Why Regular Eating Works: Hunger, Hormones and the Binge-Restrict Cycle

Most eating advice is about what to eat. This is about when.

Regular eating means eating something every three to four hours while you're awake. For most people that's breakfast, a mid-morning snack, lunch, an afternoon snack, dinner and an evening snack. Written down like that, it looks almost too plain to bother with. But it's the first practical step in the best-studied talking therapy for binge eating, and there's a surprising amount of research behind it: on the binge-restrict cycle, on hunger hormones that learn your routine, on metabolism, and on what happens to appetite when people come off a GLP-1 medicine.

We've gone through that research properly here, from the original papers wherever we could. Some of it is solid. Some of it is a sensible idea that's only been tested in small studies. And a few popular claims, "starvation mode" among them, don't hold up the way they're usually told. We'll say which is which.

What regular eating means

The clearest description comes from enhanced cognitive behaviour therapy, or CBT-E, the eating disorder treatment developed by Fairburn and colleagues at Oxford. Early in treatment, usually by the third session, people are asked to eat three planned meals and two or three planned snacks each day, so that there's rarely more than about four hours between them, and to keep their eating to those times (Murphy et al., 2010).

A couple of details surprise people. The treatment lets people choose what they eat. The one condition in the manual is that meals and snacks aren't followed by vomiting or laxatives (Murphy et al., 2010). So regular eating is about timing, and changes to the food itself, if they're needed at all, come later. The pattern is also meant to take priority over other things in the day, while staying flexible enough to move a meal around a meeting.

The NHS, the UK's public health service, describes the same approach. Its page on treating binge eating disorder says CBT will help you "plan out the meals and snacks you should have during the day, to help you adopt regular eating habits", and that you shouldn't try to diet during treatment, "because this can make it more difficult to stop binge eating" (NHS). The Centre for Clinical Interventions in Western Australia, which publishes free patient handouts, sums up the work involved as plan, prepare and prioritise (CCI).

Why treatment starts here

In CBT-E, regular eating is introduced in the first stage of treatment, around eight sessions over four weeks, and the treatment's authors describe the changes made in that stage as "the foundation on which other changes are built" (Murphy et al., 2010). They give reasons for putting it first. It tackles a form of dieting that's easy to miss, which they call delayed eating: putting off eating for as long as you can. It displaces most episodes of binge eating. And it gives the day a structure. There's a practical logic too. How much people change early in treatment is a good predictor of how well they do overall, so it helps to start with something concrete that can shift quickly.

National guidance builds this in. In the UK, NICE, which writes the treatment standards the NHS follows, tells clinicians giving adults one-to-one CBT for binge eating disorder to "advise people to eat regular meals and snacks to avoid feeling hungry", and to advise them not to try to lose weight, for example by dieting, during treatment, "because this is likely to trigger binge eating". For bulimia, the first phase of CBT focuses on "establishing a pattern of regular eating" (NICE NG69, recommendations 1.4.7 and 1.5.5).

Does the regular eating part do anything by itself, or is it just the first box on the form? A few studies have tried to find out. In an Oxford trial of 130 adults with a range of eating disorders, researchers tested which changes carried the benefit of CBT-E. Regular eating was the one the data backed: it mediated the fall in binge eating (Sivyer et al., 2020). In a smaller US study of 38 people doing guided self-help, weeks when people kept more closely to three meals and two or three planned snacks had fewer binges (Zendegui et al., 2014). And among 173 people with binge eating disorder, the third who ate three meals a day binged less than the rest, though a snapshot like that can't tell you which came first (Masheb and Grilo, 2006).

CBT-E as a whole works well for a lot of people. In that same Oxford trial, 65.5% of the people who had CBT-E met criteria for remission by the end of treatment, compared with 33.3% of those who had interpersonal therapy (Fairburn et al., 2015). A 2017 review of 79 trials found that therapist-led CBT did better than waiting lists and than other talking therapies taken together, for bulimia and binge eating disorder, though many of the trials were of poor quality (Linardon et al., 2017). The rest of the Fairburn model is in our guide to CBT for eating disorders.

The binge-restrict cycle, and where it comes from

Most people who binge don't only binge. There's usually a restrictive half to the pattern, and psychologists have been studying it since the 1970s.

In 1975, Herman and Mack gave people a set amount of food before letting them eat freely. People who scored low on a dieting questionnaire ate less afterwards when that first portion had been bigger, which is what you'd expect. People who scored high did the opposite: the bigger the first portion, the more they went on to eat (Herman and Mack, 1975). A decade later, Polivy and Herman set out the case that the dieting usually comes first and the bingeing follows (Polivy and Herman, 1985).

Evidence from outside the lab points the same way. Former prisoners of war who had lost a lot of weight in captivity were more likely to binge after the war than combat veterans who hadn't been starved (Polivy et al., 1994). In a five-year study of 496 teenage girls, going a whole day without eating to control weight predicted who went on to develop binge eating and bulimia, and generally predicted it more strongly than self-reported dieting did (Stice et al., 2008). When 133 women with bulimia logged their eating in real time for two weeks, the odds of a binge went up on days they'd restricted, and on the following day (Zunker et al., 2011). In another real-time study, the longer a run of restriction lasted, the better it predicted a binge (Holmes et al., 2014).

