Sleep

Chang and colleagues randomised 12 adults to read from an iPad or a printed book for four hours before bed over five consecutive nights. The iPad suppressed evening melatonin by 55%, delayed melatonin onset by 1.5 hours, shifted circadian phase by 1.5 hours, reduced REM sleep and produced significantly greater next-morning sleepiness even after eight hours in bed. The printed book produced none of these effects.

The mechanism is photosensitive retinal ganglion cells containing melanopsin, which are maximally sensitive to the 446-477 nanometre blue light wavelength that LED screens emit in abundance. These cells project directly to the suprachiasmatic nucleus, the master circadian clock, suppressing pineal melatonin synthesis. A screen at arm's length delivers sufficient irradiance to produce the full suppressive effect.

University of Colorado sent participants camping for one week with no artificial light. Melatonin onset shifted two hours earlier. Even self-identified night owls normalised.

This is because the suprachiasmatic nucleus responds to light intensity detected by melanopsin-containing retinal ganglion cells. Natural light at intensities indoor lighting cannot match.

Xie and colleagues discovered in 2013 that the interstitial space of the brain expands by 60% during sleep, allowing cerebrospinal fluid to flush through neural tissue and clear metabolic waste. This glymphatic system removed amyloid-beta, the protein implicated in Alzheimer's disease, twice as fast during sleep as during wakefulness. The mechanism depends on noradrenergic tone: when locus coeruleus activity drops during sleep, astrocyte endfeet relax and open the drainage channels.

A single night of sleep deprivation produced measurable increases in amyloid-beta accumulation in the human hippocampus and thalamus, as demonstrated by Shokri-Kojori using PET imaging. The glymphatic system operates primarily during deep sleep and is impaired by sleeping position, alcohol consumption and head elevation. The brain produces waste continuously but can only remove it efficiently during one specific physiological state.

Adenosine is a somnogenic metabolite that accumulates in the basal forebrain during wakefulness, creating homeostatic sleep pressure. Caffeine is a competitive adenosine A1 and A2A receptor antagonist with a mean half-life of 5.7 hours (range 1.5 to 9.5 hours depending on CYP1A2 genotype). Drake et al. (2013, JCSM) showed that 400mg caffeine consumed 6 hours before bed reduced total sleep time by 41 minutes and reduced sleep efficiency, despite subjects reporting "no effect" on sleep quality.

This is because caffeine does not supply ATP or metabolic energy. It blocks the receptor that detects tiredness. Adenosine continues accumulating behind the blockade. When caffeine is metabolised, the full accumulated adenosine load hits the receptors simultaneously, producing the "crash." Chronic caffeine consumption upregulates adenosine receptors (Svikis et al., 2005, Psychopharmacology), requiring higher doses for the same effect and producing withdrawal fatigue that is indistinguishable from genuine sleep deprivation.

A 2pm coffee (peak plasma at 2:45pm) maintains 25% receptor occupancy at 10pm and 12% at 1am. Objective polysomnography detects disruption that subjective reports miss: reduced slow-wave sleep depth, increased stage transitions and delayed REM onset. The person "functions fine on coffee" is a person who has forgotten what rested actually feels like.

Leproult and Van Cauter restricted ten healthy young men aged 24 on average to five hours of sleep per night for one week. Daytime testosterone fell 10 to 15%, a magnitude equivalent to 10 to 15 years of normal ageing compressed into seven nights. The lowest testosterone levels occurred in the afternoon and evening, precisely when men are most likely to exercise, socialise and be sexually active.

Testosterone secretion is pulsatile and entrained to the sleep cycle, with the largest pulse occurring during the first episode of slow-wave sleep. Penev confirmed the dose-response relationship: total sleep time correlated with morning testosterone levels at r = 0.31, with men sleeping under 6 hours showing levels comparable to men 15 years their senior. Sleep is not rest. It is the primary endocrine manufacturing window.

The largest growth hormone pulse occurs in the first 90 minutes after sleep onset. Cutting from 9 to 7 hours costs disproportionately more because the final hours contain critical slow-wave phases.

For those under 25 with open growth plates, 9 to 10 hours is the protocol. The growth window closes permanently. Every night of insufficient sleep is permanently lost potential.

The International Agency for Research on Cancer classified night shift work as Group 2A, "probably carcinogenic to humans," in 2019 based on limited evidence of breast cancer and sufficient evidence of circadian disruption in experimental animals. A meta-analysis of 2,062,567 participants across 42 studies found shift workers had 9% higher overall cancer incidence, 19% higher breast cancer risk and 27% higher risk of digestive tract cancers.

Vyas and colleagues separately found shift work associated with a 23% increased risk of myocardial infarction and a 5% increase in ischaemic stroke across 34 studies covering 2,011,935 participants. Approximately 14% of the UK workforce regularly works nights. The circadian system regulates cell division, DNA repair, immune surveillance and tumour suppression. Chronic misalignment of these processes over decades produces measurable disease.

