Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2019 Nov 1.
Published in final edited form as: Neurobiol Learn Mem. 2018 Aug 3;155:208–215. doi: 10.1016/j.nlm.2018.08.002

The role of sleep in emotional memory processing in middle age

Bethany J Jones 1, Alix Mackay 1,2, Janna Mantua 1, Kurt S Schultz 1, Rebecca M C Spencer 1,3
PMCID: PMC6360141  NIHMSID: NIHMS1503472  PMID: 30081153

Abstract

Sleep benefits memory in young adults, and this effect may be particularly strong for representations associated with negative emotion. Many aspects of sleep important for memory consolidation change with aging, particularly by middle age, suggesting that sleep-related consolidation may be reduced. However, the influence of sleep on memory has rarely been investigated in a middle-aged population. In the current study, young and middle-aged adults viewed negative and neutral pictures and underwent a recognition test after sleep or wake. Subjective emotional reactivity was also measured. Compared to waking, sleep benefited memory in young adults. Performance did not differ between sleep and wake groups in middle-aged adults, and it matched the level of young adults who slept. The effect of sleep versus wake was not influenced by memory valence in either age group. These results suggest the relative influence of sleep compared to wake on memory declines with aging, specifically by middle age, and that this decline extends to negative memory.

Keywords: memory, aging, sleep, middle age, emotional memory, consolidation

1. Introduction

Sleep consolidates memory, reinforcing recently formed representations and integrating them into existing knowledge networks. Memory retention is greater after periods containing sleep than equivalent time spent awake, and specific sleep properties, such as slow wave activity and sleep spindles, have been linked to memory enhancement, supporting an active role of sleep (Born and Wilhelm, 2012; Buzsaki, 1996; Rasch and Born, 2013; Stickgold, 2005).

Memory consolidation does not extend to all daily experiences but depends on factors such as emotional salience (LaBar and Cabeza, 2006). Memory for emotional content, especially when negative in valence, is particularly strong and often selectively benefitted by sleep in young adults (Nishida et al., 2009; Payne et al., 2008; Wagner et al., 2006). Information associated with negative emotions may be particularly advantaged in memory due to its potential relevance for survival. Arousal during encoding may “tag” negative representations for subsequent processing during sleep (Bennion et al., 2015; Cunningham et al., 2014; Richter-Levin and Akirav, 2003). The emotional response, or reactivity, associated with memories may also be processed during sleep, particularly REM sleep (Genzel et al., 2015; Walker and van der Helm, 2008).

Sleep changes with aging. Sleep duration and quality decline, and the proportion of time spent in light sleep stages increases while that spent in slow wave and REM stages decreases (Ohayon et al., 2004). Such changes may contribute to age-related memory decline, though sleep-related memory processing appears to be preserved in older adults in some cases (Aly and Moscovitch, 2010; Jones et al., 2016; Wilson et al., 2012). Many changes in sleep and memory are already apparent in middle age (~40-60 years; Carrier et al., 2001, 2011; Martin et al., 2013). Sleep deficits in middle age may not only underlie memory impairment contemporarily but also set the stage for cognitive impairment later in life (Musiek and Holtzman, 2016). Thus, determining when and under what circumstances the memory function of sleep begins to falter may help identify the optimal window to intervene to ameliorate age-related memory decline.

Given that negative emotionality enhances the memory benefit of sleep in young adults, one might expect sleep-related processing of negative memories to be relatively spared with aging. However, recent evidence suggests sleep may not influence negative memories or associated reactivity in older adults 55-80 years of age (though notably a sleep benefit for positive memories was observed; Jones et al., 2016). Whether this change in sleep function occurs earlier during the lifespan in middle age is unknown.

Thus, the objective of the current study was to determine the influence of sleep on negative memories and reactivity in middle-aged adults. Young adults were included for comparison. Given age-related changes in sleep, we hypothesized that the effect of sleep compared to wake on negative memories and reactivity would be reduced or absent in middle-aged adults compared to young adults.

2. Materials and methods

2.1. Participants

Forty young adults between 18 and 30 years of age (sleep group: n=20, M=19.88, SD=1.26, 12 females; wake group: n=20, M=20.0, SD=1.17, 13 females) and 41 middleaged adults between 35 and 50 years of age (sleep group: n=21, M=45.65, SD=3.20, 15 females; wake group: n=20, M=44.0, SD=5.04, 12 females) participated in this study. Participants had normal or corrected-to-normal vision and no history of neurological disease, sleep disorders, head injury, or use of medications known to affect sleep or cognitive function. All participants were compensated with payment or course credit. Experimental procedures were approved by the University of Massachusetts, Amherst Institutional Review Board, and written informed consent was obtained before the experiment.

2.2. Materials

Ninety emotionally negative and 90 emotionally neutral pictures were used in this experiment. The majority of stimuli (110) were obtained from the International Affective Picture System (IAPS; Lang et al., 2005). The rest were from an in-house set and were chosen to match the IAPS pictures in content and emotionality (Baran et al., 2012). Based on normative data and previous studies in our lab (Jones et al., 2016), negative pictures were moderate to high in arousal, and neutral pictures were low in arousal.

2.3. Procedure

Procedures used in this study have been described previously (Jones et al., 2016). The experiment consisted of two sessions (Fig. 1A). The first session (Encoding) took place either in the evening between 20:00 and 22:00 (sleep groups) or in the morning between 8:00 and 10:00 (wake groups), followed by the second session (Recognition) 12 hours later. The wake groups were asked not to nap and to limit caffeine between sessions, and all participants were asked not to consume alcoholic beverages. In the first session, participants completed the Pittsburg Sleep Quality Index (Buysse et al., 1989), which measures subjective habitual sleep quality. Additionally, participants completed a sleep/wake diary and the Stanford Sleepiness Scale (SSS) at both sessions.

Figure 1.

Figure 1.

