Uninterrupted Sleep, Staying Up All Night, and Memory Impairment

This post explores research literature to understand issues arising in a sleep-related situation that I experienced.

First, the setting. I lived in San Antonio. According to Accuweather, the historical average daytime high/low temperatures in San Antonio reached 90/70°F on May 25 of a typical year, remained at a peak of 98/74°F from August 2 to 22, and did not drop below 90/70°F until September 16. Of course, some people found even 90/70°F too hot. For instance, if I wished to run at a strenuous pace, I found that I needed a temperature below 60°F to avoid undesirable cardiac symptoms.

Thus, for me, the hot season would last for half of the year, because the average daytime low (experienced by running in the early morning) was at or above 60°F from April 19 to October 20. In the hot season, I could run at a slower pace, but knee problems suggested that running was best mixed with other exercise in any case. Since the experts recommended cross-training, and since I had unfortunately found that psychopathic Texas drivers made bicycling dangerous at any hour, walking was my primary alternative form of aerobic exercise. To get a decent workout, and to avoid heat, traffic, and distractions, I preferred to walk for several hours in the middle of the night. I found that San Antonio could actually be rather pleasant, or at least its heat and humidity less intolerable, at 3 AM.

Thus I found myself gravitating toward a nocturnal schedule. The times of going to bed and getting up varied, but typically I was sleeping around five hours, from maybe 5 to 10 AM, and then having two to three naps during the balance of the day, ranging from 15 minutes to three hours each. To remain acclimated, I used fans, but did not use air conditioning at any time, over that period of several years. The question was whether this sleep schedule was healthy.


Healthy Sleep
Threats to Healthy Sleep
Consequences of Unhealthy Sleep
Interpretation and Conclusion


Healthy Sleep

There seemed to be several possible definitions of “healthy” sleep. For many working people, the best that one could hope for was enough sleep to get through the day. During my working career, that tended to mean having a job where I could take a nap in midafternoon, because I rarely experienced full nights of sleep amid the anxieties of the workplace, the vagaries of social life, and the effects of drinking caffeine and a moderate amount of wine (which, at the time, I thought was healthy and helpful for purposes of sleeping well).

Overall, my sleep pattern did not seem to mesh with the concept of healthy sleep preferred by sleep experts. MedlinePlus and others said that adults should get seven to eight hours of uninterrupted sleep. NSF said,

On average, adults should optimally receive between seven and nine hours of sleep each night, but those needs vary individually. For example, some people feel best with eight consecutive hours of sleep, while others do well with six to seven hours at night and daytime napping.

It seemed that sleep interruptions made a big difference. Mendonca et al. (2019, p. 24528) defined sleep fragmentation as including “arousals, awakenings and transitions from deep to light sleep.” The Tuck mattress company (2018) distinguished micro, brief, and full arousals or awakenings. A microarousal, they said, was a brief awakening that you wouldn’t remember. A brief arousal was a brief awakening that you would remember. A full arousal would be actually staying awake for a while. Some said that full awakening in the middle of the night was not necessarily bad if it was biphasic — if, that is, it divided two sleep phases lasting at least three to four hours each. As I noted in a previous post, sleep historians had found evidence that it was common, especially in northerly latitudes in the pre-electrical era, to go to sleep after it got dark outside, get up for several hours in the middle of the night, and then go back to sleep for several more hours until dawn.

I didn’t get into the question of whether those historical peoples lived long enough to experience memory problems in old age due to their lifetimes of biphasic sleep. I also wasn’t sure whether getting up to pee in the middle of the night, and then going back to sleep, would count as a brief arousal. The advice on that was to avoid drinking liquids for several hours before bedtime — especially including liquids that would increase the need to pee, such as alcoholic and caffeinated beverages. Regarding NREM and SWS sleep phases discussed below, Mander et al. (2017) said that men experience “far greater relative disruption and impairment in NREM sleep than women later in life,” that men aged 70+ (but not women) experience a 50% reduction in SWS compared to their peers aged 55, and that older men experience greater sleep fragmentation and higher daytime nap propensity — and yet, paradoxically, despite these aging men’s more severe reductions in sleep quantity and quality, aging women were reportedly more likely to register complaints of poor sleep. Somewhere, I saw a suggestion that this difference could result from a greater sensitivity toward sleep deprivation on the part of women.

Another practical way of identifying fragmented sleep, presented by the NSF, was to ask whether the person had “long periods of lying awake when you wish to be sleeping” or other signs of disrupted sleep (e.g., snoring, pauses in breathing, restlessness). Baron et al. (2017) calculated a Sleep Fragmentation Index by counting periods of excessive wrist movement (as measured by a wristwatch-like device) during sleep. They found that results produced by this measurement were comparable to research participants’ own subjective reports. In other words, sleepers who had a lot of wrist movement during the night were more likely to say, the next morning, that they did not feel well-rested. This sort of research seemed to be included in what Mendonca et al. (2019) described as analysis of the sleep microstructure, which often used brainwave (i.e., electroencephalogram (EEG)) measurements to identify periods of rest and movement, sometimes lasting less than 30 seconds.

