Effects of a Month-long Napping Regimen in Older Individuals

Scott S. Campbell, PhD; Michele D. Stanchina, BA; Joelle R. Schlang, BS; Patricia J. Murphy, PhD

J Am Geriatr Soc. 2011;59(2):224-232. 

Abstract and Introduction


Objectives: To examine the effects of a month-long nap regimen using one of two durations (45 minutes or 2 hours) on nighttime sleep and waking function in a group of healthy older participants and to assess the degree to which healthy older individuals are willing and able to adhere to such napping regimens.
Design: Three laboratory sessions, with 2-week at-home recording interspersed, using a between-participants approach.
Setting: Laboratory of Human Chronobiology at Weill Cornell Medical College and participants' homes.
Participants: Twenty-two healthy men and women aged 50 to 88 (mean 70).
Measurements: Polysomnography (sleep electroencephalography), actigraphy, sleep diaries, neurobehavioral performance, sleep latency tests.
Results: With the exception of adherence to the protocol, there were few differences between short and long nap conditions. Napping had no negative effect on subsequent nighttime sleep quality or duration, resulting in a significant increase in 24-hour sleep amounts. Such increased sleep was associated with enhanced cognitive performance but had no effect on simple reaction time. Participants were generally able to adhere better to the 45-minute than the 2-hour nap regimen.
Conclusion: A month-long, daily nap regimen may enhance waking function without negatively affecting nighttime sleep. Using 2-hour naps in such a regimen is unlikely to meet with acceptable adherence; a regimen of daily 1-hour naps may be more desirable for effectiveness and adherence.


Aging is associated with significant changes in the structure and quality of sleep. Individuals aged 60 and older exhibit less slow-wave sleep, less spectral power of electroencephalographic delta activity, shorter rapid eye movement (REM) onset latencies, more and longer within-sleep awakenings, and significantly earlier terminal waking time than healthy young adults. These latter two events result in an average nighttime sleep duration almost 2 hours shorter than that of healthy young adults.[1–5]

Efforts to extend the nighttime sleep of older people have been largely unsuccessful, leading several authors to conclude that many older individuals may be incapable of obtaining more than approximately 6 hours of sleep per 24 hours.[3] Such speculation has led researchers to investigate the possibility that 24-hour sleep amounts could be increased significantly by introducing an afternoon nap into the sleep schedules of older participants.[6–11] It was found, for example, that a 2-hour afternoon sleep opportunity resulted in an average increase in 24-hour sleep amounts of 81 minutes (from 6.05 hours to 7.4 hours). The effect of the nap on subsequent nighttime sleep quality was limited to a slight increase in sleep onset latency (21.8 minutes vs 15.5 minutes in the control condition). No changes in sleep efficiency or amounts or proportions of non-REM or REM sleep were observed.[10]

A longer (17 day) study of napping in healthy older participants, in which subjective and objective data were examined,[11] reported mixed results. Sleep logs maintained by participants at home revealed significantly greater 24-hour sleep amounts, with no significant effect on nighttime sleep duration or quality. In contrast, laboratory-based polysomnography (PSG) showed a nonsignificant decline in 24-hour sleep amounts as a consequence of significant reductions in nighttime sleep duration and efficiency.

A study[8,9] examined the effects of a month of afternoon naps plus evening exercise on nighttime sleep in 11 elderly individuals with chronic sleep difficulties. Although the relative influence of naps versus exercise could not be determined, the treatment combination consolidated nighttime sleep, and sleep efficiency increased from an average of 75% at baseline to 90% during the last week of the study.

In the authors' earlier study, the increase in total sleep time per 24 hours was associated with significant improvements in a number of cognitive performance measures immediately after the nap, as well as throughout the following day. Similar results were reported[7] for a group of healthy, older habitual nappers who, on one occasion, took a 30-minute nap and, on another, remained sedentary, but awake, for 30 minutes. Performance on a visual detection task was significantly better under the nap condition than the sedentary condition. In contrast to these findings, another study reported no significant effect of naps on several measures of cognitive and psychomotor performance.[11]

In light of these somewhat mixed findings and because of the relative paucity of data addressing the effects of naps in older participants, this study was designed to build on the authors' previous study by examining the longer-term effect of a napping regimen on nighttime sleep quality and daytime cognitive performance. This study also sought to determine the feasibility of such a longer-term regimen in terms of participant adherence. This article describes the results of a 6-week protocol that involved an at-home napping regimen interspersed with objective laboratory assessments of sleep and neurobehavioral function.