Open in a new tab

Experimental design and procedure. A, Encoding took place either in the evening (sleep group) or the morning (wake group), followed by Recognition 12 hr later. The sleep group underwent PSG overnight. B, During Encoding participants viewed 60 pictures (targets) and rated the valence and arousal of each on 9-point Likert-type scales. During Recognition, participants viewed 180 pictures, a mixture of target and novel foil pictures, and rated each one on valence and arousal. Participants indicated whether or not they recognized the picture by responding yes/no.

During Encoding, participants viewed 60 target stimuli - 30 negative and 30 neutral - in pseudorandom order (Fig. 1B). Each picture appeared on the computer screen for 1000 ms, and the inter-stimulus interval was set to 1500 ms. After each picture, participants were first prompted to rate its valence on a nine-item self-assessment manikin (SAM) valence scale (1=negative, 5=neutral, 9=positive), and then prompted to rate its arousability on a nine-item SAM arousal scale (1=no arousal, 9=highly arousing). These ratings were used as measures of emotional reactivity (valence reactivity and arousal reactivity, respectively). Ratings were entered using numbers on a keyboard without any time limit. Participants were not informed that their memory for the pictures would be tested later.

During Recognition, participants were shown 180 pictures: the same 60 targets intermixed with 120 novel pictures (foils; 60 neutral and 60 negative). Pictures were displayed for 1000 ms, and participants again rated each on valence and arousal. They were then prompted to indicate whether they had seen each picture before by pressing “y” for yes and “n” for no.

2.4. Polysomnography

Polysomnography (sleep groups only) was recorded in participants’ homes using the Aura PSG ambulatory system (Grass Technologies). Electrodes applied after completion of the Encoding Phase included two EOG (right and left ocular canthi), two chin EMG, and six cortical EEG leads (O1, O2, C3, C4, F3, F4), with all channels referenced to the contralateral mastoid. Recordings were obtained and scored according to the specifications provided by the American Academy of Sleep Medicine (Iber et al., 2007). Due to technical difficulty and loss of data (stolen laptop), successful recording and sleep-stage scoring was possible for 14 young adults and 16 middle-aged adults.

2.5. Data analysis

Participants’ individual valence ratings of the pictures were used to categorize stimuli for analyses (St. Jacques et al., 2009). Targets were categorized based on ratings during the Encoding Phase, and foils were categorized based on ratings during the Recognition Phase. Negative and neutral pictures were defined as those rated 1-3 and 4-6, respectively. Hence, the analyzed picture sets were unique for each participant. One young adult (sleep group) was excluded from all analyses due to rating only one target picture as negative. The average number of target pictures used in analyses is reported in Table 1. On average, 56.63+/−12.48 foil pictures were rated as neutral, and 37.00+/−11.05 were rated as negative. These numbers did not significantly differ between young and middle-aged adults or between sleep and wake groups (p’s>0.16) with the exception that, in middle-aged participants, the sleep group rated significantly fewer foil pictures as negative compared to the wake group (35.19+/−8.10 vs. 42.35+/−12.35, t=−2.206, p=0.033).

Table 1.

Target picture ratings in session 1 (mean (SEM))

# Neg # Neu Val - Neg Val - Neu Aro - Neg Aro - Neu
MA
Sleep 22.62 (1.05) 28.76 (1.41) 1.87 (0.08) 5.02 (0.03) 5.93 (0.35) 2.38 (0.22)
Wake 25.45 (0.74) 25.20 (1.49) 1.63 (0.10) 5.02 (0.04) 6.78 (0.30) 2.32 (0.27)
p-value 0.035* 0.090 0.077 0.982 0.072 0.866
YA
Sleep 20.37 (1.10) 31.95 (1.87) 1.94 (0.11) 4.94 (0.02) 5.39 (0.46) 2.20 (0.20)
Wake 20.30 (1.41) 29.75 (2.37) 1.74 (0.10) 4.96 (0.03) 5.08 (0.50) 2.72 (0.34)
p-value 0.970 0.474 0.199 0.677 0.659 0.200
p MA vs. YA 0.001* 0.041* 0.380 0.030* 0.009* 0.650
Open in a new tab

MA 1 = middle-aged adults, YA = young adults, Neg = negative, Neu = neutral, Val = valence, Aro = arousal, p = p-values,

*

denotes statistical significance

Corrected recognition was calculated by subtracting the false alarm rate, defined as the percentage of foil pictures incorrectly identified as previously seen, from the hit rate, defined as the percentage of target pictures correctly identified as previously seen. Memory discrimination (d’) was also calculated based on hit rate and false alarm rate, with perfect scores of zero (false alarm rate) replaced by 1/(2N) and one (hit rate) replaced by 1-1/(2N), with N being the number of foil trials and target trials, respectively (Macmillan and Kaplan, 1985). Corrected recognition was chosen as the primary measure of memory, whereas the other variables are reported as secondary measures. Changes in valence and arousal ratings were calculated for target pictures (shown at both sessions) as follows: change in valence = session 2 valence rating — session 1 valence rating; change in arousal = session 2 arousal rating — session 1 arousal rating. A positive change in valence score for a negative picture thus indicates a decrease of the initial negative reaction (toward neutrality). A negative change in arousal score for a negative or neutral picture indicates a decrease in arousal from the first to the second session.

Comparisons of means were conducted using Analyses of Variance (ANOVAs), Analyses of Covariance (ANCOVAs), and Student’s independent t-tests. Levene’s test for homogeneity of variances was applied, and adjusted p-values are reported when applicable. Pearson’s r was used to assess linear relationships between variables. Significance levels were set to p < 0.05. A “trend-level” effect was defined as having a p value ≥ 0.05 and < 0.075.

3. Results

Portions of data from young adults have been reported previously (Mantua et al., 2017; non-TBI groups). They were re-analyzed and reported here to serve as a comparison to middle-aged adults. One young adult participant was excluded from all analyses due to an insufficient number of target pictures (only 1 target picture rated as negative). Another young adult participant was excluded from memory analyses due to exceptionally poor memory performance (>3 SD below the mean).

