The second force is Process S, sleep pressure: the longer someone stays awake, the more pressure builds to fall asleep. This is sometimes described as the homeostatic sleep drive, because when there is more need for sleep, Process S is supposed to return a human or animal to a state of balance. When someone does get to sleep after extra time awake, an abundance of slow wave or delta sleep, the deepest form of non-REM sleep, is a sign of relief of that pressure. We know that people with IH experience sleep pressure. Indeed, it makes them miserable when they are awake. They bear the imprint of one half of Process S but not the other: its release.

  MEASURING MELATONIN

  One reason to suspect an alteration of circadian rhythms in IH is a change in the nightly pattern of melatonin, the hormone that signals nighttime to the body and helps set our day-night schedule. Despite its common use as a sleep aid, melatonin should be considered a “hormone of darkness” rather than a sleep-induction molecule like adenosine. Nocturnal animals such as mice have higher melatonin levels at night, even though they are awake and active at that time. So changes in melatonin reflect a distortion of the underlying circadian rhythm.

  In two studies of people with IH, the production of melatonin appears to be weaker and more stretched out, compared with healthy sleepers. Investigators in Prague, led by Bedřich Roth’s associates, were the first to show this among people with IH in a publication from 2000.29 The results resemble melatonin profiles in habitual long sleepers.30 Normally, melatonin levels in saliva begin to increase at around 9 or 10 pm before bed, peak at around 2 am, and fall off as morning approaches. According to the findings from Prague, in people with IH, melatonin levels began to increase later, peaked at a level that was less than half normal, and then remained at an elevated level into the early afternoon. The rhythms of cortisol, a stress- and alertness-related hormone, were also shifted and delayed in IH patients. The Czech researchers focused on people with IH who reported long sleep times and sleep drunkenness; their results for people with narcolepsy with cataplexy and nonidiopathic hypersomnia were described only as “heterogeneous.”

  In Boston, Thomas has made similar observations. In his brief report published in 2020, most IHers (forty out of fifty) displayed an extended period when melatonin was elevated, compared to their awake state.31 Thomas has proposed that a long biological night has implications for how IH could be treated: with bright light therapy or the carefully timed application of melatonin-suppressing beta blockers. While the success of his clinical approach is currently anecdotal, it makes use of existing tools that are relatively accessible.

  THE SUPRACHIASMATIC NUCLEUS

  The long biological night theory could potentially fit with the proposal that an injury to part of the hypothalamus causes IH. The hypothalamus contains a structure called the suprachiasmatic nucleus (SCN), which sets circadian rhythms and transmits information about environmental light to the rest of the body. Its firing rate peaks in the middle of the day and reaches a minimum in the middle of the night. When stimulated by light, through a series of connections involving the retina and the spinal cord, the SCN inhibits production of melatonin. Beyond inhibiting melatonin, the SCN acts as a master regulator of circadian rhythms, sending signals to the rest of the hypothalamus and other parts of the brain. The SCN does not connect with most of the brain’s sleep-wake circuits directly, and its influence is exerted through other parts of the hypothalamus.

  The SCN’s importance first emerged in the 1970s, when researchers showed that destroying the SCN in rats eliminated the rats’ circadian rhythms: they drank water and ran around without regard to the time of day.32 Some studies of the SCN in animals showed that lesions in the SCN dramatically increase the total amount for sleep. In squirrel monkeys, animals that display consolidated sleep and wake periods similar to humans, SCN lesions resulted in them sleeping four hours more every day, and their average time awake without naps lasted only fifteen minutes.33 However, the same type of effect has not been observed consistently in rodents; this discrepancy may come from neuroanatomical differences, imprecise experimental surgeries, or the environments the animals were kept in.

