# Why you crave junk food when you're tired

July 16, 2026 · Sleep & Weight · 8 min read · https://burnweek.fit/blog/sleep-and-cravings/

> They'd already had breakfast. Their ghrelin was up and predicted nothing. Sleep-deprived men still bought 9% more calories — so the pull isn't hunger at all.

**Key takeaways**

- It isn't hunger: given a standardized breakfast first, sleep-deprived men still bought 9% more calories and 18% more food by weight in a mock supermarket.
- Ghrelin rose after sleep deprivation — and the rise didn't correlate with what participants actually bought. The hunger-hormone story doesn't carry the behavior.
- Sleep loss moves two systems at once: amygdala activity amplifies while frontal and insular control activity drops. The wanting rises as the braking weakens.
- The endocannabinoid 2-AG peaks in the early afternoon; short sleep amplifies that peak and extends it, alongside a reduced ability to inhibit palatable snacks.
- Three different methods — a supermarket, a scanner, a blood assay — all describe failed inhibition, not raised appetite. Decide the afternoon before it arrives.

## It isn't hunger — they had already eaten breakfast

The tired pull toward chocolate and chips feels like hunger, and it mostly isn't. The neatest demonstration is a mock supermarket. Fourteen normal-weight men were kept awake for a full night, then handed a fixed budget of 300 Swedish kronor — about $50 — and told to buy what they could from 40 items, half of them calorie-dense and half not ([Chapman et al., 2013](https://pubmed.ncbi.nlm.nih.gov/23908148/)). The crucial detail is what happened first: they were given a standardized breakfast before shopping, specifically, in the authors' words, "minimizing the potential confound produced by hunger."

They shopped fed. And they still bought 9% more calories and 18% more food by weight than they did after a normal night's sleep (both P < 0.05), regardless of which foods were on offer or how the prices were manipulated. Their morning ghrelin — the hunger hormone that carries most of the blame in popular write-ups — was indeed elevated. It also failed to correlate with a single thing they bought.

> They had already eaten breakfast. Their ghrelin was up, and it predicted nothing. Whatever pulled them toward the calorie-dense shelf, it wasn't hunger.

That is a small study of fourteen men after total sleep deprivation, which is a harsher exposure than most real Tuesdays. But it isolates something the bigger literature keeps gesturing at: sleep loss changes what you *choose*, not just what you feel in your stomach. That the choosing adds up to real weight is the [pillar's](/blog/sleep-and-weight-loss) argument; this is the question of why it's chips and never carrots.

## What people think is happening, and what the evidence shows

| The popular explanation | What the studies actually found |
| --- | --- |
| "You're hungrier when you're tired" | Fed a full breakfast first, sleep-deprived men still bought 9% more calories ([Chapman, 2013](https://pubmed.ncbi.nlm.nih.gov/23908148/)) |
| "It's ghrelin — the hunger hormone" | Ghrelin rose, and the rise did not correlate with what they purchased ([Chapman, 2013](https://pubmed.ncbi.nlm.nih.gov/23908148/)) |
| "You just have no willpower" | Frontal and insular control activity drops while amygdala activity amplifies ([Greer, 2013](https://pubmed.ncbi.nlm.nih.gov/23922121/)) |
| "It's all in your head" | A reward-signalling lipid, 2-AG, has its daily rhythm amplified and extended ([Hanlon, 2016](https://pubmed.ncbi.nlm.nih.gov/26612385/)) |

The pattern across that column is consistent, and it is not the one people expect. The appetite dial is not the thing that moved.

## The wanting goes up while the braking goes down

Brain imaging puts a mechanism under the shopping data. When sleep-deprived adults rated how much they wanted various foods inside an fMRI scanner, two things happened at once: activity fell in the frontal and insular cortex — the regions that evaluate and weigh a food against your intentions — while activity in the amygdala amplified. The pairing tracked a measurable increase in desire for high-calorie, weight-gain-promoting foods ([Greer et al., 2013](https://pubmed.ncbi.nlm.nih.gov/23922121/)).

Call it a bi-directional change, because that's the phrase that matters. It isn't that a tired brain wants food more. It's that a tired brain wants food more *and* is simultaneously worse at the arithmetic that normally talks it out of the biscuit. Both halves move, and they move in the directions that compound. The deeper imaging story — which circuits, which foods, how reliably it replicates — belongs to sleep and the food reward brain.

> Sleep loss doesn't just make you want the junk food more. It makes you worse at saying no to it — at the same time.

This reframes the moral question people bring to a badly slept Tuesday. "No willpower" is a description of the outcome, not a diagnosis of the cause. The brakes were measurably weaker.

## The chemical that makes the afternoon dangerous

There's a candidate carrier for all this, and it has a familiar cousin. The endocannabinoid system is a signalling network your body runs on its own — and it's the same receptor family that cannabis acts through, which is why its role in appetite is not a surprise to anyone who has heard of the munchies. Its main messenger is a lipid called 2-arachidonoylglycerol, mercifully abbreviated to 2-AG.

In a randomized crossover study, healthy young adults spent four nights on 8.5 hours of sleep and four on 4.5, with 24-hour profiles of circulating 2-AG measured under controlled conditions ([Hanlon et al., 2016](https://pubmed.ncbi.nlm.nih.gov/26612385/)). 2-AG turns out to have a robust daily rhythm in everyone: a trough in the middle of the overnight fast, then a steady climb to an early-afternoon peak. Sleep restriction didn't create that rhythm. It *amplified* it — with maximum values arriving later and lasting longer.