3.1. Sleep characteristics

Habitual sleep quality and self-rated sleepiness did not significantly differ between young and middle-aged adults, nor did these measures significantly differ between sleep and wake groups in either age group (independent-samples t-tests; all p’s>0.1), On the night of the experiment, young adults had significantly lower total sleep time (TST) than did middle-aged adults. No other sleep characteristics significantly differed between young and middle-aged adults (Table 2).

Table 2.

Sleep characteristics (mean (SEM))

TST (min) SL (min) SE (%) RL (min) NREM1(%) NREM2 (%) SWS (%) REM (%)
MA 424.41
(11.63)		 17.13
(9.07)		 94.42
(0.54)		 75.31 (7.93) 5.18 (0.48) 50.98 (1.28) 22.23
(1.72)		 21.62
(1.15)
YA 364.77
(17.85)		 6.75 (1.40)
 93.99
(1.41)		 88.21
(13.04)		 6.73 (1.00) 48.95 (1.54) 25.49
(2.14)		 18.82
(1.25)
p 0.010* 0.299 0.783 0.392 0.176 0.317 0.240 0.109
Open in a new tab

MA = middle-aged adults, YA = young adults, TST=total sleep time, SE=sleep efficiency, RL=REM latency, p = p-value when comparing MA and YA,

*

denotes statistical significance

3.2. Memory

Hit rate, false alarm rate, corrected recognition, and d’ averages are reported in Table 3. To determine whether the influence of sleep differed between young adults and middleaged adults, a three-way ANOVA was conducted on corrected recognition with Age (young vs. middle-aged) and Group (sleep vs. wake) as between-subjects factors and Valence (negative vs. neutral) as a within-subjects factor. There were significant main effects of Age (F(1,75)=5.097, p=0.027; higher in middle-aged adults), Group (F(1,75)=12.261, p=0.001; higher in sleep groups), and Valence (F(1,75)=12.066, p=0.001; higher for negative pictures). Additionally, there was a significant Age X Group interaction (F(1,75)=7.398, p=0.008). Other effects were non-significant (p’s>0.37). To decompose the Age X Group interaction, one-way ANOVAs were used to compare corrected recognition between sleep and wake groups in young and middle-aged adults separately. Because there was no interaction with Valence, corrected recognition was collapsed across negative and neutral pictures. In young adults, sleep benefited corrected recognition (F(1,36)=20.519, p<0.001, Fig. 2 (memory for negative and neutral pictures plotted separately for ease of comparison to other studies)). In contrast, there was no significant influence of sleep on corrected recognition in middle-aged adults (F(1,39)=0.293, p=0.591).

Table 3.

Recognition memory performance (mean (SEM))

Group HR – Neg
(%)		 HR – Neu
(%)		 FAR – Neg
(%)		 FAR – Neu
(%)		 CR – Neg
(%)		 CR – Neu
(%)		 d’ – Neg d’ – Neu
MA
 Sleep 85.99 (1.69) 81.66 (2.44) 12.87 (2.58) 13.26 (2.43) 73.12 (3.60) 68.40 (3.96) 2.43 (0.18) 2.24 (0.17)
 Wake 85.06 (1.67) 75.74 (3.26) 12.61 (3.14) 11.67 (2.54) 72.46 (3.44) 64.07 (3.42) 2.34 (0.14) 2.12 (0.14)
YA
 Sleep 83.65 (2.00) 77.42 (3.85) 8.49 (1.62) 8.12 (1.50) 75.16 (2.62) 69.31 (3.92) 2.46 (0.13) 2.41 (0.17)
 Wake 71.64 (3.52) 65.83 (2.94) 17.42 (2.70) 15.31 (2.66) 54.22 (3.74) 50.51 (3.60) 1.64 (0.14) 1.56 (0.13)
Open in a new tab

MA = middle-aged adults, YA = young adults, HR = hit rate, FAR = false alarm rate, CR = corrected recognition (HR-FAR), Neg = negative, Neu = neutral

Figure 2.

Figure 2.

Open in a new tab

Influence of sleep and wake on negative and neutral memory in young and middleaged adults. Memory is plotted as the average percent corrected recognition (hit rate minus false alarm rate). Error bars represent standard errors of means.

3.3. Emotional reactivity

3.3.1. Baseline picture ratings

Picture ratings in session 1 did not significantly differ between sleep and wake groups in young adults (Table 1). However, middle-aged participants in the wake group rated more pictures as negative and negative pictures higher in arousal (trend-level) compared to the sleep group. Compared to young adults, middle-aged participants rated more pictures as negative and fewer pictures as neutral. Additionally, they rated negative pictures as more arousing and neutral pictures slightly lower in valence (slightly more negative) than did young adults.

3.3.2. Change in valence ratings

Valence ratings of negative pictures significantly increased (became less negative) between encoding and recognition sessions in ah groups (p’s<0.05, one-sample t-tests using 0 as the reference point; Fig. 3). Valence ratings of neutral pictures decreased slightly but significantly between encoding and recognition sessions in all groups (p’s<0,05) except middle-aged adults in the sleep group (p=0.090). These pictures may have taken on a slightly less neutral tone as a consequence of being intermixed with negative pictures during learning. To compare changes among groups, a three-way ANOVA with Age and Group as between-subjects factors and Valence as a within-subjects factor was conducted. There was a significant main effect of Valence (F(1,76)=87.460, p<0.001) and trend-level interactions of Valence X Group (F(1,76)=3.410, p=0.069), Age X Group (F(1,76)=3.426, p=0.068),and Valence X Age X Group (F(1,76)=3.894, p=0.052). Other effects were non-significant (p’s>0.19).

Figure 3.

Figure 3.

Open in a new tab

Influence of sleep and wake on valence reactivity in young and middle-aged adults. Average inter-session change in valence ratings is plotted. Valence was rated on a 9-point Likerttype scale: 1=most negative, 5=neutral, 9=most positive. A positive change in valence indicates ratings became less negative. Error bars represent standard errors of means.