Slow-wave sleep declines approximately two percent per decade from early adulthood, resulting in a cumulative 70% reduction by the seventh decade. Mander and colleagues demonstrated this decline in slow-wave activity causally drives cognitive deterioration independent of age. Disrupted slow-wave activity in the medial prefrontal cortex predicted impaired overnight hippocampal-dependent memory consolidation and increased amyloid-beta deposition.

The relationship is bidirectional. Amyloid plaques preferentially accumulate in the same prefrontal regions that generate deep sleep, degrading the very mechanism that would clear them. Modern sleep disruption from screens, artificial light and caffeine accelerates this process decades before it would otherwise begin. The brain requires deep sleep to clean itself and the cleaning system is failing earlier in each generation.

Kitamura et al. (2016, Scientific Reports) restricted volunteers to 5.5 hours of sleep for 5 nights then allowed ad libitum recovery sleep. Cognitive performance required 14 days of unrestricted sleep to return to baseline, despite subjects reporting feeling recovered after 2 to 3 days. Van Dongen et al. (2003, Sleep) showed that chronic restriction to 6 hours per night for 14 days produced cognitive impairment equivalent to 2 full nights of total sleep deprivation, while subjects rated themselves as only "slightly sleepy."

This is because acute sleep debt depletes glycogen stores, adenosine clearance capacity and BDNF-dependent synaptic consolidation. The brain can partially compensate through increased sleep intensity (more delta power) during recovery, but the time required for full restoration exceeds the debt period by 2 to 3 fold. Six hours of sleep for 6 months cannot be repaired by one weekend. The molecular damage to telomeres (Prather et al., 2015, PLoS ONE), inflammatory markers and cortisol rhythms accumulates independently of perceived sleepiness.

The subjective adaptation to sleep restriction is the most dangerous feature: the impaired brain loses the capacity to detect its own impairment. Van Dongen's subjects at day 14 of 6-hour sleep performed as badly as subjects after 48 hours awake, but rated themselves as minimally affected. The sleep-deprived population does not know it is sleep-deprived.

Wang et al. (2014, Occupational and Environmental Medicine) meta-analysed 12 studies totalling 226,652 participants and found that shift workers had a 42 per cent increased risk of metabolic syndrome (RR 1.42, 95% CI: 1.20 to 1.67) compared to day workers. The risk was dose-dependent: each five-year increment of night shift work increased type 2 diabetes risk by 9 per cent (Pan et al., 2011, PLoS Medicine, following 177,000 nurses for 18 to 20 years).

This is because circadian misalignment disrupts the temporal coordination of insulin secretion, glucose tolerance, leptin, ghrelin and cortisol rhythms. Scheer et al. (2009, PNAS) induced circadian misalignment in healthy volunteers and produced a 6 per cent increase in glucose (despite 22 per cent higher insulin), a 17 per cent decrease in leptin and complete inversion of the cortisol rhythm within three days. The metabolic damage begins immediately.

Approximately 3.2 million people in the UK work night shifts (ONS Labour Force Survey, 2022). The NHS is the single largest night-shift employer in the country. Night shift workers are significantly over-represented in lower socioeconomic groups, creating a health inequality that compounds the diet and stress disadvantages already experienced by these populations. IARC classified night shift work as Group 2A (probable carcinogen) in 2019.

Ebrahim et al. (2013, Alcoholism: Clinical and Experimental Research) meta-analysed 27 studies and found that even moderate alcohol (1 to 2 drinks) reduced REM sleep by 9.3% of total sleep time in the first half of the night, with rebound disruption in the second half. At higher doses (>4 drinks), total REM suppression reached 20 to 39%. REM onset latency increased and REM fragmentation doubled. EEG power in the delta and theta bands was artificially elevated, mimicking deep sleep without its restorative functions.

This is because alcohol is a GABA-A receptor agonist that suppresses the cholinergic REM-generating nuclei in the pons (pedunculopontine and laterodorsal tegmental nuclei). Alcohol-induced "sleep" is pharmacological sedation: consciousness is abolished but the neurotransmitter cycling between NREM and REM that consolidates memory, processes emotion and prunes synapses is disrupted. Yoo et al. (2007, Nature Neuroscience) showed one night of REM deprivation increased negative emotional reactivity by 60%.

The cultural belief that a nightcap aids sleep is pharmacological illiteracy. Alcohol reduces sleep onset latency by 4 to 7 minutes (the perceived benefit) while destroying the REM architecture that makes sleep restorative (the hidden cost). A person who drinks 3 to 4 nights per week loses approximately 90 minutes of REM per week, equivalent to losing an entire night's REM every two weeks.

Cajochen et al.'s randomised controlled trial found that five hours of LED-backlit screen exposure in the evening reduced melatonin levels by twenty-two per cent compared to non-LED controls, with significant phase-delay effects on sleep onset.

LED screens emit short-wavelength blue light in the 450-480nm range, which has the highest sensitivity for intrinsically photosensitive retinal ganglion cells. These cells signal the suprachiasmatic nucleus directly, suppressing pineal melatonin synthesis.

The average British adult now spends over four hours per evening on screens. The accumulated circadian disruption across a lifetime of screen use represents a chronic, self-administered melatonin suppression treatment without medical oversight.