The data reported here are from 22 participants (11 men, 11 women) aged 50 and older (mean 70 ± 10, range 50–83). The Weill Cornell Medical College institutional review board approved the protocol. All participants provided written informed consent and were compensated for their participation.

After responding to advertisements to the general public targeting older individuals with or without sleep disturbance, age-eligible individuals were screened over the telephone for additional inclusion and exclusion criteria relating to medication usage, sleep and napping habits, and subjective sleep ratings. This was followed by an in-person physical and mental health screening and a tour of the laboratory facilities.

Thirty-seven potential participants underwent the in-person screening interview. Of these, three were ineligible because of medication usage, and an additional five declined to participate, leaving 29 participants enrolled in the study. Two enrolled participants completed only the baseline session, and two were excluded after the first night spent in the laboratory because of suspected, significant periodic limb movement disorder (see below). Of the 25 who completed the protocol, three had data sets that were not usable: one because of Actiwatch malfunction, one because the participant took melatonin for sleep problems sporadically throughout the study, and one because she started smoking during the study.

Participants were in self-reported good physical health and, at the time of the study, were not taking psychotropic medications or any other medications known to interfere with normal sleep. Minor, controlled health problems such as mild hypertension and mild arthritis were not grounds for exclusion from the study. A brief psychiatric screening (17-item Hamilton Depression Rating Scale (HDRS-17[12]) was completed on each participant. A score less than 7 was required for participation. In addition, participants were excluded if they had a score greater than 5 on the global Pittsburg Sleep Quality Index (PSQI[1]) or greater than 2 on the sleep latency or use of sleeping medication PSQI subscales.

Only people who did not report habitual napping were enrolled. Eligible participants had to report taking an average of fewer than two naps per week in the last 6 months. Nevertheless, all participants consented to following the protocol, including the requirement to incorporate a daily nap into their routines.

Although all participants reported age-related sleep problems, primarily in the form of sleep maintenance difficulties, insufficient sleep duration, or both, a previous diagnosis or current evidence of other sleep disorders (e.g., sleep apnea, periodic limb movement disorder (PLMD), narcolepsy, REM behavior disorder, circadian rhythm sleep disorder, restless legs syndrome, primary insomnia) excluded potential participants. An extensive sleep history was obtained during the screening examination, and questionnaires, including the Sleep Disorders Questionnaire,[13] Epworth Sleepiness Scale,[14] and a separate questionnaire probing for symptoms of restless legs syndrome,[15] were used to screen out those with significant sleep pathologies. Also, on the first night in the laboratory, pulse oximetry, and bilateral leg electromyelography recordings were used to further screen participants for sleep apnea or PLMD (see below).


Baseline: In-home and Laboratory Session Participants maintained daily sleep logs, and actigraphy was recorded continuously for 1 to 2 weeks before the baseline laboratory session. These data were used to confirm the self-reported sleep schedules and frequency of napping and to calculate habitual bedtimes and waketimes for use during the laboratory session. On the day immediately after the conclusion of the in-home baseline phase, participants reported to the sleep laboratory by 7:00 p.m. for the baseline laboratory session. This session consisted of 3 consecutive nights and the 2 intervening days during which participants remained in the laboratory. After they settled into private bedrooms, electrodes were applied for PSG recording. A trained research assistant then introduced participants to the neurobehavioral performance assessment battery (PAB). The familiarization session continued until the participant had an understanding of how to perform each task.

Bedtimes and waketimes in the laboratory were individualized for each participant, as calculated from the average weekday sleep times reported in the sleep logs maintained during the prebaseline in-home phase. In the laboratory, each participant slept in a private, darkened, sound-attenuated bedroom.

On the first laboratory night, participants were screened for sleep pathology using pulse oximetry and bilateral leg electromyelography. Participants whose oxygen saturation levels dropped below 90% or who had an index of more than 10 leg movements per hour associated with an electroencephalographic (EEG)-defined arousal were excluded from continued participation. Based on these criteria, two potential participants were excluded from the study.

Beginning 2 hours after awakening from Night 1 and then every 2 hours for the next 10 hours, participants practiced the PAB. Participants were required to practice the PAB until their trial-to-trial deviation in accuracy was less than 5% across three consecutive trials on each of the four individual tasks making up the battery. The data from one participant who was unable to achieve this level of performance on the logical reasoning task (Data Analysis, below) were excluded from subsequent analyses involving that task.