Given that our primary interest was comparing across ages, two-way ANOVAs with Age and Group as factors were conducted for negative and neutral pictures separately in order to decompose the three-way interaction. These yielded a significant Age X Group interaction for negative pictures (F(1,76)=3.988, p=0.049) and a significant main effect of group for neutral pictures (/’n.7r>)=4.491, p=0.037; smaller decrease in sleep group). No other effects were significant (p’s>0.13). The increase in valence of negative pictures was smaller in the sleep group than wake group in young adults (t=−1.839, p=0.074, trend-level) and similar between sleep and wake groups in middle-aged adults (t=0.290, p=0.774). Thus, sleep appears to preserve valence reactivity to negative pictures in young adults but not middle-aged adults, while the effect of sleep on reactivity for neutral pictures was similar between age groups.

3.3.3. Change in arousal ratings

To compare inter-session changes in arousal ratings among groups, a three-way ANOVA with Age and Group as between-subjects factors and Valence as a within-subjects factor was conducted. This test yielded a significant main effect of Age (F(1,76)=4.283, p=0.042), with greater decrease in arousal in middle-aged adults. Higher baseline arousal ratings in middle-aged adults may have contributed to this effect. No other effects were significant (p’s>0.25; Table 4). Thus, in this study, sleep does not appear to influence arousal reactivity.

Table 4.

Change in arousal ratings (mean (SEM))

Group Negative Neutral
MA
 Sleep −0.312 (0.217) −0.387 (0.177)
 Wake −0.442 (0.147) −0.471 (0.177)
YA
 Sleep 0.162 (0.329) −0.234 (0.281)
 Wake 0.089 (0.258) −0.151 (0.289)
Open in a new tab

MA = middle-aged adults, YA = young adults

3.4. Control analyses

In order to control for group differences in emotional reactivity at encoding, memory analyses were repeated with average baseline valence and arousal ratings of negative pictures included as covariates (responses to neutral pictures were not included given their lower magnitude and lower variability). The Age X Group X Valence ANCOVA yielded a significant Age X Group interaction (F(1,73)=5.819, p=0.018) as well as significant main effects of Age (F(1,73)=4.403, p=0.039), Group (F(1,73)=8.264, p=0.005), and baseline valence ratings (F(1,73)=6.252, p=0.015; higher (less negative) valence ratings associated with higher corrected recognition). Other effects were non-significant (p’s>0.095). In young adults, a one-way ANCOVA with Group as a between-subjects factor yielded significant main effects of Group (F(1,34)=14.820, p<0.001; higher corrected recognition in sleep group) and baseline arousal ratings (F(1,34)=4.554, p=0.040; higher arousal ratings associated with higher corrected recognition) and a non-significant effect of baseline valence ratings (F(1,37)=1.922, p=0.175). In middle-aged adults, a oneway ANCOVA yielded no effect of Group (F(1,37)<0.001, p=0.989) or baseline arousal ratings (F(1,37)=0.017, p=0.897), but there was a significant effect of baseline valence ratings (F(1,37)=4.133, p=0.049). These results suggest the memory benefit of sleep observed in young adults may not extend to middle-aged adults, and this pattern cannot be explained by differences in emotional response at encoding.

Given the age difference in total sleep time, we sought to determine whether this difference could be driving the pattern of results. We first investigated relationships between total sleep time and memory (collapsed across negative and neutral pictures) and change in valence for negative pictures (since an age difference was only present for negative pictures). In the sleep groups, there was no significant relationship between total sleep time measured with PSG and corrected recognition in young (r=0.50, p=0.082) or middle-aged adults (r=−0.036, p=0.902), though we note the marginal effect in young adults. There was also no significant relationship between total sleep time and change in valence for negative pictures in young (r=0.002, p=0.994) or middle-aged adults (r=0.057, p=0.846). In the wake groups, we calculated total sleep time the night before the experiment based on sleep diary reports. As in sleep groups, there was a significant difference between age groups (t=−3.110, p=0.004), with middle-aged adults sleeping longer (M=460.50 min, SD=77.28) than young adults (M=373.93 min, SD=83.53 min). However, there was no significant relationship between sleep duration the night before learning and corrected recognition in young (r=−0.057, p=0.847) or middle-aged adults (r=−0.013, p=0.957), or between sleep duration and change in valence in young (r=0.112, p=0.704) or middle-aged adults (r=0.094, p=0.693).

Finally, we sought to equate young and middle-aged adults on TST the night of the experiment (sleep groups) or the night before the experiment (wake groups). If PSG was not available for sleep participants, we used TST from their sleep diary reports. We subsampled young adults with the longest sleep durations and middle-aged adults with the shortest sleep durations, such that age groups were evenly matched on TST (sleep groups: n=12, p=0.986; wake groups: n=11, p=0.968). Despite reduced power, a two-way ANOVA still yielded a trend-level Group X Age interaction (F(1,42)=3.325, p=0.075) for corrected recognition, reflecting significantly better memory in the sleep group than wake group in young adults (t=3.301, p=0.003) but not middle-aged adults (t=0.403, p=0.691). With regard to change in valence for negative pictures, the Group X Age interaction no longer reached significance using this reduced sample size (F(1,42)=2.403, p=0.129), but the pattern of results was very similar, with the increase in valence approximately 2.75 times greater in the wake group (M=0.742, SD=0.704) than sleep group (M=0.267, SD=0.650) in young adults (t=−1.683, p=0.107) and slightly less in the wake group (M=0.413, SD=0.464) than the sleep group (M=0.525, SD=0.708) in middle-aged adults (t=0.443, p=0.662). Thus, our pattern of results does not appear to be explained by age differences in sleep duration.

4. Discussion

Here we show that young adults who slept had higher memory performance than those who remained awake, regardless of negative or neutral memory valence (as previously reported in Mantua et al., 2017). In middle-aged adults, wake and sleep groups performed at similar levels, also regardless of negative or neutral memory valence. This pattern is very similar to that which we previously observed in older adults (Jones et al., 2016). Thus, the relative influence of sleep compared to wake on declarative memory performance appears to change between young adulthood and middle age, and this change extends to negative memories, as depicted in Fig. 4.

Figure 4.

Figure 4.