Mendonca et al. (p. 24528) found, however, that most sleep studies are instead interested in the sleep macrostructure, defined as a series of sleep phases, measured in minutes. Full-night polysomnography (PSG) was the ideal tool for identifying the start and end of such phases, using many factors (e.g., EEG, ECG, EOG, EMG, body position, respiratory movement, breath airflow, oxygen saturation). WebMD (2018) said that the sleep macrostructure typically consists of cycles of sleep with and without rapid eye movements (REM) under one’s eyelids. Wikipedia described REM sleep as a phase in which birds and mammals experience atonia (i.e., loss of muscle strength; extreme relaxation to the point of paralysis) and vivid dreams. By contrast, Wikipedia said, non-REM (NREM) or “quiescent” sleep entails little to no dreaming, REM, or atonia. Unlike the bizarre mental content of REM sleep, non-REM sleep features more normal and organized mental activity. VeryWellHealth (Cherry, 2019) and others described NREM sleep as consisting of three (until 2007, four) stages:

  • NREM Stage 1: relaxed wakefulness, lasting five to ten minutes. May begin with a hypnogogic (a/k/a myoclonic or hypnic) jerk (i.e., an involuntary muscle contraction). People awakened at this point may think they were already awake.
  • NREM Stage 2: a period of about 20 minutes of being asleep, but easily awakened, with reduced body temperature and more regular breathing and heartrate. Short bursts of brain activity. Accounts for about 50% of total sleep time.
  • NREM Stage 3: deep (a/k/a delta or) slow-wave sleep (SWS), detected in slow-wave (brain) activity (SWA). Some dreaming — but disconnected, less vivid, less memorable. More present in the first third of the night. Accounts for about 10% of total sleep time. Most common time for parasomnias (e.g., sleepwalking, sleep terrors, bedwetting). Hardest to awaken from; groggy when awakened. Rebound into this stage after sleep deprivation suggests a particular need for this stage.

According to WebMD (2018), sleepers typically go through the three NREM stages first, and then reach REM sleep about 90 minutes after falling asleep, and remain in REM sleep for about ten minutes. But VeryWellHealth (Cherry, 2019) said the actual sequence is typically more like NREM 1, NREM 2, NREM 3, then go back to NREM 2, then REM, then NREM 2 again, then back to NREM 1. Then the NREM-REM cycle is repeated, for a total of four or five cycles per full night’s sleep. Each subsequent REM phase lasts longer — up to an hour, in the final cycle. Medscape (Stevens, 2015) explained that the first full NREM-REM cycle totals 70 to 100 minutes, while the last one is more like 90 to 120 minutes. Altogether, adults spend about 20% of their sleep in REM sleep. People awakened from REM sleep typically feel refreshed, but the night’s sleep is not entirely complete until the sleeper returns to NREM 1 and finally awakens naturally. It wasn’t clear how much interruption the brain would tolerate, before it dropped out of its current sleep phase and went back to start over at NREM Stage 1.


It was not clear what to think about napping. On one hand, Mander et al. (2017) seemed to say that napping was unusual. They indicated that, even among adults aged 55-64, only 10% reported taking naps in the daytime — though, of course, people of that age would still tend to be in the workforce and subject to its limits on napping. On the other hand, it seemed like I had encountered a number of people who felt a strong need for naps, and quite a few who enjoyed them. My dad had always napped in the daytime, though I don’t know whether he would have done so if he hadn’t worked on the railroad, with its calls to work at all hours of day and night. The National Sleep Foundation (NSF) seemed to feel that there was nothing wrong with a nap of 20-30 minutes — indeed, that it could be lifesaving, for the tired driver. But NSF said that a nap — especially one exceeding 30 minutes — could also leave people with sleep inertia (i.e., grogginess and disorientation), could increase insomnia, and could be problematic for people at risk of heart failure.

Thus it appeared that shorter nights of sleep, with long daytime naps as a cause and/or compensation, would not be ideal because there would be no opportunity for the person to reach the point of experiencing the later cycles, with their longer REM phases. But if circumstances made daytime napping unavoidable, it appeared that long daytime naps could help to a limited extent. The Guardian (Ackerman, 2009) asserted that long naps (i.e., 90-120 minutes) in the morning yielded more REM sleep, while those in the afternoon yielded more SWS. Agreeing with that, Mednick offered an interactive Take a Nap Wheel (requires Adobe Flash; hit Ctrl-+ a few times to make it large enough to read) that purported to indicate the best time for a nap maximizing REM or SWS.

Sleep.org agreed that a 90-minute nap could be long enough to permit a full daytime NREM-REM sleep cycle, and would thus boost memory and creativity. I liked those claims, because 90 minutes happened to be the approximate length of some of my naps. Indeed, some were not quite twice as long, especially in my nocturnal schedule. One could contend that one’s mind and body know what they need, especially where napping has been a lifelong tendency. Indeed, if (as indicated above) SWS is more present in the first third of the night, it seemed that a strategy to maximize SWS (albeit at the expense of REM sleep) could include one or two long naps along with the main period of, say, five hours of (usually nighttime) sleep. The point wasn’t just that the total would add up to seven or eight hours; it was that each 24-hour period would thus include several instances of the SWS-heavy sleep provided in the first third of the night.

But NSF didn’t like this. They said, “Sleeping for an hour or more is too much during the day and will likely set you up for nighttime troubles.” The troubles they had in mind were, apparently, that I would sleep less at night. Merck agreed: “Patients who compensate for lost sleep by sleeping late or by napping further fragment nocturnal sleep.” And I believed them; it was just that I had never been able to do what they wanted me to do. Going short on naps in my working years hadn’t made me a good nighttime sleeper back then, and I sure didn’t seem to be tending in that direction in my later years.