Sleep on baseline Nights 2 and 3 was again recorded polygraphically. Parameters derived from the scored EEG records were averaged across the 2 nights to yield single baseline measures of sleep composition and quality. Two-night averages were chosen because of the well-documented night-to-night variability in sleep quality that older individuals frequently exhibit.[4,16,17] It was, therefore, felt that such an approach provided a more-accurate reflection of participants' typical sleep.

On the day between Nights 2 and 3, performance was again measured 2 hours after waketime and each subsequent 2 hours until five trials were completed. Performance measures obtained on this day were used to establish baseline performance levels. One hour after each PAB trial, participants underwent a 20-minute sleep latency test (SLT), in which they were asked to try to fall asleep while lying in bed in a darkened room. Detection of a sleep spindle or K-complex resulted in the termination of the SLT. These measures were employed to establish baseline sleepiness levels.

Between PAB trials and SLTs, participants were permitted to engage in leisure activities (e.g., reading, watching television or movies) within the laboratory area. They were not permitted to nap. Continuous EEG and closed-circuit television monitoring were used to ensure wakefulness. Meals and snacks were available as desired, except during the performance and sleepiness assessment intervals.

In-home Napping Phase

After the baseline laboratory visit, participants began the 4-week-long in-home phase of the study. They were randomly assigned to a 45-minute nap condition (short) or a 2-hour nap condition (long). They were instructed to nap at least 5 days per week and strongly encouraged to nap daily. They were further instructed to nap only one time per day and to complete their naps by no later than 6:00 p.m. Participants completed sleep logs twice a day (bedtime and waketime) throughout the 4-week interval.

Mid and End Laboratory Sessions Two subsequent laboratory sessions, 2 weeks and 4 weeks after the initial visit ended, were identical to the baseline session, with the exception that participants stayed in the laboratory for only 2 nights and the intervening day and that they napped on the day in the laboratory. The in-laboratory nap opportunity was scheduled to start 6.5 hours after wake-up from Night 1 and to continue for 45 minutes, or 2 hours, depending on group assignment. Scheduling the in-laboratory nap at the same time that each participant napped most frequently during the preceding 2 weeks at home or at their average nap time across the 2 weeks was considered. Allowing participants to self-select the time of in-laboratory naps, to mimic the in-home regimen, was also considered, but scheduling nap times based on time from waking had two important methodological advantages. First, this approach avoided the problem of missed performance trials by participants who might be napping at the time of a scheduled trial. Second, the approach permitted direct within-participant comparisons of performance and sleepiness data at baseline and after implementation of the napping regimen. That is, all participants completed four PAB trials and three SLTs that corresponded to the times of testing at their baseline laboratory session, anchored to morning wake time.

Data Analysis

Polysomnographic Sleep A trained sleep scorer scored all EEG records from nighttime and nap sleep in 30-second epochs according to Rechtschaffen and Kales criteria,[18] with the exception that no amplitude criterion was applied to evaluate slow-wave sleep (SWS) Stages 3 and 4.[19] The following variables were derived from these records: sleep onset latency (SOL—interval from bedtime until the first epoch of Stage 2, SWS, or REM); minutes and percentage of Stages 1, 2, SWS, and REM during the sleep period time (SPT—interval from sleep onset until scheduled wake time); minutes and percentage of wakefulness after sleep onset (WASO); total sleep time (TST—sum of minutes of all sleep stages); and sleep efficiency (SE), calculated as the ratio of TST during the SPT and as the ratio of TST during time in bed (TIB—interval from scheduled bedtime until scheduled wake time).

Neurobehavioral Performance The PAB comprised four tasks from the Automated Neuropsychiatric Assessment Metrics:[20] Logical Reasoning—Symbolic (LOG), Mathematical Processing (MTH), Sternberg 6-letter Memory Search (ST6), and 2-Choice Reaction Time (2CH). The measure of throughput (calculated as response time/accuracy) was used as the outcome variable for each task. For most analyses, a daily average throughput value was calculated for each participant for each session. To account for high interparticipant variability in absolute performance levels and to evaluate how performance changed across the study, the percentage change in throughput from baseline to mid sessions and from baseline to end sessions was determined for each participant. For analyses comparing performance before with performance after the nap at mid and end sessions, the absolute throughput value of the two prenap PAB trials were averaged, and the percentage change from the prenap baseline to the single postnap PAB trial was calculated for each participant. For correlations between performance and sleep measures (see below), absolute throughput levels were used.

Sleep Latency Tests As with performance measures, daily average sleep onset latency was calculated for the SLTs at baseline, mid, and end sessions. If a participant did not fall asleep during the 20-minute SLT, a latency of 20 minutes was assigned for that test.