Open in a new tab

Influence of sleep and wake on negative memory and valence reactivity to negative pictures in young (YA), middle-aged (MA), and older (OA) adults. Memory (left) is plotted as the average percent corrected recognition (hit rate minus false alarm rate). Change in valence reactivity (right) is plotted as the average inter-session change in valence ratings. Valence was rated on a 9-point Likert-type scale: 1=most negative, 5=neutral, 9=most positive. A positive change in valence indicates ratings became less negative. Error bars represent standard errors of means. Older adult data is taken from Jones et al. (2016).

One potential interpretation of these results is that sleep-related processes benefit memory consolidation in young adults but that these processes are impaired with aging, resulting in no advantage of sleep over wake in middle-aged adults. Impairment would be consistent with previous evidence that reduced slow wave sleep and slow wave activity are associated with poorer declarative memory in middle-aged and older adults (Backhaus et al., 2007; Mander et al., 2013; Scullin, 2013). Perhaps surprisingly, in the current study we found no difference between age groups in the time spent in slow wave sleep or other sleep stages. This result may reflect relatively low sleep duration in our young adult sample. Regardless, age-related changes in the distribution of sleep stages does not appear to account for the memory differences observed here. However, we did not analyze microstructure of sleep such as slow wave activity, so we cannot rule out more fine-grained sleep differences between groups.

Another potential interpretation is that sleep-related processes are still intact with aging but that a performance benefit is more difficult to uncover because there is less memory decay over wake. In other words, there is an age-related memory improvement over waking. Our results would seem to support this interpretation since middle-aged adults performed to the level of young adults who slept. However, we believe this account is unlikely for a few reasons. First, previous studies typically see either a detrimental effect or no effect of aging on memory retention across wakefulness (Cherdieu et al., 2014; Giambra and Arenberg, 1993; Mander et al., 2013; Scullin, 2013; Wilson et al., 2012). Second, we believe other factors could more easily explain the relatively good memory performance in middle-aged compared to young adults in this study. Research participation is likely a more novel experience for middle-aged adults compared to young adult college students who participate in many studies to earn class credit. Middle-aged individuals may thus be more interested and engaged during the task and may also be more motivated to perform well. These factors may have boosted general performance, resulting in better memory in the wake group of middle-aged compared to young adults. On the other hand, sleep restriction in young adults, as evidenced by short sleep duration, may have led to a relatively low baseline performance in young adults in this study. Nonetheless, it should be kept in mind that we cannot rule out age-related memory improvement over wake based on our data. Future studies should investigate and control for the amount of interference experienced over the day in different age groups, as this factor could influence memory retention.

Our interpretation is that memory consolidation processes become more selective with aging. Socioemotional selectivity theory posits that time horizons (perceived future time) influence goals such that young adults, perceiving their futures to be relatively open-ended, prioritize novelty and knowledge acquisition. As individuals perceive their futures becoming more limited with aging, they begin prioritizing emotional well-being (Carstensen, 2006; Carstensen et al., 1999). In support of this theory, older adults often show a positivity bias in attention and memory (Mather and Carstensen, 2005; Reed et al., 2014). A positive bias may begin to emerge in middle age (Charles et al., 2003). Neuroimaging research has shown that aging leads to chronic activation of an emotional regulation network (Dolcos et al., 2014; Leclerc and Kensinger, 2010; Sakaki et al., 2013). Thus, positive memories may be tagged at encoding for subsequent consolidation during sleep, while negative and neutral information is left untagged. In line with this view, we recently found that older adults display a sleep benefit for positive but not negative memories (Jones et al., 2016). Future studies could include a positive picture condition to determine whether sleep benefits positive memories in middle-aged adults.

We did not find a selective sleep benefit for negative memories over neutral memories in young adults, as some past research has shown (Nishida et al., 2009; Payne et al., 2008; Wagner et al., 2006). This may indicate that sleep consolidates emotional and neutral memories equally, as previous evidence suggests (Cairney et al., 2014; Cellini et al., 2016; Lehmann et al., 2016; Lewis et al., 2011). However, it should be noted that using the same memory task employed here, we previously found that the influence of sleep on memory for neutral pictures was driven by the effect on memory for negative pictures, since there was no sleep benefit when neutral pictures were encoded intermixed with positive pictures (Jones et al., 2016). Thus, it is possible that the benefit for neutral memories in this study likewise reflects a carry-over effect. As in young adults, negative emotion did not interact with sleep group in middle-aged adults, suggesting that if there is an age-related decline in efficacy or engagement of sleep-related processing, it extends to negative memories. Alternatively, if there is an age-related improvement in memory over waking, then there is no influence of negative valence on this improvement.

With regard to reactivity, we found that valence ratings for negative pictures were preserved in young adults who slept compared to those who remained awake, although it should be kept in mind that this effect was statistically marginal. This result is consistent with past research suggesting sleep consolidates not only memory contents but also aspects of associated emotionality in young adulthood (Baran et al., 2012; Groch et al., 2013; Lara-Carrasco et al., 2009; Werner et al., 2015; though sleep may depotentiate emotional reactivity in some cases (Cunningham et al., 2014; Pace-Schott et al., 2011; van der Helm et al., 2011)). This pattern was not observed in middle-aged adults, wherein negative valence declined similarly in sleep and wake groups. This finding is consistent with results previously observed in older adults (Jones et al., 2016), perhaps suggesting an age-related change in the impact of sleep on reactivity associated with negative memories (see Fig. 4). Alternatively, there may be enhanced consolidation of reactivity over wake intervals with aging. Interestingly, ratings for neutral pictures became slightly less neutral and more negative. Negativity may have generalized to neutral pictures due to intermixed stimulus presentation at encoding. Across age groups, sleep tended to mitigate this effect, preserving neutrality.