I did want to progress toward what NSF considered ideal. At the same time, I contemplated indications that NREM sleep quantity and quality is important for memory processing. For instance, according to Mander et al. (2017), “Sleep after initial encoding supports the subsequent long-term consolidation of recent hippocampal-dependent memory representations.” I did wonder, that is, whether a daytime nap after learning something, long enough to allow for a complete and uninterrupted NREM-REM cycle, and timed (by my own fatigue, rather than conscious planning) to maximize NREM, might actually be better, for purposes of memory consolidation, than nighttime sleep occurring hours later. Supporting that thought, Baran et al. (2016) found that the mind does work with memories even during a brief nap after a learning session, though the mechanism appears to be different and less effective in older than in younger adults.

I thought my views on napping would be supported by evolutionary psychology and anthropology. That is, I assumed that, during the tens of thousands of years of tribal living that still shape human life, napping must have been pretty common. It is not clear that’s true. There is the Sentinel Hypothesis which, according to Psychology Today (Pardi, 2017), proposes that teenagers sleep late while old people get up early because this made sure that someone in the tribe was always awake, standing guard against potential threats (New Scientist, 2017). One might expect that older people would then want a nap in midday. But in at least some of the few remaining hunter-gatherer tribes (mostly located in the tropics), it appears napping is infrequent. Then again, biphasic sleep is also infrequent among such tribes, even though they do apparently tend to awaken for a while, once or more, during the night. It could be that my people, hailing from the ice caves of the far north, did find biphasic sleep and midday naps salutary, for reasons that would not add up in the land of the tiger.

In any case, APA (Weir, 2016) cites research suggesting that naps provide great benefits, especially for people who are inclined to nap: it seems “people who choose to nap regularly are predisposed to get more out of it.” In this scenario, the problem is just that the non-napping majority mistakenly believe that our needs must be like theirs. We might retort that they are too manic and caffeinated to see that, actually, they are wrong about naps. For instance, Liu et al. (2019) found that children who nap 30 to 60 minutes experience greater happiness, self-control, and grit, higher IQ, and fewer behavioral problems.

Threats to Healthy Sleep

Wikipedia said that there were three distinct classifications of sleep disorders. Two were oriented toward psychiatry: the World Health Organization’s International Classification of Diseases (ICD) and the American Psychiatric Association’s (APA) Diagnostic and Statistical Manual, fifth edition (DSM-5), available for ~$100 from Amazon). Wikipedia said those two tended to combine distinct disorders into single categories, while the International Classification of Sleep Disorders (ICSD) produced by the American Academy of Sleep Medicine (AASM) tended to separate them.

Merck’s summary of the DSM indicated that poor sleep hygiene was one cause of poor sleep. Within the area of sleep hygiene, numerous sources (e.g., the NSF, the AASM, the Mayo Clinic) listed measures that would reportedly help people to fall asleep and stay asleep. Their lists included these: avoid substances (e.g., caffeine, alcohol, nicotine) that can disrupt sleep; create a restful (e.g., cool, dark, quiet, maybe with earplugs or a sound conditioner) sleep environment; develop a routine of calming activities just before bedtime (e.g., take a bath, meditate); avoid daytime napping, especially naps that are long or late in the day; engage in exercise (but not too close to bedtime); resolve worries before bedtime, or make notes and set them aside until tomorrow; maintain a consistent sleep schedule, so your body can develop a natural sleep-wake cycle; don’t go to bed until you are sleepy, and get up if you haven’t fallen asleep after 20 minutes; use your bed only for sleep and sex; stop eating and drinking three hours before bedtime, and be particularly cautious with foods that could trigger indigestion or heartburn; turn off electronic devices at least a half-hour before bedtime; and avoid excitement (e.g., disturbing movies) near bedtime. While these all sounded good, at a certain point a person might sympathize with Wignall’s objection that these lists of tips and techniques could become cumulatively oppressive — that, in many lives, some of them would not be feasible, leaving the person with worry or guilt that could actually make it harder to sleep.

That objection seemed consistent with psychophysiologic insomnia (i.e., being unable to sleep because you’re worried about being unable to fall asleep). That was one of several sleep problems, identified by Merck, that I had experienced at one point or other. Others included physical problems that could impair sleep (e.g., broken collarbone, poison ivy, leg cramps from distance running); insufficient sleep syndrome (i.e., social or work commitments impinging upon sleep hours); and sleep disruption caused by depression, anxiety, or acute emotional stress due to disruptive events (e.g., job loss, divorce). These did not seem to be especially relevant to my present situation. I did think, however, that light exposure could make a difference. Some of the foregoing sources recommended limiting exposure to bright lights in the evening (see e.g., f.lux), and other sources (e.g., HelpGuide) encouraged getting a few hours of exposure to bright sunlight during the daytime. I had heard that a few days of camping could bring a person’s circadian rhythm back to normal, and I did wonder whether my own situation was compounded by the fact that I had long been working in a room facing north, where I could see sunshine on the trees and ground outside, but could not see the sun itself.

Wikipedia said the 2014 revision of the ICSD identified 83 sleep disorders, divided into seven main categories: insomnia, sleep-related breathing disorders (e.g., sleep apnea), hypersomnolence disorders (especially narcolepsy and hypersomnia), circadian rhythm disorders, parasomnias, sleep-related movement disorders, and other (i.e., atypical) sleep-related disorders. Hypersomnia was the persistent or recurrent tendencies to sleep too much. Narcolepsy (i.e., excessive sleepiness) could feature such things as hallucinations, cataplexy (i.e., suddenly feeling paralyzed or weak, especially after excitement or laughing), falling asleep during normal activities, and a tendency to go into REM sleep within ten minutes. Parasomnias included abnormal sleep-related behaviors, emotions, perceptions, and dreams. Among the various sleep-related movement disorders, the only things I had experienced were the occasional hypnic jerk and bruxism (i.e., tooth grinding), the latter remedied by a plastic tooth block purchased in the sporting goods section of Walmart for one dollar. I had audiotaped myself sleeping, and concluded that I didn’t seem to have any breathing disorders. Really, among those seven ICSD categories, my problems mostly seemed to involve insomnia and circadian rhythm issues.