  It is possible that defects in SCN function weaken circadian rhythms in people with idiopathic hypersomnia. Degeneration of the SCN is also thought to lie behind disrupted circadian rhythms in Parkinson’s disease and other neurodegenerative diseases.34 Sleep researchers at Northwestern have observed a weakened, extended nighttime profile of melatonin in people with Parkinson’s, particularly in those who had stronger symptoms of excessive daytime sleepiness.35

  CIRCADIAN RHYTHM GENES

  The 2017 Nobel Prize in Medicine or Physiology was awarded for the discovery of circadian clock genes in fruit flies, and similar genes exist in humans too. Mutations in genes encoding parts of the circadian rhythm’s cellular clock have been found in “morning lark” humans, who have an advanced sleep phase. Other mutations appear in their “night owl” opposites, who have a delayed sleep phase.36 The mutations compress or extend the clock’s timing, producing a circadian clock that would fit better on a planet with a twenty-three- or twenty-five-hour day, respectively. These mutations’ effects go beyond what we see in older people, who tend to go to sleep early and wake up early, or in many teenagers, who display the reverse. The mutations shift the time when people carrying them go to bed and wake up compared to typical sleepers, but their total amount of sleep need does not change.

  Additional support for altered circadian rhythms in IH has come from studies conducted at the University of Münster in Germany. Researchers asked people with IH to donate skin cells, which were cultured and artificially synchronized to follow the oscillations of their circadian rhythm genes. In IHers’ cells, the activity of core clock genes swung up and down markedly less, compared to controls.37 In addition, the overall activity of the core clock gene BMAL1 was reduced. A follow-up study showed that circadian period length was slightly longer in IHers’ skin cells.38

  These were intriguing results because it looked like the alterations in circadian rhythms were independent of the nervous system and present in every cell in the body. However, signals carried in the blood are continually corralling the clocks of peripheral tissues, so weakened circadian rhythms in IH may not be “cell-intrinsic.”39 In addition, the Münster authors left out information on study participants’ circadian rhythms and sleep habits. Did some exhibit abnormally long sleep periods at home? What about their melatonin profiles? All we know is that they were diagnosed with IH according to current criteria and thus fell asleep quickly during an MSLT. Altered circadian rhythms have been detected in several other disorders, ranging from diabetes type 2 to schizophrenia.40 Side-by-side comparisons are necessary to assess whether circadian alterations are more pronounced in IH compared with other conditions.

  The reports from Prague, Boston, and Münster highlight the need to more thoroughly examine circadian rhythms in IH patients, including measurements of melatonin, cortisol, and body temperature, along with the activity of their circadian rhythm genes. Measurements of melatonin are inherently tricky because of dim light requirements, but a major focus of sleep researchers’ efforts over the last few years has been to develop robust blood biomarkers for circadian phase.41 Future studies of IHers could incorporate measurements of such biomarkers, as well as examining the effects of medications on their circadian rhythms.

  TOWARD GENETIC STUDIES OF IH

  Along with probing circadian rhythms, future studies of IH should include genetic sequencing of “multiplex” families. For IH, musing about a genetic link extends back to the 1950s. Four members of a family with similar symptoms occupied a central place in Bedřich Roth’s initial report. Later sleep researchers encountered families having several members with IH, suggesting strong inheritance effects.42 For example, Michel Billiard and Yves Dauvilliers reported a patient whose sister, mother, and maternal uncle all had IH.43 Most clinicians who have studied IH in detail have found that a large fraction of IH patients report having relatives with a history of excessive daytime sleepiness, if not outright IH. Examples in this book include Anna Sumner Pieschel and her brother James.

  Then why has progress been so meager? Partly, the lack of scientific agreement about how to define IH clinically and the absence of research funding. Those obstacles are beginning to change, at a time when genomics could make progress more rapid. Today, modern genetics should allow the historic strands connecting narcolepsy and idiopathic hypersomnia to be unraveled.44

  So far, no mutations closely analogous to those in Yanagisawa’s sleepy mice have been found in humans, but technology has advanced to a point that might astonish Bedřich Roth. It is now possible to sequence an entire human genome within a week, for less than $1,000.45 Whole exome sequencing—a limited survey of the protein-coding regions of the genome—is even more accessible. If a few members of a family are affected by the same disease and donate DNA samples, researchers can compare patients’ genomes to unaffected relatives. By sorting through a list of genetic changes present in the patients’ genomes, it is possible to identify mutations that may be responsible for their disease. This approach is now used to diagnose cases of autism spectrum disorder and early-onset epilepsy, although the diagnostic yield (the fraction of cases in which a cause can be assigned) is generally below 40 percent.46 Similar studies on multiplex IH families could be critical for establishing IH’s biological basis.