Middle-aged adults rated negative pictures more arousing at encoding than did young adults, and middle-aged adults in the wake group rated arousal marginally higher than did those in the sleep group. It is important to consider whether differences in arousal levels impacted performance, since arousal has been shown to enhance memory for emotional and neutral information (McGaugh, 2004). Indeed, in young adults, baseline arousal ratings were positively associated with subsequent memory. Arousal during encoding may “tag” memories for consolidation in young adults via amygdalar modulation of hippocampal activity (Bergado et al., 2011; McReynolds and McIntyre, 2012). However, arousal ratings were not related to memory in middle-aged adults, and statistically controlling for arousal during encoding did not alter results. Interestingly, neuroimaging research suggests that amygdala activity and amygdala-hippocampal functional connectivity is reduced when older adults view negative pictures (Leclerc and Kensinger, 2010; Mather et al., 2004; St. Jacques et al., 2009). Thus, the impact of arousal on the encoding and consolidation of memories may be reduced with age. For these reasons, we believe it is unlikely that group differences in arousal contributed to differences in sleep/wake memory patterns between young and middle-aged adults.

Another important consideration is whether circadian influences contributed to the age differences observed in this study. Young adults tend to be “evening-types”, and a shift toward “morning-type” occurs with age, particularly around age 50 (Adan et al., 2012; Ishihara et al., 1991). Although our middle-aged participants were relatively young (35-50 years), it is nonetheless possible that chronotype differed between age groups. Since cognitive performance may vary between peak and off-peak times of day (Schmidt et al., 2007), this could have led young adults to perform relatively better in the evening and middle-aged adults to perform relatively better in the morning. However, an age difference in chronotype is unlikely to account for our results for several reasons. First, young and middle-aged adults did not differ on subjective sleepiness in either the morning or evening, which might be expected if there were a meaningful difference in chronotype. Second, previous work using the identical task as the current study demonstrated no short-term memory differences between young adults who learned in the morning and evening (Baran et al., 2012), suggesting time of day does not influence encoding or recognition performance of this task, at least in young adults. Third, if time of day affected processes occurring during the recognition task, then we would expect young adults in the wake group to out-perform the sleep group, and vice-versa in middle-aged adults, which was not the case. Finally, with regard to the time of day of learning/encoding, past studies have shown that middle-aged and older adults who learn in the evening and sleep overnight perform better than those who learn in the morning and remain awake during the day (e.g. Wilson et al., 2012), suggesting that any time of day effect on learning would not occlude the effect of sleep-related consolidation in these age groups. Thus, the pattern of results observed here cannot easily be explained by potential differences in chronotype.

Finally, young adults in both sleep and wake groups slept less than middle-aged adults, and slept less than recommended for their age, suggesting they were acutely and perhaps chronically sleep restricted. However, there was no difference between ages or between sleep and wake groups on self-rated sleepiness, suggesting sleepiness during testing cannot explain the pattern of results. Furthermore, results were unchanged when ages were matched on sleep duration. Sleep restriction the night before learning would be expected to hinder memory encoding in both sleep and wake groups, perhaps moreso in the sleep group where encoding occurred after a day of wakefulness (Mander et al., 2011). Thus, while it is quite possible that sleep restriction impacted performance generally in young adults, it cannot easily explain the sleep benefit in this age group. Nonetheless, future studies could be improved with better controlled sleep schedules leading up to and during experiments, which may be especially pertinent in a student population.

5. Conclusions

The effect of sleep on memory consolidation is well documented in young adults. There is reason to believe this process may change in middle-age, but few studies have empirically investigated this possibility. The current study provides evidence that the relative influence of sleep compared to wake on declarative memory declines between young adulthood and middle age, and this decline extends to negative memory. As no age differences in sleep stage parameters were observed, this decline may reflect changes in the microstructure of sleep, changes in processing over wake, or an age-related shift toward more selective processing, consistent with socioemotional selectivity theory. Future work is needed to further understand how sleep-related memory processing changes across adulthood.

Highlights.

  • Compared to time spent awake, sleep benefited memory in young adults

  • Memory did not differ between sleep and wake groups in middle-aged adults

  • Sleep vs. wake memory patterns were similar for negative and neutral memory

Acknowledgements

This work was supported by NIH R01 AG040133 and an Honors Research Grant from Commonwealth Honors College. We thank Ahren Fitzroy for helpful comments on the manuscript.

Footnotes

Disclosure statement

The authors declare no conflicts of interest. All procedures were approved by the Institutional Review Board of the University of Massachusetts, Amherst.

Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. 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.