Actigraphic Sleep Actiwatches (Mini-mitter Respironics, Inc., Bend, OR) were worn continuously during the 4-week in-home napping phase of the study. The following sleep variables for nighttime and nap sleep were obtained from the combination of actigraphy records and daily sleep logs: SPT, TST, and SE. (SOL was not reliably measured using actigraphy and sleep logs and was not calculated for nighttime or nap periods during the in-home study phase.) The actigraphy records were analyzed using the Actiware 5.0 algorithm, set at a medium threshold for sleep versus wake detection for all participants.

Statistical Analyses

Mixed and multivariate analyses of variance (ANOVAs) were used to compare measures between conditions and across laboratory sessions (e.g., nighttime PSG, nap PSG, performance) or across intervals of the in-home napping phase (e.g., actigraphy-derived sleep measures from baseline to mid). Post hoc comparisons were used to examine significant interactions. Possible relationships between sleep and performance measures were examined using Pearson product-moment correlations.


Nighttime Sleep in Laboratory (PSG)

Randomized assignment to the two nap conditions resulted in 11 participants each in the short and long groups. As shown in , there were no significant differences between the groups with respect to age, sex distribution, or baseline sleep and neurobehavioral performance measures.

Table 1.  Baseline Variables × Condition

Characteristic Short Long
Age, mean ± SD 72.3 ± 9.3 67.55 ± 10.3
Male/female, n 6/5 5/6
Bedtime 11:23 p.m. ± 56 minutes 12:02 a.m. ± 1 hour 15 minutes
Waketime 7:06 a.m. ± 48 minutes 7:45 a.m. ± 1 hour 15 minutes
Polysomnographic sleep
   Sleep efficiency, % (TST/SPT) 83.0 ± 9.9 82.0 ± 8.4
   Sleep onset latency, minutes* 16 ± 13 19 ± 12
   SPT, minutes 451 ± 68 445 ± 38
   TST, minutes 370 ± 49 364 ± 47
   Wakefulness after sleep onset, minutes (%), mean ± SD 80 ± 51 (17.0 ± 9.9) 81 ± 39 (18.0 ± 8.4)
   Stage 1, minutes (%), mean ± SD 24 ± 25 (5.4 ± 4.7) 25 ± 14 (5.5 ± 3.0)
   Stage 2, minutes (%), mean ± SD 181 ± 57 (40.0 ± 11.8) 172 ± 32 (38.3 ± 5.0)
   SWS, minutes (%), mean ± SD 84 ± 35 (19.4 ± 8.9) 85 ± 23 (19.3 ± 5.9)
   REM, minutes (%), mean ± SD 80 ± 43 (18.3 ± 10.2) 83 ± 23 (18.8 ± 4.6)

*To first epoch of Stage 2 sleep, slow wave sleep (SWS), or rapid eye movement (REM) sleep.
TST=total sleep time; SPT=sleep period time (from sleep onset to wake-up time); SD=standard deviation.

Neither long nor short naps had a significant effect on any measure of subsequent nighttime sleep (Figure 1). Postnap nighttime sleep at mid or end was not significantly different from baseline on any measure. SOL, SE, and TST remained essentially unchanged across laboratory sessions. Also, as at baseline, there were no significant differences between the long and short nap groups on any measure of nighttime sleep at mid or at end.

Figure 1.


Postnap nighttime polysomnographic sleep variables for short and long conditions at (A) mid and (B) end laboratory sessions. Sleep stage percentages, including wakefulness after sleep onset (WASO), are expressed as a percentage of the interval from sleep onset to wakeup time, or sleep period time (SPT). The duration of nighttime SPT varied between participants but did not differ between the short and long groups. A mixed multivariate analysis of variance (condition × session, including all sleep stage percentage variables). SWS= slow-wave sleep; REM=rapid eye movement; ns=no main effects for condition or session and no interaction effects.

Naps in Laboratory (PSG)

Neither the short nor the long group showed a significant change in nap sleep measures between the mid and end sessions (Figure 2). As would be expected, at the mid and end sessions, the long nap participants obtained significantly more average total sleep during their naps (TST) than did the short nappers (mid: 74 ± 32 minutes for long vs 17 ± 13 minutes for short, P<.001; end: 65 ± 36 minutes for long vs 20 ± 18 minutes for short, P<.001). As a consequence of longer sleep times, the long group also obtained significantly more minutes of REM sleep than the short group (mid: 14 ± 7 minutes long vs 0 ± 0 minutes short, P<.001; end: 11 ± 12 minutes long vs 1 ± 4 minutes short, P=02).