Under the ICSD, Wikipedia listed seven circadian rhythm issues, and those could be summarized as phase disorders, rhythm disorders, shift work disorder, jet lag disorder, and the catch-all NOS (i.e., not otherwise specified) circadian disorder. These seemed to be reducible to a few primary categories:

  • Phase disorders could be either delayed or advanced — that is, I could be falling asleep and wanting to wake up later, or earlier, than America preferred. Proffered examples included bedtimes of roughly 4 AM vs. 7 PM, respectively. Both could have a genetic origin, but the delayed disorder would typically become visible by adolescence, whereas the advanced version might not become evident until middle or late adulthood. I did seem to have a delayed phase disorder, but I wasn’t sure I would qualify for the full diagnosis: I hadn’t heard of it affecting other family members, and it had only emerged under these San Antonio conditions.
  • Rhythm disorders included irregular and non-24-hour varieties. The former, apparently quite rare, entailed no main (typically nighttime) sleep episode, numerous naps throughout day and night, and no particular pattern from one day to the next. The non-24-hour version usually involved free-running sleep, in which the body acted as if the day was longer (or, very rarely, shorter) than 24 hours. In other words, the person might go to bed at 6 AM today, 7 or 8 AM tomorrow or the next day, and even later next week.
  • Extrinsic disorders. The ICSD considered both phase and rhythm disorders (above) to be “intrinsic” circadian disorders, insofar as the problem was due to the person’s inner clock: it was out of sync with the classic 9 AM to 5 PM working schedule of American life. By contrast, in an “extrinsic” circadian disorder, the person’s inner clock was in sync with the nine-to-five; the problem was that the workplace was expecting the person to handle something different. Shift work sleep disorder (SWSD) was apparently the most common example, arising when the person’s work hours overlapped with his/her natural sleep hours. (A radical thinker might suggest that SWSD could be considered extrinsic to the person because the actual disorder was in the organization or the economy requiring this unnatural scheduling, from a worker who just wasn’t built for it.) SWSD entailed elevated risks for cancer, heart disease, digestive disorders, menstrual irregularities, hazardous driving, depression, cognitive performance, and so forth. Jet lag was another kind of extrinsic circadian disorder.

Circadian disorders would commonly entail insomnia of one form or another. Sources (i.e., Wikipedia, NCBI, and Medscape) said the DSM-5 essentially defined primary insomnia as difficulty falling asleep, returning to sleep after interruption, or staying asleep until an appropriate rising time, occurring at least three nights per week for at least three months, resulting in significant distress or impairment in social, occupational, behavioral, or other important areas of functioning. It wouldn’t be insomnia if it was due to inadequate opportunity for sleep, or if it was caused by drugs, medication, some other sleep disorder (e.g., parasomnia, breathing disorder, circadian disorder), or other mental disorder or medical condition.

It seemed I would qualify for primary insomnia, because for many years I had been getting significantly less than seven hours of sleep, most nights, not because of any of those other causes (e.g., inadequate opportunity to sleep, drugs, breathing disorder), and regardless of what was happening with my circadian rhythm (i.e., regardless of when I felt like going to bed).

Aside from primary insomnia, the diagnostic sources allowed for some variants. Transient insomnia apparently lasted less than a week, and resembled sleep deprivation. Acute (a/k/a short-term or stress-related) insomnia lasted for less than a month. Chronic insomnia was like transient insomnia, but it lasted longer than a month and, not surprisingly, it could result in hallucinations, double vision, and muscle weariness. Adjustment insomnia could be a response to acute emotional stress (e.g., job loss), and would typically fade after stressors were removed.

My desire to nap would apparently not qualify as Excessive Daytime Sleepiness (EDS). Wikipedia said EDS was “characterized by persistent sleepiness and often a general lack of energy, even during the day after apparently adequate or even prolonged nighttime sleep.” Stanford Health Care said that EDS was the core symptom of hypersomnias and narcolepsy. The Epworth Sleepiness Scale (ESS) and the Multiple Sleep Latency Test (MSLT) were used to assess EDS, narcolepsy, and hypersomnia. SleepEducation.org explained that the MSLT would take all day, whereas I was able to use Lauren Clowney’s online version of the ESS within a few minutes. It didn’t ask for responses during a day when I’d had adequate sleep the night before; it just asked for my ordinary responses, which would vary according to whether I’d had a nap anytime recently.

To summarize, there seemed to be many possible threats to healthy sleep. Sleep hygiene could make a difference. There were some areas of improvement for me there. But even under the best circumstances, I seemed to have some persisting problems. For one thing, I seemed to be experiencing a circadian rhythm disorder (specifically, of the delayed phase variety), but it wasn’t a big worry because I didn’t have to get up at any particular hour, and I believed and hoped I could go back to normal sleep hours when San Antonio’s hot season ended. That seemed reasonable: According to Wikipedia, “[H]aving a preference for extreme early or late wake times are not related to a circadian rhythm sleep disorder diagnosis. There must be distinct impairment of biological rhythms that affects the person’s desired work and daily behavior.”