  JAPANESE GENETIC STUDIES

  Before whole exome or genome sequencing became widespread, another approach was in common use at the beginning of the twenty-first century. A GWAS (genome-wide association study) makes use of single nucleotide polymorphisms (SNPs), places where one letter of the genetic code varies in a fraction of the population. In a GWAS, researchers examine hundreds or thousands of human genomes in people with a particular disease—reading SNPs but not the entire genome. The SNPs don’t necessarily cause the disease in question, but they offer hints of where to dig deeper. GWAS doesn’t require a strong inheritance pattern; it can pick up weaker influences.

  The only published attempt to find genetic risk factors for IH using GWAS techniques provided limited results. In this study, Japanese researchers used GWAS techniques to study 408 patients with “essential hypersomnia,” a category that corresponds to narcolepsy without cataplexy and IH without long sleep combined. What emerged was just one marker, close to a gene encoding an enzyme called carnitine O-acetyltransferase (CRAT).47 Carnitine helps cells metabolize fatty acids as fuel, but how CRAT fit into the IH puzzle was not clear, and the CRAT mutations’ effect was not strong.48

  As this book was being revised for publication, the same Japanese team published the first study of IH to identify a mutation with a strong link to the disorder.49 The mutation may account for only a sliver of IH cases in Japan, since only twenty out of 598 IH-diagnosed patients had the identified mutations. Still, the study was significant because it was the first to find anything potentially causative. The mutation was in the prepro-orexin gene, encoding the neuropeptide orexin/hypocretin (discussed in the next chapter). The mutation did not appear to induce sleep-onset REM, even though the overall effect of an orexin/hypocretin mutation should be like a limited version of narcolepsy type 1.

  LONG SLEEP IH MAY BE DISTINCT

  Future genetic studies may need to separate out patients with the long sleep form of IH. Supporting this point, a 2021 paper from Nevšímalová and her colleague Karel Šonka argued that IH with long sleep should be considered a distinct clinical entity, based on stability of the symptoms and clinical course of the disorder. In their paper, the Czech authors reexamined IH patients they had seen over the last twenty years.50 Those in the long sleep duration group could sleep for an average of more than twelve hours if given the chance to do so. They reported a higher frequency of sleep inertia and autonomic nervous system symptoms, along with a poorer response to conventional medications. Other investigators, such as Isabelle Arnulf, have also found that people with long sleep IH tend to display severe sleep inertia or sleep drunkenness more than IH without long sleep.

  Longitudinal and genetic studies could resolve the issue. Individuals and especially families with long sleep IH have not been studied in a way that would allow researchers to say that someone who habitually sleeps seventy or more hours per week needed the same amount one or five years ago. Most previous studies of IHers have relied on patients’ recall of their habitual sleep duration. Also, IHers’ attempts to manage their symptoms with stimulants may obscure shifts in their underlying sleep need.

  WIDE VERSUS DEEP

  Studies on families and twins indicate that at the population level, a large part of habitual sleep duration is genetically controlled. From an evolutionary point of view, humans sleep far less per twenty-four hours than expected based on comparisons with other primates, so the genetic changes that facilitated that difference may be relatively recent.51 Based on recent studies, it looks like dozens of genes can influence sleep duration, each one having a small effect. In the general population, excessive daytime sleepiness is more likely to come from insomnia or sleep apnea, not from an inborn increased need for sleep.52

  One of the largest studies to probe genetic influences on sleep duration comes from the Broad Institute, where investigators were able to tap into the UK Biobank project, which has collected DNA from close to half a million people.53 In this analysis, the 5 percent of people who carried the most sleep-increasing genetic variants reported sleeping about twenty minutes more than those carrying the least. The largest effect that any single variant had was to increase sleep duration by 2.4 minutes. Genetics explained a small part of the total variation in sleep duration; other factors—age, job, or previous health history—influenced sleep duration more.