Figure 2.


Polysomnographic sleep variables from naps for short and long conditions at (A) mid and (B) end laboratory sessions. Minutes to sleep onset (sleep onset latency; SOL) and minutes of each sleep stage, including wakefulness after sleep onset (WASO) are presented. Nap opportunity duration was 45 minutes for short and 120 minutes for long. **P<.001, *P<.05. SWS= slow-wave sleep; REM=rapid eye movement; TST=total sleep time.

Mixed ANOVA (condition × session) revealed that the addition of nap TSTs to nighttime TSTs resulted in significantly greater TST per 24 hours (TST/24) for the long and the short nap groups combined at the mid and end laboratory sessions (Figure 3). Post hoc analyses indicated that average TST/24 of the short group increased significantly over baseline only at the end session (34.7 minutes; P<.001), whereas the long group showed significant increases at both laboratory sessions (87.5 minutes, P=.009, at mid; 70.3 minutes, P<.001, at end).

Figure 3.


Total sleep time per 24 hours (TST/24 hours). A mixed analysis of variance (condition × session) revealed a main effect for session (P=.01) and the interaction between condition and session (P=.04). Post hoc comparisons indicated that TST/24 hours in the long group increased significantly from baseline to mid and baseline to end, TST/24 hours in the Short group increased significantly from baseline to end only, and TST/24 hours in the long group was significantly greater than in the short group at the mid session (P=.04) but not at the end session.

Neurobehavioral Function The following results are reported for 21 participants (11 short, 10 long) because significant amounts of data from one participant were lost because of equipment difficulties.

Throughput on the PAB tasks at each session is shown in Figure 4. As was the case for nighttime sleep measures, no significant differences were found between long and short nappers at baseline on any of the tasks constituting the PAB. Performance on three of the four tasks improved significantly from baseline to mid and from baseline to end. There were also significant improvements on these three tasks from mid to end. Two-Choice Reaction Time remained essentially unchanged across the three laboratory sessions for the long and short nap groups, although there was a consistent (nonsignificant) trend for those in the long group to show greater improvement at the mid and end sessions.

Figure 4.


Throughput ([accuracy × speed] × 100) on four neurobehavioral performance tasks (A–D). Top portion of each graph shows percentage change from baseline for short (- - • - -) and long (—□—) conditions. Bottom portion of each graph shows absolute levels of throughput for short (open bars) and long (solid bars) conditions. Multivariate analyses of variance (ANOVAs) for condition × session including all four tasks revealed no main effect for condition, a significant main effect for session (Wilks lambda P=.003), and no condition × session interaction. Univariate ANOVAs indicated that throughput increased significantly across sessions for all but (D) Two Choice Reaction Time. Post hoc comparisons indicated that throughput increases were significant from baseline to mid, mid to end, and baseline to end on each of the other three tasks.

To determine possible explanations for the improvements in performance, relationships between changes in performance and sleep measures were conducted. These analyses revealed that there were no significant correlations between any performance measures and nap duration or 24-hour sleep amounts (nap+nighttime TST). Likewise, no significant correlations were found between performance and adherence (see below) during the at-home intervals of the study. With respect to the possible influence of sleep architecture on performance measures, improvement on the Logical Reasoning task at the end session was associated with minutes of SWS during naps (correlation coefficient=.64, P=.007). No other correlations between sleep stages and changes in performance were significant. Finally, there were no significant relationships between nap variables (TST, minutes of SWS, minutes of REM) and performance on the single PAB trial that occurred in the late afternoon and early evening hours, 2 hours after the nap opportunity.

Sleep Latency Tests Although SOL on the SLTs did not indicate that these participants were sleepy during the daytime hours, short and long naps increased SOL in both groups. Mean SOL at baseline for the short group was 14.0 ± 5.0 minutes, compared with 15.5 ± 3.5 minutes for the long group (P=.42). Average SOL for the all participants for the entire day (averaging the 2 prenap and 1 postnap SLTs) was significantly longer at the mid (18.0 ± 5.0 minutes) and end (18.5 ± 5.0 minutes) sessions than at baseline; these day averages did not differ between the short and long groups. Short and long nappers exhibited significantly longer SOLs on the SLT after the nap than the average of the two SLTs that preceded the nap (short mid prenap 14.0 ± 6.0 minutes vs postnap 18.0 ± 4.0 minutes, P<.001; long mid prenap 15.0 ± 5.5 minutes vs postnap: 20.0 ± 0 minutes, P<.001; short end prenap 13.0 ± 6.0 minutes vs postnap 16.5+6.5 minutes, P<.001; long end prenap: 14.5 ± 5.0 minutes vs postnap 20.0 ± 1.0 minutes, P<.001). In the long group, no participants fell asleep during the postnap SLT at mid, and only two fell asleep during the postnap SLT at end.