In addition, I seemed to have primary insomnia, because I didn’t have a good excuse for my difficulty in staying asleep for seven or more hours per night. That is, I wasn’t deprived of sufficient sleep opportunity; my lack of sleep wasn’t due to any physical or mental problem, nor to medication or drugs; and so forth.

It didn’t seem that sleep experts would be highly enthusiastic about the theory that I was actually tending toward a biphasic sleep schedule, like those ancient people, with four or five hours of sleep, a few hours of being awake, and then a three-hour nap. I also wasn’t sure how far I’d get with an argument that I couldn’t have primary insomnia because going short on nighttime sleep was not causing me significant distress or impairment in social, occupational, behavioral, or other important areas of functioning. True, I could get away with long naps because I was retired. But there was an additional risk area I hadn’t previously worried too much about. It appeared that, even if things seemed fine now, I might be working toward significant problems that might only become obvious much later, as long-term consequences of short sleep.

Consequences of Unhealthy Sleep

I had a funny experience one time. After my divorce in 2002, I went back to school at the University of Missouri. The divorce experience impelled me to start seeing a student counselor, a PhD student working in the University’s counseling center. Students also had coverage for a few visits to a psychiatrist. I had already been reading in the DSM for some time. I wanted whatever help was available, and I was also curious how the psychiatrist’s perspective might differ from that of the student counselor I had been seeing.

What was funny about that was that, when I told the psychiatrist that I had been going short on sleep for the past ten years or more, his response was to wonder whether maybe I was bipolar (i.e., manic-depressive). This guy was a big wheel, head of the psychiatry department as I recall. I still remember him wondering, aloud, whether a person could have a manic phase lasting ten years. I don’t remember if my mouth dropped open. Any freshman student, reading the DSM, would know that mania involves feelings of energy, invincibility, being on top of the world. It does not involve feeling tired, needing naps, often feeling like you haven’t had enough sleep — which was what I had just described to him. I did follow up with another psychiatrist, and also with two psychologists there in the university system, to confirm that there was simply no reason to think that I might be manic. I still don’t know what to make of that conversation. Maybe the guy had slept through the part of his psychiatric training where they covered sleep disorders. Fortunately, I think they promoted him to administration. That way, he wouldn’t be working with actual clients anymore.

So if you’re having a hard time with sleep, and you go to see a mental health professional, and they tell you some weird sh*t that doesn’t make any sense at all, you just need to realize that, as the experts tell us, going short on sleep can have a variety of adverse consequences, possibly including meeting with mental health professionals who don’t know their ass from their elbow.

Some adverse consequences of short sleep are obvious immediately: low energy; reduced mental sharpness; the risk of falling asleep in a place where doing so may be embarrassing or possibly fatal (e.g., traffic). Other consequences may become manifest in the short to medium term, but it may not occur to you that they are linked to sleep. Examples cited by WebMD (Peri, 2014) and Healthline include depression, loss of libido, skin aging, forgetfulness, weight gain, poor balance, weakened immunity, and impaired judgment. Beyond that, Medic et al. (2017) identified a number of potential longer-term consequences of sleep disruption, including hypertension, cardiovascular disease, weight issues, cancer, and premature death. Johns Hopkins offered a graphic depicting a variety of consequences:

I was particularly interested in that graphic’s indication that sleep deprivation entails a 33% increase in dementia risk. According to Mander et al. (2017), starting in one’s 40s, sleep tends to change in a number of mostly undesirable ways, including advanced timing (i.e., earlier bedtime and rising time), being slower to fall asleep, getting less sleep, more fragmented sleep, more fragile sleep (i.e., more easily disturbed), reduced NREM SWS sleep, fewer NREM-REM sleep cycles, and increased time spent awake during the night. Mander et al. said that REM sleep impairments could be a problem too, though often they would not appear until one’s 80s, or would appear only as a symptom of dementia.

Mander et al. (2017) addressed the question of whether older adults get less sleep than younger people because they need less sleep, or rather because they need more sleep than they are getting. There were arguments on both sides. Facts favoring the former view included (a) older adults sleep much less than younger adults even when put into situations virtually compelling them to stay in bed, (b) after sleep deprivation, older adults display less of a catch-up need, and (c) by both objective and subjective measures, older adults deprived of sleep exhibit less relative impairment on sleep-sensitive (e.g., vigilance) tasks. But facts favoring the latter view included (d) molecular evidence suggests that older adults do have the sleep need one would expect, in situations like those just described, but they are less sensitive to that need, (e) sleep-deprived younger adults revert to lower levels of sleepiness in later days of sleep deprivation, suggesting that older adults are similarly inclined to stop noticing their sleepiness, (f) rested older adults show levels of cognitive impairment similar to sleep-deprived younger adults, so that sleep deprivation in research situations simply has less of an impact on older adults, and (g) older adults perform much worse than younger adults on such sleep-dependent cognitive functions as learning and long-term memory consolidation. These arguments suggested that an older person might be better able to deal with a situation requiring an ability to cope with sleep deprivation — but that, as Mander et al. concluded, older adults have impaired abilities to register that they need sleep and/or to get that sleep.

