When, not how long

Every number in this course so far has been an amount. Four, six or eight hours. Seven or more. A relative risk attached to short sleep, wherever each study drew the line.

None of them says anything about when.

That's a second quantity, it's measured by different instruments, and in this course's experience very little of what people are told about sleep is about it.⁴ This lesson is that quantity, and it's where this course's own question comes from: is this claim about how much, or about when?

Chronotype, and what it is a measurement of

People differ in when their bodies are set to sleep, not only in how much. The technical name for that difference is chronotype, and the field defines it as a property of how you synchronise with the world rather than as a preference.

In the words of a 2020 paper on how to measure it:

"Individuals vary in how their circadian system synchronizes with the cyclic environment (zeitgeber). Assessing these differences in phase of entrainment—often referred to as chronotype—is an important procedure in laboratory experiments and epidemiological studies but is also increasingly applied in circadian medicine, both in diagnosis and therapy."¹

This course read that paper's abstract and the section defining the measure, and nothing else of it.

A zeitgeber is a time-giver: something in the world that your internal clock lines itself up with, light being the main one. Phase of entrainment is where your clock has settled relative to it.

Two more words this lesson leans on, glossed here because no earlier course defines them. Circadian means roughly a day long, and describes the internal cycle this lesson is about. Your biological night is the stretch of that cycle your clock treats as night. It usually overlaps your sleep, but it doesn't have to line up with the clock on the wall. Both glosses are this course's plain-language background rather than definitions quoted from a source.⁴

The gold standard, and the thing everybody uses instead

The same abstract says how it's properly measured: "While biochemical measurements (e.g., dim-light melatonin onset [DLMO]) of internal time are still the gold standard, they are laborious, expensive, and mostly rely on special conditions (e.g., dim light)."¹

So the real instrument is a hormone assay taken in near darkness, which is the fifth instrument in this course after lesson 1's four. And the reason you have never had it done is in the same sentence: laborious, expensive, special conditions.

What is used instead is a questionnaire, and the abstract is careful about what that buys: "Chronotype estimation in the form of questionnaires is useful in approximating the timing of an individual's circadian clock."¹ Approximating.

The marker the questionnaire produces is the midpoint of your sleep on free days, the halfway point between falling asleep and waking on a night when nothing wakes you. The researchers then correct it, in their words, "for potential oversleep on free days (to compensate for sleep debt accumulated over the workweek, MSFsc), therefore accounting for the homeostatic process influencing sleep."¹ This course didn't read the formula for that correction and so doesn't print one.

What light does

The clock moves, and light moves it. Here's what that looks like when somebody measures it.

A 2012 experiment gave participants one hour of bright white light, about 8000 lux (lux is a unit of how much light reaches a surface), at one of eighteen different points around the circadian cycle, and measured how far their clocks shifted.² Phase was measured as the change in dim-light melatonin onset, under constant routine in under three lux. A matched group got the same hour in dim light, under three lux, and "No discernible PRC was observed" for them,² which is what shows it was the bright light doing the work. This course read the abstract in full and one sentence of the introduction, and nothing else.

The size of the effect: "Exposure to 1 h of bright white light induced a Type 1 PRC with a fitted peak-to-trough amplitude of 2.20 h."² A PRC is a phase response curve: the size and direction of the shift, plotted against when in the cycle the light landed. "Type 1" is a technical class this lesson doesn't need. One hour of light, and on the fitted curve the difference between the biggest delay and the biggest advance was over two hours of clock time.

And the direction, as the paper's introduction states it: "Light exposure occurring in the early biological night induces a phase delay shift of the circadian pacemaker, whereas light exposure occurring in the late biological night induces a phase advance shift."²

The same light does opposite things depending on when it lands. Early in your biological night it pushes your clock later. Late in your biological night it pulls it earlier. And since the curve runs from one to the other, somewhere between them it crosses zero, which is this course's reading of the shape rather than a figure the abstract gives.⁴

The trap in that sentence

Try this one, and notice that the obvious answer is wrong.

Noticing which advice about light quietly assumes otherwise is most of what this lesson is for.