References

  1. Adan A, Archer SN, Hidalgo MP, Di Milla L, Natale V, Randler C, 2012. Circadian typology: A comprehensive review. Chronobiol. Int 29, 1153–1175. [DOI] [PubMed] [Google Scholar]
  2. Aly M, Moscovitch M, 2010. The effects of sleep on episodic memory in older and younger adults. Memory. 18, 327–334. [DOI] [PubMed] [Google Scholar]
  3. Baran B, Pace-Schott EF, Ericson C, Spencer RMC, 2012. Processing of emotional reactivity and emotional memory over sleep. J. Neurosci 32, 1035–1042. [DOI] [PMC free article] [PubMed] [Google Scholar]
  4. Backhaus J, Born J, Hoeckesfeld R, Fokuhl S, Hohagen F, & Junghanns K, 2007. Midlife decline in declarative memory consolidation is correlated with a decline in slow wave sleep. Learn. Mem 14, 336–341. [DOI] [PMC free article] [PubMed] [Google Scholar]
  5. Bennion KA, Payne JD, Kensinger EA, 2015. Sleep and cortisol interact to support memory consolidation. Cereb. Cortex 25, 646–657. [DOI] [PubMed] [Google Scholar]
  6. Bergado JA, Lucas M, Richter-Levin G 2011. Emotional tagging - a simple hypothesis in a complex reality. Prog. Neurobiol 94, 64–76. [DOI] [PubMed] [Google Scholar]
  7. Born J, Wilhelm F, 2012. System consolidation of memory during sleep. Psychol. Res 76, 192–203. [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. Buysse DJ, Reynolds CF 3rd, Monk TH, Berman SR, Kupfer DJ, 1989. The Pittsburgh Sleep Quality Index: A new instrument for psychiatric practice and research. Psychiatry Res 28, 193–213. [DOI] [PubMed] [Google Scholar]
  9. Cairney SA, Durrant SJ, Jackson R, Lewis PA, 2014. Sleep spindles provide indirect support to the consolidation of emotional encoding contexts. Neuropsychologia 63, 285–292. [DOI] [PubMed] [Google Scholar]
  10. Carrier J, Land S, Buysse DJ, Kupfer DJ, Monk TH, 2001. The effects of age and gender on sleep EEG power spectral density in the middle years of life (ages 20-60 years old). Psychophysiology 38, 232–42. [PubMed] [Google Scholar]
  11. Carrier J, Viens I, Poirier G, Robillard R, Lafortune M, Vandewalle G, Martin N, Barakat M, Paquet J, Filipini D, 2011. Sleep slow wave changes during the middle years of life. Eur. J. Neurosci 33, 758–766. [DOI] [PubMed] [Google Scholar]
  12. Carstensen LL, 2006. The influence of a sense of time on human development. Science. 312, 1913–1915. [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Carstensen LL, Isaacowitz DM, Charles ST, 1999. Taking time seriously: a theory of socioemotional selectivity. Am. Psychol 54, 165–81. [DOI] [PubMed] [Google Scholar]
  14. Cellini N, Torre I, Stegagno L, Sarlo M, 2016. Sleep before and after learning promotes the consolidation of both neutral and emotional information regardless of REM presence. Neurobiol. Learn. Mem 133, 136–144. [DOI] [PubMed] [Google Scholar]
  15. Charles ST, Mather M, Carstensen LL, 2003. Aging and emotional memory: The forgettable nature of negative images for older adults. J. Exp. Psychol. Gen 132, 310–324. [DOI] [PubMed] [Google Scholar]
  16. Cherdieu M, Reynaud E, Uhlrich J, Versace R, Mazza S, 2014. Does age worsen sleep-dependent memory consolidation? J. Sleep Res 23, 53–60. [DOI] [PubMed] [Google Scholar]
  17. Cunningham TJ, Crowell CR, Alger SE, Kensinger EA, Villano MA, Mattingly SM, Payne JD, 2014. Psychophysiological arousal at encoding leads to reduced reactivity but enhanced emotional memory following sleep. Neurobiol. Learn. Mem 114, 155–64. [DOI] [PubMed] [Google Scholar]
  18. Dolcos S, Katsumi Y, Dixon RA, 2014. The role of arousal in the spontaneous regulation of emotions in healthy aging: A fMRI investigation. Front. Psychol 5: 681. doi: 10.3389/fpsyg.2014.00681. eCollection 2014. [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Genzel L, Spoormaker VI, Konrad BN, Dresler M, 2015. The role of rapid eye movement sleep for amygdala-related memory processing. Neurobiol. Learn. Mem 122, 110–21. [DOI] [PubMed] [Google Scholar]
  20. Giambra LM, Arenberg D, 1993. Adult age differences in forgetting sentences. Psychol. Aging 8, 451–462. [DOI] [PubMed] [Google Scholar]
  21. Groch S, Wilhelm L, Diekelmann S, Sayk F, Gais S, Born J, 2013. The role of REM sleep in the processing of emotional memories: Evidence from behavior and event-related potentials. Neurbiol. Learn. Mem 99, 1–9. [DOI] [PubMed] [Google Scholar]
  22. Iber C, Ancoli-Israel S, Chesson A, Quan SF, 2007. The AASM manual for the scoring of sleep and associated events: rules, terminology, and technical specification. (American Academy of Sleep Medicine, Westchester, IL, 1st ed). [Google Scholar]
  23. Ishihara K, Miyake S, Miyasita A, Miyata Y, 1991. Morningness—eveningness preference and sleep habits in Japanese office workers of different ages. Chronobiologia. 18, 9–16. [PubMed] [Google Scholar]
  24. Jones BJ, Schultz KS, Adams S, Baran B, Spencer RMC, 2016. Emotional bias of sleep-dependent processing shifts from negative to positive with aging. Neurobiol. Aging 45, 178–189. [DOI] [PMC free article] [PubMed] [Google Scholar]
  25. LaBar KS, Cabeza R, 2006. Cognitive neuroscience of emotional memory. Nat. Rev. Neurosci 7, 54–64. [DOI] [PubMed] [Google Scholar]
  26. Lang PJ, Bradley MM, Cuthbert BN, 2005. “International affective picture system (IAPS): Affective ratings of pictures and instruction manual” (Tech. Rep. A-6, Univ. Florida, Gainesville, FL: ). [Google Scholar]
  27. Lara-Carrasco J, Nielsen TA, Solomonova E, Levrier K, Popova A, 2009. Overnight emotional adaptation to negative stimuli is altered by REM sleep deprivation and is correlated with intervening dream emotions. J. Sleep Res 18, 178–87. [DOI] [PubMed] [Google Scholar]
  28. Leclerc CM, Kensinger EA, 2010. Age-related valence-based reversal in recruitment of medial prefrontal cortex on a visual search task. Soc. Neurosci 5, 560–76. [DOI] [PubMed] [Google Scholar]