Mander et al (2016, p. 2) observed that “older adults with superior sleep quality have a significantly lower risk” of developing Alzheimer’s, and also that they maintain cognitive function longer. My interest in the link between sleep deprivation and dementia — indeed, my motivation for writing this post — stemmed from a Frontiers in Pharmacology article (Cordone et al., 2019) containing this statement:

The growing interest in the preclinical stage of Alzheimer’s disease (AD) led investigators to identify modifiable risk and predictive factors …. There is a consensus on the importance of sleep within this context: the bidirectional relationship between sleep and AD pathology is supported by growing evidence that proved that the occurrence of sleep changes, starting from the preclinical stage of AD, many years before the onset of cognitive decline. … Growing evidence supports the notion that insomnia, excessive daytime sleepiness (EDS), sleep-disordered breathing (SDB), and circadian sleep–wake alterations all seem to increase the risk of AD.

Well, that was good news. I had only two of those four precursors — only insomnia and a circadian issue. So I wasn’t a complete poster child for Summer of 2030, Alzheimer’s Edition. Still — just to be a responsible adult about it, on behalf of those who would be pushing my wheelchair — I did think I might want to try to develop a clear understanding of this, to the extent possible, considering how sleepy I was. For the only thing better than a life well lived is being able to remember some of it.

When Cordone et al. (2019) referred to a “bidirectional relationship,” they meant that Alzheimer’s impairs sleep and, in turn, impaired sleep seems to enhance the risk of Alzheimer’s. Since I didn’t (yet) seem to have Alzheimer’s, I was only interested in the second of those two propositions. That is, I wondered whether my apparently impaired sleep would enhance my risk of Alzheimer’s.

The Cordone article was especially preoccupied with the accumulation of the A-beta (i.e., Aβ, AB, or β-amyloid) protein fragment. I didn’t know anything about that; I just needed to surf it and see if I could get a sense of what they were saying.

It seems that, when Aβ lingers in the brain, eventually it becomes aggregated to form the plaques and tangles of Alzheimer’s. Vanderheyden et al. (2018) explained that you would want to find a healthy amount of Aβ in the brain, because this would mean that Aβ was not being immediately absorbed into such Alzheimer-related formations, but was instead waiting for sleeptime to clear it. This was evidently a very long-term affair: it appeared that reduced Aβ levels could precede the onset of Alzheimer’s by 15 years or more. So if you knew of people who started to show signs of Alzheimer’s at 65, you might want to start thinking about this by the time you turn 50.

It seems that Aβ clearance proceeds much more rapidly during SWS — which, as noted above, occurs during NREM Stage 3 sleep. Unfortunately, according to Medscape (Stevens, 2015), older people spend less time in NREM 3. Mander et al. (2013) reported evidence that, even without Alzheimer’s, ordinary age-related atrophy in the brain’s prefrontal cortex diminishes SWS sleep, which in turn means impaired long-term memory. Ohayon (2004, pp. 1263, 1270) found that minutes of SWS sleep per night dropped sharply, through adolescence, and more slowly thereafter — from a high of nearly 200, at age 5, to a low of approximately 15, at age 85, with SWS accounting for about 25% of total sleep at age 5 but less than 10% after age 65.

Some of those declines seem to be due to decreased sleep continuity, which appears to mean increased sleep fragmentation (e.g., Stamatakis & Punjabi, 2010). Stevens (2015) noted that fragmentation “may be exacerbated [in older people] by the increasing number of geriatric medical conditions, including sleep apnea, musculoskeletal disorders, and cardiopulmonary disease.” In other words, the more highly interrupted sleep of old people may not give them enough time in NREM Stage 3 to achieve adequate Aβ clearance. Sleep apnea may be the worst-case offender, insofar as its interruptions can occur many times each night (e.g., Mayo Clinic).


So I had some things to think about. For one thing, they were advising me to move toward longer nights of sleep, with less reliance on naps. I might also want to do something about the circadian cycle. In addition, there were some sleep hygiene considerations. And I had this particular concern with how much SWS sleep I was getting.

In the area of sleep hygiene, alcohol and caffeine were not much of a factor in my life these days. As far as I could tell from my audio recordings of myself (viewed in an audio editor like Audacity, where the waveform would quickly point me toward the noisy spots), my principal sleep interruptions were when I got up to pee. Those interruptions could hopefully be reduced by being more careful not to drink liquids for several hours before bedtime, as part of good sleep hygiene (above). It could also be worth considering lifestyle changes (e.g., changes in household rooms or in residence address) to reduce noises, distractions, and anything else that might interfere with good sleep. On a comparably dramatic level, some men might decide to adopt an aggressive stance toward prostate problems, bearing whatever consequences that might bring.

It seemed that one way of determining whether I was getting enough SWS sleep was just to decide whether I felt good when I got out of bed. Cordone et al. (2019) cited research finding that “human subjective reports of reduced sleep duration and diminished sleep quality correlate with cortical Aβ burden in healthy older adults.” In other words, we come pre-equipped with an ability to evaluate the quality of our sleep. This ability is apparently not perfect: I saw, somewhere (perhaps in Ohayon, 2004), an indication that women’s subjective evaluations of their sleep quality tends to correspond more closely than men’s to objective measures of sleep quality (as determined by e.g., accuracy in responding to various word tests). For men and women alike, however, it seemed that our subjective sleep evaluation ability might provide at least a rough indication of how effectively our brains are clearing Aβ. So a person who was waking up tired, as I was, after a five-hour night’s sleep, could evidently take that as a reasonably good hint that Aβ was not being adequately cleared, thereby putting him/her at elevated risk of Alzheimer’s.