The second hour is worth less than the first

One more line from that abstract: "The fitted peak-to-trough amplitude of the 1 h bright light PRC was ∼40% of that for the 6.7 h PRC despite representing only 15% of the light exposure duration, consistent with previous studies showing a non-linear duration–response function for the effects of light on circadian resetting."²

Fifteen per cent of the exposure bought forty per cent of the range. Working backwards, the longer exposure's amplitude comes out around five and a half hours, which is this course's arithmetic on the abstract's own two figures and not a number the paper prints.⁴

What that shape means: the first hour does far more than its share. Fifteen per cent of the time produced forty per cent of the effect. The abstract gives two points, one hour and 6.7, and nothing in between, so this course can't tell you what the second or the fourth hour is worth, only that together the hours after the first did less, hour for hour, than the first did.

Screens, once

This course has read no measurement of what a screen does to sleep or to circadian timing. Not a weak one, not a contested one: none.

So it has nothing to say about screens before bed, in either direction, and it will not mention them again. If you meet a confident claim about them, lesson 8 is the method, and one of the first questions is which of the five instruments produced the number.

When timing meets an outcome

There is one place where sleep timing has been set against a hard outcome at scale, and it is the only part of this lesson that touches somebody's livelihood.

In 2019 the International Agency for Research on Cancer, the World Health Organization's cancer agency, classified night shift work as "probably carcinogenic to humans", which is its Group 2A.³ It was the second time: the first assessment, in 2007, reached the same group.⁵ This course did not read the monograph. What it has is the classification's grades as secondary summaries report them, and the announcement from ISGlobal, the Barcelona institute whose researcher Manolis Kogevinas chaired the working group's epidemiology subgroup. That's an institute reporting its own researcher's part, and worth knowing once.⁵

The grades, as the summaries give them: limited evidence in humans, with positive associations reported for cancers of the breast, prostate, colon and rectum, which are "associations" in lesson 3's sense of the word; sufficient evidence in experimental animals for alteration of the light and dark schedule; and strong mechanistic evidence.³

What Group 2A says. In the announcement's words, "The IARC classifies agents into four risk categories on the basis of the strength of the evidence on whether the substance or agent causes cancer. Group 2A is the second highest category in terms of the level of certainty of the evidence."⁵ Put in the working group's own terms, having looked at the human, animal and mechanistic evidence together, they judged that night shift work probably causes cancer in people. The word "probably" is theirs. That is a judgement about a literature.

What Group 2A does not say is how much anybody's risk changes. No number attaches to the label, this course has read none, and a classification and a risk estimate are different kinds of thing.

Kogevinas put both halves in one breath: "We consider that the evidence suggests a link between night shift work and breast, prostate and colorectal cancer. The evidence in animals is conclusive, but the mechanisms underpinning the association between night shift work and cancer in humans have still not been fully established."⁵ This course can't fully reconcile "strong mechanistic evidence" with that caveat from what it read. The grade covers the mechanistic evidence as a whole, and the caveat is about humans in particular.

He also said what he thinks follows. The category, in his words, "implies that more research is needed and also that preventive measures should be taken", and the measures he named were to "avoid long-term night shift work over many years and also eliminate shifts that require people to work many nights a week", while saying in the same breath that "we need people to work night shifts in many sectors".⁵ That's one working-group chair's view about how schedules are designed, reported by his own institute. It isn't the classification, and it isn't advice to anybody about their job.

This course has no advice for anybody who works nights. Nurses, drivers, bakers and police officers mostly don't choose their schedules, and a course that treated night work as a lifestyle error would be making a claim about people's lives that nothing in its sources supports.

Five things people get wrong about timing

"Night owls are just undisciplined." Chronotype is where your clock settles relative to light, and its gold standard is a hormone assay, not the time you choose to get up.¹ Where anybody's comes from is outside what this course read, and the light section above shows the clock does move with light, so this lesson can't tell you how much of anybody's lateness is choice. What it can say is that the same bedtime falls at different points in two people's biological night.

"Light before bed is bad for you." The evidence here supports something narrower. Bright light early in your biological night delays the clock,² which is the wrong direction if you're trying to sleep earlier and may be the right one before a night shift. What this course can't tell you is whether the light in your bedroom is bright enough to matter, because nothing here measured it.

"My body clock is fixed, so there's nothing to be done about when I sleep." One hour of bright light moved the clock across a fitted range of over two hours in a laboratory.² Whether the light you actually get moves yours enough to matter is what nobody here measured.