  29. Lehmann M, Seifritz E, Rasch B, 2016. Sleep benefits emotional and neutral associative memories equally. Somnologie 20, 47–53. [Google Scholar]
  30. Lewis PA, Cairney S, Manning L, Critchley HD, 2011. The impact of overnight consolidation upon memory for emotional and neutral encoding contexts. Neuropsychologia 49, 2619–2629. [DOI] [PMC free article] [PubMed] [Google Scholar]
  31. Macmillan NA, Kaplan HL, 1985. Detection theory analysis of group data: Estimating sensitivity from average hit and false-alarm rates. Psychol. Bull 98, 185–199. [PubMed] [Google Scholar]
  32. Mander BA, Rao V, Lu B, Saletin JM, Lindquist JR, Ancoli-Israel S, Jagust W, Walker MP, 2013. Prefrontal atrophy, disrupted NREM slow waves and impaired hippocampal-dependent memory in aging. Nat. Neurosci 16, 357–64. [DOI] [PMC free article] [PubMed] [Google Scholar]
  33. Mander BA, Santhanam S, Saletin JM, Walker ΜP, 2011. Wake deterioration and sleep restoration of human learning. Curr. Biol 21, R183–R184. [DOI] [PMC free article] [PubMed] [Google Scholar]
  34. Mantua J, Henry OS, Garskovas NF, Spencer RMC, 2017. Mild traumatic brain injury chronically impairs sleep- and wake-dependent emotional processing. Sleep. 40, doi: 10.1093/sleep/zsx062. [DOI] [PMC free article] [PubMed] [Google Scholar]
  35. Martin N, Lafortune M, Godbout J, Barakat M, Robillard R, Poirier G, Bastien C, Carrier J 2013. Topography of age-related changes in sleep spindles. Neurobiol. Aging. 34, 468–76. [DOI] [PubMed] [Google Scholar]
  36. Mather M, Canli T, English T, Whitfield SL, Wais PE, Ochsner KN, John DEG, Carstensen LL, 2004. Amygdala responses to emotionally valenced stimuli in older and younger adults. Psychol. Sci 15, 259–263. [DOI] [PubMed] [Google Scholar]
  37. Mather M, Carstensen LL, 2005. Aging and motivated cognition: The positivity effect in attention and memory. Trends Cogn. Sci 9, 496–502. [DOI] [PubMed] [Google Scholar]
  38. McGaugh JL, 2004. The amygdala modulates the consolidation of memories of emotionally arousing experiences. Annu. Rev. Neurosci 27, 1–28. [DOI] [PubMed] [Google Scholar]
  39. McReynolds JR, McIntyre CK, 2012. Emotional modulation of the synapse. Rev. Neurosci 23, 449–461. [DOI] [PMC free article] [PubMed] [Google Scholar]
  40. Musiek ES, Holtzman DM, 2016. Mechanisms linking circadian clocks, sleep, and neurodegeneration. Science. 354, 1004–1008. [DOI] [PMC free article] [PubMed] [Google Scholar]
  41. Nishida M, Pearsall I, Buckner RL, Walker MP, 2009. REM sleep, prefrontal theta, and the consolidation of human emotional memory. Cereb. Cortex 19, 1158–1166. [DOI] [PMC free article] [PubMed] [Google Scholar]
  42. Ohayon MM, Carskadon MA, Guilleminault C, Vitiello MV, 2004. Meta-analysis of quantitative sleep parameters from childhood to old age in healthy individuals: Developing normative sleep values across the human lifespan. Sleep. 27, 1255–73. [DOI] [PubMed] [Google Scholar]
  43. Pace-Schott EF, Shepherd E, Spencer RM, Marcello M, Tucker M, Propper RE, Stickgold R, 2011. Napping promotes inter-session habituation to emotional stimuli. Neurobiol. Learn. Mem 95, 24–36. [DOI] [PMC free article] [PubMed] [Google Scholar]
  44. Payne JD, Stickgold R, Swanberg K, Kensinger EA, 2008. Sleep preferentially enhances memory for emotional components of scenes. Psych. Sci 19, 781–788. [DOI] [PMC free article] [PubMed] [Google Scholar]
  45. Rasch B, Born I, 2013. About sleep’s role in memory. Physiol. Rev 93, 681–766. [DOI] [PMC free article] [PubMed] [Google Scholar]
  46. Reed AE, Chan L, Mikels JA, 2014. Meta-analysis of the age-related positivity effect: age differences in preferences for positive over negative information. Psychol. Aging. 29, 1–15. [DOI] [PubMed] [Google Scholar]
  47. Richter-Levin G, Akirav I, 2003. Emotional tagging of memory formation—in the search for neural mechanisms. Brain Res. Brain Res. Rev 43, 247–256. [DOI] [PubMed] [Google Scholar]
  48. Sakaki M, Nga L Mather M, 2013. Amygdala functional connectivity with medial prefrontal cortex at rest predicts the positivity effect in older adults’ memory. J. Cogn. Neurosci 25, 1206–1224. [DOI] [PMC free article] [PubMed] [Google Scholar]
  49. Schmidt C, Collette F, Cajochen C, Peigneux P, 2007. A time to think: Circadian rhythms in human cognition. Cogn. Neuropsychol 24, 755–789. [DOI] [PubMed] [Google Scholar]
  50. Scullin MK, 2013. Sleep, memory, and aging: The link between slow-wave sleep and episodic memory changes from younger to older adults. Psychol. Aging 28, 105–114. [DOI] [PMC free article] [PubMed] [Google Scholar]
  51. Stickgold R, 2005. Sleep-dependent memory consolidation. Nature. 437, 1272–8. [DOI] [PubMed] [Google Scholar]
  52. Jacques PL, Dolcos F, Cabeza R, 2009. Effects of aging on functional connectivity of the amygdala for subsequent memory of negative pictures: A network analysis of functional magnetic resonance imaging data. Psych. Sci 20, 74–84. [DOI] [PMC free article] [PubMed] [Google Scholar]
  53. van der Helm E, Yao J, Dutt S, Rao V, Saletin JM, Walker MP, 2011. REM sleep depotentiates amygdala activity to previous emotional experiences. Curr. Biol 21, 2029–2032. [DOI] [PMC free article] [PubMed] [Google Scholar]
  54. Wagner El., Hallschmid M, Rasch B, Born J, 2006. Brief sleep after learning keeps emotional memories alive for years. Biol. Psychiatry. 60, 788–790. [DOI] [PubMed] [Google Scholar]
  55. Werner GG, Schabus M, Blechert J, Kolodyazhniy V, Wilhelm FH, 2015. Pre-to postsleep change in psychophysiological reactivity to emotional films: Late-night REM sleep is associated with attenuated emotional processing. Psychophysiology. 52, 813–25. [DOI] [PubMed] [Google Scholar]
  56. Wilson JK, Baran B, Pace-Schott EF, Ivry RB, Spencer RM, 2012. Sleep modulates word-pair learning but not motor sequence learning in healthy older adults. Neurobiol. Aging 33, 991–1000. [DOI] [PMC free article] [PubMed] [Google Scholar]

RESOURCES