  Genome-wide association studies like this rely on tracking common variants, which usually don’t make a large difference in gene function. GWAS can spot interesting gene variants that influence a universal trait—such as sleep duration—or a common disease.54 GWAS is less applicable toward getting to the bottom of a rare sleep disorder. Mutations driving IH are unlikely to be present at a sufficient frequency in the general population. For rare disorders such as IH, going deep—looking for rare variants with a large effect on sleep duration—may be better than going wide. The long sleep form of IH would be more unambiguous and better suited for this type of study.

  SHORT SLEEP FAMILIES

  The type of study I am proposing has already been performed with families who exhibit the opposite phenotype from IH. The geneticists Louis Ptacek and Ying-Hui Fu, a husband-and-wife team at University of California–San Francisco, have had extraordinary success finding natural short sleep families and isolating the mutations responsible. If this is possible for short sleep, it should be possible for long sleep IH as well.

  Ptacek and Fu came to this topic through their work on advanced sleep phase syndrome: people who both wake up and go to bed several hours earlier than others, without sleeping more overall. Their investigation of a Utah family, sparked by a colleague’s encounter with a woman with pronounced advanced sleep phase, began before the Human Genome Project and other advances made such studies less burdensome.55 Until they were close to their quarry, their effort was independent of knowledge of circadian rhythm genes in fruit flies. Discovery of the responsible mutation was a confirmation that clock genes isolated through fly genetics were relevant for human biology.56 Since then, DNA sequencing has become easier and less expensive, enabling the investigation of smaller families with just a few affected individuals.

  Fu and Ptacek’s initial reports on the Utah advanced sleep phase family had led to a stream of people contacting them. In their initial investigations of other advanced sleep phase families, the mutations were easier to find because of clues about where to look.57 Instead of having to search across the entire genome, the scientists took an educated guess about what might be affected: the other components of the circadian clock. Their guesses were correct, at least for some families. However, the alterations in sleep did not manifest in the same way in each family—in some families, affected individuals displayed signs of seasonal depression or migraine headache.58

  Other families who contacted Fu and Ptacek were different, in that affected members woke up early but didn’t go to bed early to compensate. These individuals needed less sleep than other people in their families—usually less than six hours every night, and they didn’t seem to suffer negative health consequences.59 In Fu and Ptacek’s first paper on a short sleep family, the “candidate gene” (or educated guess) approach led them to a mutation in a component of the circadian clock called DEC2.60 But the other short sleep families they identified did not carry mutations in circadian clock components. Instead, affected family members have other mutations affecting neuronal signaling. Fu and Ptacek have identified more than fifty short sleep families and have published genetics on only a few so far, and they have not described the characteristics of those families’ sleep in detail.61

  From research elsewhere, there is evidence that habitual short sleepers display a higher proportion of slow-wave sleep; they may be packing the same level of restorative sleep into fewer hours spent in bed.62 When they’re awake, genetic short sleepers appear to be more efficient and optimistic than other people. Fu told Scientific American: “They like to keep busy. They don’t sit around wasting time.”63

  It is striking how some genetic mutations in humans shift sleep cycles forward or back, while others decrease the need for sleep, but none have been found that markedly increase sleep need—yet. Fu and Ptacek said their team had only begun to collect long sleepers and had not begun to screen for mutations. I asked Ptacek whether it would be possible for a mutation in a circadian rhythm clock gene to lengthen the nighttime phase, producing long sleep IH. This hypothetical mutation would be analogous to the DEC2 mutation but would have the opposite effect. “I believe the answer is yes, but we have not found a mutation that does it,” Ptacek said.