"Social jetlag is a diagnosis." Social jetlag is a researcher's name for the difference between your mid-sleep on work days and on free days. This course holds that definition at search-summary level only, has no figure for how large it usually is, and has read nothing linking it to any health outcome.⁴ It's a measurement, not a condition.

"Group 2A means night work gives you cancer." It grades the strength of the evidence that something can cause cancer, and carries no risk figure at all.³

Practice

Before the exercise, one worked through, with a step left for you.

Connections

Back. Lesson 1 named four instruments and this lesson adds a fifth, a hormone assay under constant routine, which measures a different quantity from any of them. Lesson 3's reading of "associated with" is what the IARC grading needs.

Forward. Lesson 8 is where the clock question joins the institute's standing sort, and it is the question this lesson exists to hand over: before anything else, ask whether a claim is about how much or about when.

Go deeper

• Human phase response curve to a 1 h pulse of bright white light, St Hilaire and colleagues, 2012. Open access. This course read the abstract in full and one sentence of the introduction. Worth it for the design alone: eighteen circadian phases, one hour of light each, and a hormone measured before and after under constant routine.
• The µMCTQ paper, Ghotbi and colleagues, 2020. Open access. This course read the abstract and the section defining the measure. It shortens the standard Munich ChronoType Questionnaire from 17 questions to 6. The full questionnaire wasn't read here, and if you want to answer it properly, that is where to go rather than to any summary of it.

Sources

1. Neda Ghotbi and colleagues, "The µMCTQ: An Ultra-Short Version of the Munich ChronoType Questionnaire", Journal of Biological Rhythms 35(1), 2020, pp. 98 to 110, doi 10.1177/0748730419886986. Open access. Read in part: the abstract and the section defining the measure, from the open PDF. The validation results were not read, which the body says. Supports: the definition of chronotype and phase of entrainment, dim-light melatonin onset as the gold standard, the word "approximating" about questionnaires, the mid-sleep marker, and the existence of the oversleep correction. The formula for that correction is not printed here, because this course did not read it.
2. Melissa A. St Hilaire and colleagues, "Human phase response curve to a 1 h pulse of bright white light", The Journal of Physiology 590(13), 2012, pp. 3035 to 3045. Read at abstract level, plus one sentence of the introduction. Supports: the 8000 lux, the one-hour exposure at eighteen circadian phases, the 2.20 hour fitted peak-to-trough amplitude, the 40 per cent against 15 per cent comparison and the word non-linear, and the direction of the shift in early and late biological night.
3. International Agency for Research on Cancer, IARC Monographs Volume 124, "Night Shift Work", 2019, and the working group's summary, "Carcinogenicity of night shift work", The Lancet Oncology, 2019, doi 10.1016/S1470-2045(19)30455-3. Neither was read. The classification's grades reach this course through secondary summaries, which the body says. Supports: the Group 2A classification, the three evidence grades and the four cancer sites named. No risk figure appears in this lesson because none was read.
4. The course's own constructions, each labelled where it appears in the body. Ana and Ben are invented and the body says so. The reading of "biological night" as meaning that one clock time is a different circadian phase in two people is this course's, said as such at the point of use. The five-and-a-half-hour figure for the longer exposure is this course's arithmetic on the abstract's own two figures and is labelled as such where it appears. The expectation that most light advice specifies neither timing nor intensity is an expectation, and the exercise says so. So is the opening remark that little of what people are told about sleep is about timing. The glosses of circadian and biological night are plain-language background, not quotations. The crossing point on the curve is the course's reading of its shape, and the guess that most people on a conventional schedule wake in their late biological night is a guess. The definition of social jetlag is held at search-summary level only, from summaries rather than from the paper that introduced the term, and no figure or outcome is given for it because none was read.
5. ISGlobal, "IARC Retains 'Probably Carcinogenic' Classification for Night Shift Work", 15 July 2019. Read in full. An institute's announcement about its own researcher, Manolis Kogevinas, who chaired the working group's epidemiology subgroup, which the body says. Supports: the four categories and where Group 2A sits, the 2007 first assessment, and every quotation from Kogevinas, including the mechanisms caveat and his view on preventive measures.