Naps and Nighttime Sleep at Home (Actigraphy and Logs) According to actigraphy records (combined with sleep log information), participants in the short and long groups napped approximately five times per week (short: 5.11 ± 1.23 naps/wk vs long: 5.17 ± 1.24 naps/wk, P=.91). Although the group averages for number of naps per week were greater than 5 in both groups, some participants napped every day, whereas others napped much less frequently (see Adherence, below). The mean duration of in-home naps was 58 ± 29 minutes for the short group, compared with 95 ± 35 minutes for the long group (P=.01). Sleep efficiency of the at-home naps averaged 60 ± 16% for the short, compared with 56 ± 20% for the long groups (P=.61); these sleep efficiencies were similar to PSG naps in the laboratory sessions.

While at home, the average nap start time was 3:07 p.m. ± 2:09 hours for the short group and 2:43 p.m. ± 1:32 hours for the long group (P=.62). These clock times were not significantly different from the clock times at which nap opportunities were scheduled in the laboratory (6.5 hours after T min), which were 1:37 p.m. ± 0:52 hours for the short group and 2:20 p.m. ± 1:36 hours for the long group.

To examine whether any "mismatch" between at-home nap timing and in-laboratory nap timing affected the nap, the difference between these times was calculated for each participant, separately for mid and end sessions, and then correlated with nap variables. The average mismatch was –0:59 (i.e., the nap at home was initiated, on average, an hour later than the scheduled laboratory nap). The mismatch amounts did not differ between the short and long groups and were not associated with any nap variables (all P>.10 for correlations).

To examine whether naps taken at home affected nighttime sleep at home, the duration and SE of nighttime sleep were compared for nights after naps with those of nights after a day when no nap was taken. These analyses confirmed that naps did not negatively affect nighttime sleep. Combining the short and long groups, postnap nighttime sleep duration averaged 7.6 ± 1.2 hours, versus 8.1 ± 1.5 hours after a nap-free day (P=.40). Postnap nighttime SE averaged 76 ± 8%, versus 78 ± 8% after a nap-free day (P=.56). Moreover, there were no significant relationships between nap duration or nap SE and sleep measures on the subsequent night.

Adherence Adherence to the napping protocol was operationally defined in three different dimensions: frequency, duration, and timing. Participants were considered adherent in terms of frequency if they averaged at least five naps per week, in terms of duration if at least 80% of naps were within 15 minutes of the prescribed length (45 minutes or 2 hours), and in terms of timing if 90% of naps were initiated between 10:00 a.m. and 6:00 p.m. A participant was considered "super adherent" if all three criteria were met.

Adherence was assessed based on actigraphy data supplemented by information provided by participants' daily sleep logs. The following results are based on 21 participants, because one participant in the long nap group did not complete daily logs and did not consistently wear an Actigraph during the home portion of the study.

The number of participants meeting the adherence criterion for frequency did not differ between groups (chi-square (χ2)=2.39, P=.12). Eighty-two percent (9/11) of participants in the short group were adherent for frequency (5.1 ± 1.2 naps per week) and 50% (5/10) in the long group (5.2 ± 1.2 naps per week). Participants in the short group showed a nonsignificant tendency toward better adherence in terms of nap duration, although both groups demonstrated relatively poor adherence (45% vs 30%; χ2=0.53, P=.47). In terms of timing, the long group showed a tendency, again nonsignificant, toward better adherence, with 80% of participants consistently initiating their naps within the prescribed window, compared with 45% of short nappers (χ2=2.65, P=.10). Two participants in each group met criteria for "super adherence."


These results support and extend previous findings[10] that a daytime nap may improve neurobehavioral functioning in healthy older adults without negatively affecting subsequent nighttime sleep. Whereas the previous study examined the effects of a single, 2-hour nap opportunity on performance and nighttime sleep quality, the current study focused on the effect of a longer-term napping regimen and on the relative effectiveness of a short (45 minute) and a longer (2 hour) nap opportunity. Because the study was conducted to examine the effectiveness of napping as a possible countermeasure against some of the negative consequences of age-related changes in sleep, there was also interest in the degree to which participants could adhere to the napping regimen.

