression Scale; Radloff, 1977). The sleepiness and state anxiety scales were not related to any outcome variable and not further considered.

Last came the divergent creativity tests, preceded by a short break. (The final test was of perceptual defensive strategies, preliminarily reported in Carlsson, Davidson, & Samuelsson, 2016).

Results

Descriptive statistics

The analysis of not normally distributed variables (hypothesis 3) used nonparametric tests.

The descriptive results for the variables total sleep time, range of total sleep time, slow wave sleep and REM sleep are shown in Table 1.

TABLE 1 Total Sleep Time (TST), Range of TST, Slow Wave Sleep and REM Sleep; Means (Minutes) and Standard Deviations

	Sleep (n= 28)	REM (n= 20)	No-REM (n= 8)
Total Sleep Time	80.66	91.83	52.56
SD	28.37	17.24	32.25
Range of TST	8–116.5	61–116.5	8–92
SWS	14.01	15.49	10.50
SD	12.47	11.15	15.58
REM	10.90	15.20	.00
SD	10.77	9.76	.00

Overall means (SD) for trait anxiety and depression were 39.8 (9.64) versus 33.1 (7.30). They were significantly correlated; r = .67, p= .001). Regarding the RAT variables, ANOVAs (REM, No-REM, Rest) were insignificant, both for trait anxiety and depression.

Regarding divergent creativity, trait anxiety and depression were both negatively correlated to the CFT (rho = −.42, p= .055, versus rho = −.38, p = .012).

Descriptive RAT data are shown in Table 2.

TABLE 2 Baseline RAT, Baseline RAT Improvement, Primed RAT, and Primed RAT Improvement; Means and Standard Deviations

	Total (n= 42)	REM (n= 20)	No-REM (n= 8)	Rest (n= 14)
Baseline RAT	5.07	5.15	5.00	5.00
SD	1.63	1.95	1.31	1.36
Improvement	1.40	2.20	1.25	0.36
SD	1.80	2.12	1.28	0.75
Primed RAT	6.64	7.10	6.38	6.14
SD	2.26	2.53	1.85	2.07
Improvement	1.57	1.95	1.38	1.14
SD	2.19	2.54	1.92	1.79

The overall mean on AUT flexibility was 13.00 (SD = 2.78). The median on CFT was 2. The AUT and the CFT were not significantly related (rho = .014, p= .93). The AUT was not significantly related to either the baseline or the primed RAT (r = .13, p = .40, versus r = .03, p = .87). The CFT was not significantly related to either the baseline or the primed RAT (rho = - .21, p = .19, and rho = .08, p = .60).

Notable was that the CFT distribution was skewed. The percentage on the three highest levels was 14.4% and on the two lowest levels 57.2%.

Analysis for hypothesis 1

For the primed RAT improvement, there was no significant difference on an ANOVA (REM, No-REM, Rest); F (2,39) = 0.39. p = .56.

Analyses for hypothesis 2

The ANOVA (REM, No-REM, Rest), for retested RAT improvement was significant (F (2,39)= 5.29, p = .009, partial eta squared = .21). Follow-up tests showed that the REM group was better than the Rest group (t = 3.59, p = .001, Cohen’s d = 1.16), but not significantly better than the No-REM group. Furthermore, the No-REM group improved significantly more than the Rest group (t = 2.08, p = .05, Cohen’s d = 0.85).

Analyses for hypothesis 3

There was no significant difference between the sleep or rest groups on neither of the creativity tests.

Discussion

This study looked at the effects of a nap on the Remote Associate Test. It was studied both when the RAT was primed before the nap and when the RAT was retested after the nap. It was also investigated if a nap could be beneficial for divergent creativity.

The first hypothesis predicted that the REM group should improve more on the primed RAT than the other groups. However, this was not found; there are several possible reasons why this study did not find the result as in Cai et al. (2009).

It is possible that the participants in the Cai et al. study were higher on creative cognition because they were in familiar circumstances, as opposed to the participants in the present study. It might also be the case that those participants had even more culture-specific knowledge, relevant for the primed RAT, than the present participants.

The second hypothesis predicted improvement on the RAT taken in the morning when retested after the nap. This was shown to be the case, as those who slept improved more than those who stayed awake. Both the no-REM group and the REM group improved more than the rest group. This is in line with Whitehurst et al. (2016).

Thus, the napping made a significant difference for the improvement of the retested RAT. Contrarily, the group that stayed awake and actually had a possibility to reflect upon the baseline RAT, did significantly worse than those who slept. The time spent awake may possibly have caused cognitive interference to a greater degree than in the sleep group. However, the Non-REM group spent more time awake but nevertheless fared better than the rest group. Neither did a longer break before retesting make any difference compared to a short one. Thus, the indications point towards a functional role for the nap.