In case one’s own subjective impression of sleep quality didn’t seem reliable enough — and also for those whose conditions might not permit an immediate if not magical lurch into the eight-hour night’s sleep — it appeared that tech might be in the process of offering electronic means of tracking one’s sleep stages. For example, Lustgarten (2019) exclaimed that a PSG informed him that, at age 26, he had “the deep sleep duration of a 65 year old!”

Lustgarten also said he was using a wrist-worn device capable of measuring sleep stages. Unexpectedly, in his case, increased hours of sleep did not yield increased SWS. The device he used was the WHOOP Strap ($18/mo. for an 18-month membership). That, however, was not among the top sleep trackers listed in recent reviews by Esquire, Wareable, or Digital Trends. Rather, those and other sources distinguished smartphone apps, wearable devices, and freestanding monitors:

  • In the sleep-tracking app category, Wirecutter considered SleepScore (free or $50/year) best, because it offered advice on how to achieve sleep goals, provided detailed sleep stage data, and included a fairly effective smart alarm, timed to awaken the user at a time when s/he would not feel groggy. Wirecutter said that its runner-up, Sleep Cycle (free), did not offer sleep stage data, but ExpertReviews disagreed, naming it the best sleep-tracking app, as did Joan Lunden.
  • Moving up from apps to wearable devices, ExpertReviews and Wareable chose Fitbit’s Inspire HR ($100, 3.5 stars from 652 reviewers on Amazon) as the best available sleep-tracker wristband, due to its accuracy and its sleep stage analysis capabilities. Fitbit sponsored a study (see SleepReviewMag, 2017) finding Fitbit’s wrist devices accurate at identifying sleep stages (although the New York Times, below, indicated that the study found only 70% accuracy).
  • Finally, moving up again to the freestanding monitors, several sites (i.e., ExpertReviews, Esquire, and Digital Trends agreed on the Withings Sleep Tracking Mat ($100 from Amazon, with 3.8 stars from 89 reviewers), apparently for its accuracy in reporting sleep duration, quality, snoring patterns, and interruptions; apparently it did not offer sleep improvement advice or a smart alarm. Esquire recommended Tomorrow Sleeptracker as an alternative offering “comprehensive features at a good price” — though apparently they weren’t interested in actually selling it, as their website was unresponsive and it wasn’t listed on Amazon. Most of these sites listed additional alternatives on all three levels — apps, wearables, and freestanding monitors.

There was a possibility, however, that in this case less might be more. The New York Times (Zraick & Mervosh, 2019) warned that not every sleep tracking device was accurate. That was the finding of a study by Gruwez et al. (2019), although the items they tested were not among those recommended above. Other concerns mentioned in the Times article included bad sleep hygiene (i.e., messing around with devices in bed) and orthosomnia (i.e., unhealthy worries or obsession with achieving perfect sleep). It seems that people were worried about living up to the machines’ expectations, even when the devices in question were inaccurate.

The TImes article noted the low-tech alternative of a sleep diary. Harvard Health (2018) said that a sleep diary could track sleeping and waking time (including time awake in bed) for a week or two, as part of a sleep restriction strategy. Stanford Health Care explained that sleep restriction could be an appropriate cognitive-behavioral intervention for someone who spends a lot of time in bed but isn’t actually sleeping all that time. Falloon et al. (2015) said the purpose of sleep restriction was to consolidate fragmented sleep by reducing the sleep opportunity, thereby inducing mild sleep deprivation to enhance the person’s sleep drive. Tuck (2018) said the goal would be to get the person to fall asleep quickly and sleep through the night. A principal objective was to get the person to associate the bed with sleep, rather than with long stretches of tossing and turning.

I wasn’t sure that those thoughts were very helpful for me. My situation was more that of the person who awakens after five hours of sleep and wants to get up and get busy. For that, Harvard Health (2018) suggested relaxation techniques. I had some prior exposure to meditation and some other relaxation techniques; not so much to biofeedback. It was interesting to see that, for example, someone was now working on a study exploring the use of meditation to reduce sleep fragmentation, compared against a control group using a sleep hygiene and exercise program. I was interested in learning more — although, to do so, I was going to have to make some decisions about my priorities: be busy and feel productive, or else learn to prioritize this other stuff instead, and maybe live longer and better as a result.

In a previous post, I discussed various pills and techniques that people used to fall asleep. But in this case the need was different. I was falling asleep just fine. I just wasn’t staying asleep. There was the part about training myself to slow down and stay in bed, and I could work on that; but apparently there was also the part (above) where old people generally needed more NREM sleep than they were getting. In response to that, Cordone et al. (2019) cited research suggesting that “NREM sleep enhancement in midlife to late life may lead to a preventive positive effect that could decrease AD risk.” That research, by Mander et al (2016), found particular potential for enhanced NREM SWA in “transcranial direct current stimulation” (p. 9) — that is, electricity applied to the head. But they were also interested in auditory stimulation during NREM SWA, and by kinesthetic stimulation (i.e., slow bed rocking). More recently, Zhang and Gruber (2019) concurred that several different kinds of interventions (i.e., auditory, electrical and pharmacological) had potential to increase SWS duration.

It wasn’t clear when the researchers would get around to designing the ideal NREM enhancement mattress. In the meantime, that talk about kinesthetic stimulation reminded me of the lulling sensation of riding long distances on a train. I wondered whether a traveling lifestyle might be helpful. For one thing, I had noticed that, when traveling (for pleasure, not business), at least at first, I slept better in hotels than I did at home; and there was also the possibility that the right kind of travel could entail exposure to camping (e.g., Stothard et al., 2017; Wright et al., 2013), rainfall, and other soothing sleep environments. Few sounds in my life had been more pleasant than the sound of snowflakes sliding down the side of my tent in a deserted wintertime campground.