Sleep and Alertness

There were few differences between the short and long nap groups in terms of their effect on nighttime sleep and daytime sleepiness. Short and long nappers showed significant increases in 24-hour sleep amounts over their baseline sleep times, and neither the short nor long nap had a negative effect on nighttime sleep: SOL, SE, and sleep architecture remained essentially unchanged across the study. This finding is in agreement with a majority of studies that have examined the relationship between napping and subsequent nighttime sleep quality in older individuals[6–11,21–25] and argues against the frequently expressed notion that naps should be avoided.

Greater total sleep time per 24 hours resulted in less daytime sleepiness, as reflected in mid and end laboratory sessions compared with baseline, as well as within the mid and end sessions when pre- and postnap latencies were compared.

Neurobehavioral Function

In both groups, performance on three of the four tasks constituting the PAB showed significant improvement, with simple reaction time showing no change. There were no significant differences in the degree of improvement between the groups, although the long group showed a consistent tendency for greater improvement. Significant improvements in performance were observed not only between baseline and each subsequent laboratory session, but also between the mid and end sessions. That is, performance continued to improve across the entire study.

These findings are in contrast to those of a previous study[11] that reported no significant effect of napping on performance, although performance was measured in laboratory sessions, during which total 24-hour sleep amounts between the nap and no-nap conditions did not differ. This difference, the dissimilar average age of participants studied, or perhaps differences in the performance tasks employed may help explain the discrepancy between studies.

The findings of the current study may indicate that the beneficial effects of napping on neurobehavioral function may not be maximized even after a month of scheduled naps. It is conceivable that the negative effects of years of chronic sleep restriction associated with age-related truncation of nighttime sleep may require a longer time to be fully mitigated.

Alternatively, the results may be viewed as simply reflecting a practice effect. Despite the fact that participants were thoroughly trained on the PAB and were required to reach strict criteria demonstrating asymptotic accuracy levels on each of the tasks before baseline testing (see Procedures), the possibility that participants simply improved because of repeated exposure to the PAB cannot be excluded. Perhaps adding support to this interpretation is the finding that changes in performance were not associated with greater sleep amounts or the composition of nighttime sleep or naps (with the exception of a significant correlation between SWS in the end nap with enhanced performance on the Logical Reasoning task). Likewise, no measure of adherence was related to change in performance.

Adherence Based on the operational definition of adherence, which considered frequency, timing, and duration of naps, most older individuals would have difficulty incorporating a daily 2-hour nap into their schedules; only half of the participants in the long group napped at least five times per week, and only 30% showed average nap durations within 15 minutes of the assigned duration, with most naps failing to reach minimum length. Although the short nap group showed better adherence with regard to frequency, these participants also showed poor adherence in terms of nap duration, with a large majority exceeding the maximum length. Thus, in terms of adherence to a nap regimen, it appears that a 2-hour nap opportunity may be too long, whereas a 45-minute window may be too short.


Because neurobehavioral function was assessed only at the three laboratory visits (at baseline and after 2 and after 4 weeks of napping), relatively infrequent "snapshots" of effects of napping could be evaluated. It is impossible to determine, therefore, whether the improvements observed throughout the entire protocol would have continued across additional days and weeks of napping or whether the positive effects on cognitive performance were attenuated, or leveled out, at some point between the second and fourth weeks of the napping regimen.

Another limitation of the study, with respect to generalizing the findings to other older samples, involves the overall health, in general, of the participants and their sleep quality, in particular. All of the participants were in good physical and emotional health, and although all reported some degree of age-related sleep disturbance (e.g., early-morning awakening, disrupted nighttime sleep), none suffered from a diagnosed sleep disorder, including significant insomnia. Thus, it is unclear whether a napping regimen would be beneficial to older individuals with chronic illnesses or sleep disorders.

In summary, these findings add to a still-limited, but growing body of literature suggesting that a daily afternoon nap may be a safe and effective means of increasing 24-hour sleep total and, in so doing, improve waking function. Although a practice effect could not be excluded, these data support a previous finding that enhanced neurobehavioral functioning, as well as less daytime sleepiness, accompany greater 24-hour sleep amounts, with little negative effect on nighttime sleep. Although there were few differences between the 45-minute and 2-hour naps in terms of their effect on nighttime sleep and daytime functioning, it is clear from the adherence results that a regimen featuring the longer nap would not be widely accepted. From the current data, it is not possible to evaluate whether naps shorter than 45 minutes would also be associated with enhanced waking function, although at least one study has reported beneficial effects, in a small group of habitual nappers, of a 30-minute nap.[7] In this regard, the participants of the current study, who were not habitual nappers before the study, did not respond to the instructions to nap daily by taking "power naps" at home. Rather, the majority of participants in the short group preferred a nap of approximately an hour in duration.