Speculatively, it may have been the case that sleep helped processing the unsolved RAT items, by facilitating access to the long-term memory. Lewis and Durrant (2011) interestingly propose that, “the reactivation of newly learned memories during sleep could actively underpin both schema formation and the addition of new knowledge to existing schemata.” (p. 343). In their “memory replay” model, sleep can facilitate different recombinations and the selective strengthening of schemata, forming the basis of, for instance, insight. The nap may also have helped restore cognitive flexibility so that these participants solved the retested RAT better than those who stayed awake. Earlier research has found that REM sleep especially is amenable for cognitive flexibility (Walker et al., 2002). This line of reasoning seems to get support by results in another study from the present project, preliminarily reported in Carlsson et al. (2016). Here, those who entered REM sleep, compared to those who stayed awake, were significantly more flexible on a measure of perceptual defense strategies tested after the nap. This kind of flexible perceptual defense has previously been related to creativity (Carlsson, 2002).

Hypothesis 3 was not supported. There were no significant differences between the conditions on the creativity variables. It is possible that a nap is not enough to increase divergent creativity. In future investigations it is important to measure sleep during entire nights, to get a full picture of the sleep architecture. In addition, an improvement would be to screen for creativity. The present sample had a low rate of creative functioning overall, considering that well over half the group got scores on the second lowest or lowest level, whereas only 14 per cent reached a very high or a high CFT score.

Reflecting on this modest result on creative functioning, it may have been the case that the participants’ circumstances at the time were less conducive for being creative. They had recently come to stay for one to two years on another continent and were facing new issues in advanced university courses. The process of adapting and learning many new things may have led them to hold a strong focus on the external reality, rather than towards subjective and associative cognition. Speculatively, it may have been the case that a cognitive focus on objective reality to some extent inhibited conscious access to the implicit primes. Although indirectly, of some support for this is a study reported in Smith and Carlsson (1990a). That study used a design with different instructions plus a placebo pill, aiming to get the participant either more alert and outwardly directed, or relaxed and inwardly directed, before doing the CFT. The result showed much higher scores in the relaxation group.

Further, the present participants’ efforts to adapt to new circumstances may have caused a negative mood. This seems to have played a part, as indicated by the negative correlations found between creative functioning and anxiety and depression. The very modest creative functioning in the present participants seems to have an analog in the previous research on children’s creativity (Smith & Carlsson, 1983). Here, a large share of children tested during their first school year scored at the low end-pole in the CFT, in stark contrast to the age groups both below and above. Smith and Carlsson reasoned that maturational as well as adaptive challenges may play a part in the ”low tide”. Regarding the present sample of young academics, one would expect their creative functioning to improve in other circumstances.

No significant correlations were found between the two different creativity measures. This is in line with earlier research reporting non-significant relations between the CFT and the AUT and a creative activity questionnaire (Hoff & Carlsson, 2002). Neither did the creativity measures correlate significantly with the RATs. Since the RAT is said to measure convergent creativity (Akbari Chermahini & Hommel, 2012), to measure insight, and to have moderate correlations with logical thinking and verbal skills (Runco, 2007; Shen et al., 2016), only low correlations were expected. In the present study the general creativity level, the sample size, or both, may have been insufficient.

As related above, preliminary analyses in the present project showed that the US participants got significantly higher scores on the primed RAT than English-speaking people from other countries. In retrospect, it is quite possible that cultural factors played a part, since the RAT was constructed some decades ago in the USA, and builds on associations that are a part of this culture, but not necessarily of the culture in other English-speaking countries. This is a limitation in the present study since it reduced power. However, it seems that no previous study has investigated such a cultural difference in the RAT.

When comparing the results in Cai et al. (2009) with the present results, it can be seen that their REM group and the present REM group got very similar values for time spent in REM sleep, and thus were comparable in that respect. The results on the RAT improvement variables were unfortunately not possible to compare with Cai et al., given that the present study used the difference between the raw scores, in order not to risk inflated percentages when calculating small numbers.

Another limitation in the present study is that the number of participants got quite small in certain sub-groups. A further, design, limitation is the use of two different lengths of the break after sleep before the afternoon testing. Initially, 15 min was used, in order not to include extra awake time for the sleep group, which they might use for conscious reflection. But with null-results, a speculation was that it may be necessary for some reason with a longer break for the primed RAT. However, as described above, no difference was found between the long and the short break.

In conclusion, daytime napping improved the retested RAT both for those that had REM sleep and for those who did not get into REM sleep, which is in line with previous research. However, primes given before sleep did not help the problem solving after the nap, in contrast to Cai et al. (2009). It is possible that both contextual and individual factors made the present participants less able to get conscious access to the primes, even though they got the opportunity to sleep on it.

Acknowledgments

This work has been supported by the Bial Foundation, Portugal, Grant 102/10. Preliminary data from the project were reported in a final report to the Bial foundation (Carlsson et al., 2014).

Disclosure statement

No potential conflict of interest was reported by the authors.

Additional information

Funding

This work was supported by the Fundação Bial [102/10].