And the behavior tracked the chemistry. When sleep-deprived, participants reported more hunger and appetite in step with that afternoon elevation, and — the authors' phrasing again — were "less able to inhibit intake of palatable snacks."

*Less able to inhibit.* That is the third independent study reaching for the language of failed braking rather than raised appetite. A supermarket experiment, a brain scanner, and a blood assay each land on inhibition. When methods this different converge on the same verb, it's worth listening.

The practical read is that the tired-eating risk has a clock on it. It is not evenly spread across your day; it concentrates when 2-AG peaks, in the afternoon and beyond. Note this is a mechanism study in healthy young adults with intake assessed in a lab — it explains a pull, it doesn't predict your Wednesday.

## Why it's chips and never carrots

Put the three together and the selectivity makes sense. A reward system running hot, an evaluation system running cold, and a signalling molecule tuned specifically to *hedonic* eating — eating for pleasure rather than for fuel. None of those machinery pieces is interested in celery. They are interested in things engineered to be worth wanting: dense, salty, sweet, fatty, immediate. Fat is doing quiet work in that list — at [9 kcal per gram against protein's 4](/blog/macronutrients-explained), the foods a tired brain selects for are the ones that cost the most per mouthful. The pooled intervention data bears the selectivity out: the extra calories eaten after sleep restriction arrive as significantly more fat and significantly *less* protein, with carbohydrate unchanged ([Al Khatib et al., 2017](https://pubmed.ncbi.nlm.nih.gov/27804960/)).

This is also why a tired day's damage is disproportionate to its hunger. Where those extra calories land, how many there are, and what they add up to over weeks is the accounting in [how sleep loss drives weight gain](/blog/sleep-deprivation-and-weight-gain). And because ghrelin conspicuously failed to explain the shopping, the neat leptin-and-ghrelin story deserves more scepticism than it usually gets — the audit is in sleep and appetite hormones.

## What to do with a badly slept day

The useful move is to stop treating it as a character test. If the frontal cortex is measurably quieter, then "try harder this afternoon" is a plan that relies on the exact faculty the evidence says is impaired. Decisions made in advance don't need it.

So: decide the afternoon before the afternoon arrives, when the 2-AG curve is still at its trough and you're still the person who cares. Have something worth eating already in the building — protein is the practical lever here, since it's the macro that [does the most for fullness per calorie](/blog/protein-for-weight-loss) and the [foods that carry it well](/blog/best-protein-sources) are rarely the ones a tired brain reaches for by default. And keep [counting through the bad days rather than abandoning the log](/blog/how-to-count-calories), because a tired Tuesday is data, not a verdict — the pattern is only visible if you don't stop recording when it's unflattering.

One night of short sleep is not a moral event. It's a few hundred calories of predictable, time-locked, chemically-assisted temptation, and the reliable defense against something predictable is a decision you already made.

## FAQ

### Why do I crave sugar and junk food when I'm tired?

Because sleep loss moves the reward system and the control system in opposite directions at once: amygdala activity amplifies while frontal and insular evaluation activity drops, and desire for high-calorie foods rises with it ([Greer et al., 2013](https://pubmed.ncbi.nlm.nih.gov/23922121/)). A reward-signalling lipid called 2-AG also has its daily peak amplified and extended by short sleep, alongside a reduced ability to resist palatable snacks ([Hanlon et al., 2016](https://pubmed.ncbi.nlm.nih.gov/26612385/)). The selectivity is the point — none of that machinery is aimed at vegetables.

### Does sleep loss make you hungrier, or just worse at resisting?

The evidence leans hard toward resisting. Sleep-deprived men given a standardized breakfast beforehand still bought 9% more calories and 18% more food by weight, and their elevated ghrelin didn't correlate with any of it ([Chapman et al., 2013](https://pubmed.ncbi.nlm.nih.gov/23908148/)). Hunger rises too — but hunger alone doesn't explain the behavior, which is why three very different methods all describe the failure as one of inhibition.

### Is the afternoon really the worst time for tired snacking?

That's what the mechanism predicts. 2-AG follows a daily rhythm that bottoms out during the overnight fast and peaks in the early afternoon, and sleep restriction amplifies that peak and drags it later — with self-reported hunger and reduced snack inhibition tracking it ([Hanlon et al., 2016](https://pubmed.ncbi.nlm.nih.gov/26612385/)). It's a controlled lab study in healthy young adults, so treat it as a plausible clock rather than a personal schedule.

## Sources

- [Chapman CD, Nilsson EK, Nilsson VC, et al. Acute sleep deprivation increases food purchasing in men. Obesity (Silver Spring). 2013;21(12):E555-E560.](https://pubmed.ncbi.nlm.nih.gov/23908148/)
- [Greer SM, Goldstein AN, Walker MP. The impact of sleep deprivation on food desire in the human brain. Nat Commun. 2013;4:2259.](https://pubmed.ncbi.nlm.nih.gov/23922121/)
- [Hanlon EC, Tasali E, Leproult R, et al. Sleep restriction enhances the daily rhythm of circulating levels of endocannabinoid 2-arachidonoylglycerol. Sleep. 2016;39(3):653-664.](https://pubmed.ncbi.nlm.nih.gov/26612385/)
- [Al Khatib HK, Harding SV, Darzi J, Pot GK. The effects of partial sleep deprivation on energy balance: a systematic review and meta-analysis. Eur J Clin Nutr. 2017;71(5):614-624.](https://pubmed.ncbi.nlm.nih.gov/27804960/)