  1. Buysse DJ, Reynolds CF, Monk TH et al. Quantification of subjective sleep quality in healthy elderly men and women using the Pittsburgh Sleep Quality Index (PSQI). Sleep 1991;14:331–338.

  2. Foley DJ, Monjan AA, Brown SL et al. Sleep complaints among elderly persons: An epidemiologic study of three communities. Sleep 1995;18:425–432.

  3. Campbell SS, Dawson D, Anderson MW. Alleviation of sleep maintenance insomnia with timed exposure to bright light. J Am Geriatr Soc 1993;41:829–836.

  4. Webb WB. Sleep in older persons: Sleep structures of 50- to 60-year-old men and women. J Gerontol 1982;37:581–586.

  5. Dijk DJ, Duffy JF. Circadian regulation of human sleep and age-related changes in its timing, consolidation and EEG characteristics. Ann Med 1999;31:130–140.

  6. Tamaki M, Shirota A, Tanaka H et al. Effects of a daytime nap in the aged. Psychiatry Clin Neurosci 1999;53:273–275.

  7. Tamaki M, Shirota A, HayashiMet al. Restorative effects of a short afternoon nap (o30 minutes) in the elderly on subjective mood, performance and EEG activity. Sleep Res Online 2000;3:131–139.

  8. Tanaka H, Taira K, Arakawa M et al. Effects of short nap and exercise on elderly people having difficulty in sleeping. Psychiatry Clin Neurosci 2001;55:173–174.

  9. Tanaka H, Taira K, Arakawa M et al. Short naps and exercise improve sleep quality and mental health in the elderly. Psychiatry Clin Neurosci 2002;56:233–234.

  10. Campbell SS, Murphy PJ, Stauble TN. Effects of a nap on nighttime sleep and waking function in older subjects. J Am Geriatr Soc 2005;53:48–53.

  11. Monk TH, Buysse DJ, Carrier J et al. Effects of afternoon "siesta" naps on sleep, alertness, performance, and circadian rhythms in the elderly. Sleep 2001;24:680–687.

  12. Hamilton M. A rating scale for depression. J Neurol Neurosurg Psychiatry 1960;23:56–62.

  13. Douglass AB, Bornstein R, Nino-Murcia G et al. The Sleep Disorders Questionnaire. I: Creation and multivariate structure of SDQ. Sleep 1994;17:160–167.

  14. Johns MW. A new method for measuring daytime sleepiness: The Epworth Sleepiness Scale. Sleep 1991;14:540–545.

  15. Abetz L, Arbuckle R, Allen RP et al. The reliability, validity and responsiveness of the International Restless Legs Syndrome Study Group rating scale and subscales in a clinical-trial setting. Sleep Med 2006;7:340–349.

  16. Dement WC, Miles LE, Carskadon MA. "White paper" on sleep and aging. J Am Geriatr Soc 1982;30:25–50.

  17. Vitiello M. Sleep in normal aging. Sleep Med Clin 2006;1:171–176.

  18. Rechtschaffen A, Kales A. AManual of Standardized Terminology, Techniques and Scoring System for Sleep Stages of Human Subjects. Washington, DC: National Institute of Health, 1968.

  19. Webb WB, Dreblow LM. A modified method for scoring slow wave sleep of older subjects. Sleep 1982;5:195–199.

  20. Reeves DL, Winter KP, Bleiberg J et al. ANAM genogram: Historical perspectives, description, and current endeavors. Arch Clin Neuropsychol 2007; 22(Suppl 1):S15–S37.

  21. TakahashiM. The role of prescribed napping in sleepmedicine. Sleep Med Rev 2003;7:227–235.

  22. Evans FJ, Cook MR, Cohen HD et al. Appetitive and replacement naps: EEG and behavior. Science 1977;197:687–689.

  23. Buysse DJ, Browman KE, Monk TH et al. Napping and 24-hour sleep/wake patterns in healthy elderly and young adults. J Am Geriatr Soc 1992;40:779–786.

  24. Hayashi M, Ito S, Hori T. The effects of a 20-minutes nap at noon on sleepiness, performance and EEG activity. Int J Psychophysiol 1999;32:173–180.

  25. Yoon I, Kripke DF, Elliott JA et al. Naps and circadian rhythms in postmenopausal women. J. Gerontol. A Biol Sci Med Sci 2004;59A:844–848.