Mander et al (2016, p. 10) said that cognitive behavioral therapy (CBT) was another possible solution. In other words, if mental health care is covered by your insurance, and if you have access to someone trained in CBT, it can’t hurt to explore that option. Harvard Health (2018) explained that CBT seeks to substitute constructive thoughts for unproductive ones. In my case, CBT might work toward an understanding that staying in bed a while longer might make me more productive, over the long term, than jumping up and getting started right away on a new day. Unfortunately, according to Mander et al., medication is not effective, so far: they say the known drugs don’t necessarily help and may even cause further memory loss. There were surely other possibilities. For instance, Lustgarten (2019) contended that he was able to achieve much longer SWS sleep by increasing the level of a certain kind of bacteria in his gut.

Interpretation and Conclusion

The ideal concept seemed to be that a person should get a full night’s sleep, consisting of roughly seven to eight hours of uninterrupted sleep each night. There could be situations where a midafternoon power nap (of 15-20 minutes) could add sharpness without detracting from end-of-day sleepiness. To assist in achieving that full night’s sleep, a person should observe a number of rules of sleep hygiene, such as avoiding caffeine and alcohol for some hours before bedtime, not using tech devices (e.g., computers, smartphones) near bedtime, and avoid food and excitement (e.g., movies) before bedtime.

The more likely situation was that the f*cking boss would have given me a look I didn’t like as I was heading out the door at 5 o’clock; I would end the day by working late over a nice, spicy dinner of Chinese takeout after an argument with the girlfriend; I would calm down with a glass or two of cheap red wine; and the neighbor’s dog would be barking as I slid into a bedroom that was a little too warm because I was economizing on the air conditioning. Or if that wasn’t exactly my situation at present, I certainly had experienced enough of such things, at various stages of life. The more recent version was that, as a retired guy, I found that I had built myself a sleep pattern consisting of roughly 5.5 hours of sleep, supplemented by at least one or two naps, some of which were long, all rotated to a point about eight hours past where they should be.

My question was whether my body had perhaps worked out an adaptive, possibly even a superior, solution. The hunter-gatherers and the ancient Europeans had apparently arrived at variations on the same theme, which was that you go to bed sometime after dark, you get up for a while in the middle of the night, and then you go back to sleep until dawn — and that was a bit like what I had done at one job, where my way of getting something like a full night’s sleep (or at least enough to get through the workday with) was to go to bed at 7 or 8 PM, sleep until sometime around midnight, get up and go for a run or a bike ride (because the roads were deserted and it was cool then), come back and shower and work some more, and then go back to sleep at maybe 4 AM.

I wasn’t doing that now, because I had more freedom in retirement. I was waking up after five and a half hours, give or take, perhaps because that would be about enough time for three NREM-REM cycles, each running 90 to 120 minutes. Those cycles would be heavy on the SWS sleep that reportedly predominates in the first third of the eight-hour night’s sleep — and this was good, because SWS was what my head needed to clear Aβ, so as to avoid Alzheimer’s. I was tired — I hadn’t had enough sleep — but I was waking up from light sleep, so I wasn’t really groggy.

I awoke at that point partly because I had things on my mind, and I could perhaps work on learning to let those things wait a bit longer, in the belief that I would be more efficient with more sleep. To that end, I had found that wakeup time was perhaps the best time of day for meditation: I was probably too awake to fall back asleep, which could be a problem with attempts to meditate at some other times of day — but if the meditation did put me back to sleep, that was fine too.

Aside from my thoughts, I was inclined to awaken after five (or so) hours of sleep because I wasn’t tired enough, anymore, to sleep through things like my sore knee, which actually bothered me more in bed than when I was running, perhaps because lying on my side (which I favored) would allow the leg’s weight to press laterally on the knee. I had lately been working on extending my primary sleep time, with meditation and such, and that had potential, but I honestly wasn’t sure I had enough fatigue to carry me through another sleep cycle, and I was a bit worried that forcing myself to stay in bed, when I couldn’t sleep, would actually be bad sleep hygiene. At present, I was thinking it was probably OK as long as I was tired or mellow enough to meditate or to fall into a light sleep. And, who knows? Maybe I could eventually add a fourth sleep cycle to my usual night.

The good thing about those three sleep cycles, in my five hours of sleep, was that they did tend to be uninterrupted. I could try to verify this with a sleep tracking device. I might be surprised by the results. But assuming I was right, it seemed that my long naps had the same advantage: I seemed to be tired enough to sleep right through one (sometimes two) 90-minute cycles in a single nap — thereby insuring, again, that I was getting SWS sleep. The sleep purist’s argument was that, if I just stopped taking naps, I would naturally lapse into eight hours of sleep, and that might be true, but that wasn’t my experience. My experience was that sleep was not that predictable.

In short, it was possible that retirement would somehow enable me to achieve consistent eight-hour nights of sleep with fewer and shorter naps. This did not seem likely, given my life history and the tendency of older people to sleep less, but it was something that I could seek via incremental changes in some aspects of my current situation. In particular, I could work toward a more normal sleep schedule with, perhaps, an additional nighttime sleep cycle that would move me closer to a total of seven hours of sleep per night. I would want to continue to monitor the science press for new insights and, in the meantime, I would value naps to the extent that they contributed to my daytime performance and gave me opportunities for uninterrupted daytime sleep that I might not have at night.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.