Hunger isn't the only trigger. Studies that ask people how they feel several times a day find that low mood tends to rise before a binge, and then gets worse afterwards (Haedt-Matt and Keel, 2011). That second part is the guilt that feeds the next round of rules. It's why CBT works on thoughts and feelings as well as eating, and why regular eating is the start of treatment rather than the whole of it. Our piece on the all-or-nothing guilt cycle goes into that side.

Eating less doesn't automatically lead to bingeing

Here's the part the popular version leaves out. When adults with obesity follow supervised low or very low calorie programmes, binge eating usually falls in the people who were bingeing beforehand. In people who weren't, a systematic review found no consistent pattern: some programmes saw binge eating rise, some saw no change and some saw it drop (da Luz et al., 2015). The questionnaires the older studies used to measure "dieting" also turn out to be poor at measuring how much people actually eat (Stice et al., 2010).

So the risk doesn't seem to sit in eating less as such. It sits in particular ways of doing it: long gaps, whole days without food, and rigid rules that snap. In large German samples, including tens of thousands of people in a weight-loss programme, rigid all-or-nothing control went with more frequent and more severe binges, and flexible control went with fewer (Westenhoefer et al., 1999).

That matters if you're on a GLP-1, where eating less is part of the plan. The things worth avoiding are the long gaps and the brittle rules.

What the Minnesota Starvation Experiment does and doesn't show

The most famous study in this area is also the most misquoted.

Towards the end of the Second World War, 36 young men, conscientious objectors, volunteered for a study at the University of Minnesota led by Ancel Keys (Kalm and Semba, 2005). The aim was practical: to learn how to bring starving people in post-war Europe back to health. The men spent 24 weeks on a semi-starvation diet, followed by 12 weeks of restricted refeeding and then a period of eating freely (Dulloo and Jacquet, 1998); (Dulloo et al., 1997). Most lost a large share of their body weight, and many became tired, apathetic, irritable and very weak (Kalm and Semba, 2005). The full results fill a two-volume book, The Biology of Human Starvation, published in 1950 (Keys et al., 1950).

What matters for eating is what happened in people's heads and at the table. A review of the research on starvation and self-imposed dieting describes preoccupation with food and eating, bigger emotional swings, poor concentration, and eating binges once food became available again (Polivy, 1996). When the Minnesota men were finally allowed to eat as much as they wanted, the overeating carried on until both their fat and their lean tissue had recovered, and body fat overshot where it had started (Dulloo et al., 1997). Decades later, interviews with 19 of the men found that many had kept abnormal eating habits for months, and some for years (Eckert et al., 2018).

What it shows: severe, prolonged food restriction can create food obsession and binge eating in people who had neither before. That's why eating disorder clinicians still teach it.

What it doesn't show: this was a small group of young men, and no women, on a severe diet for six months, compared against their own earlier selves rather than a separate control group. It tells us nothing directly about skipping lunch or eating a bit less than usual. It can't separate the effect of hunger from the effect of spending most of a year inside a demanding wartime study. And it doesn't reproduce anorexia: the preserved energy and high activity that many people with anorexia show despite weight loss weren't seen in the Minnesota men (Eckert et al., 2018). Treat it as proof of what the body does under extreme conditions, and not as a forecast of what one missed meal will do to you.

Your hunger runs on a timetable

Hunger isn't only an empty stomach. It's a set of signals, and some of them learn your routine.

The best known is ghrelin, a hormone made mostly in the stomach. In a 2001 study that took blood 38 times over 24 hours from ten people eating on a fixed schedule, ghrelin nearly doubled just before each meal and dropped to its lowest point within an hour of eating (Cummings et al., 2001). A follow-up found the same rise before meals that people started of their own accord, with no clocks or food cues around, and their hunger ratings moved with it (Cummings et al., 2004).

Ghrelin also seems to learn when you usually eat. People who habitually had lunch about three hours after breakfast had their ghrelin peak before that time, while people who usually waited about six hours peaked later, before their own lunch (Frecka and Mattes, 2008). In a small study of six people fasting for 24 hours, ghrelin still rose and fell around their usual mealtimes (Natalucci et al., 2005). And when habitual breakfast eaters skipped breakfast, they were hungrier before lunch, while habitual breakfast skippers weren't (Thomas et al., 2015).

Researchers still argue about exactly what ghrelin is doing. In the timing study, hunger tended to rise before ghrelin did, which suggests ghrelin may partly be anticipating the meal rather than causing the hunger (Frecka and Mattes, 2008). Either way, the practical point stands: your appetite signals adapt to the pattern you give them. That fits with the advice in clinical handouts to start out eating "by the clock" (CCI).

On the other side are the fullness signals released as food moves through the gut, including cholecystokinin (CCK), GLP-1 and peptide YY. CCK is well established as one of the body's own brakes on eating, at least in people of a healthy weight. The roles of the others are still being worked out (Steinert et al., 2017). GLP-1 is the one the new weight-loss medicines copy. Semaglutide, sold for weight management as Wegovy, acts on the GLP-1 receptor. Tirzepatide acts on both the GIP and GLP-1 receptors, and is sold as Mounjaro in the UK, Ireland and Australia, and as Zepbound for weight management in the United States. Both are built to last for days in the body, where the body's own GLP-1 is cleared quickly (Wegovy SmPC); (Mounjaro SmPC).

Three meals compared with grazing

One tightly controlled Dutch trial shows how a meal pattern shapes hunger across a day. Twelve healthy young men spent 36 hours in a metabolic chamber on two occasions, eating identical food either as three meals or as 14 small meals, one every hour. With three meals, hunger swung: high before each meal and low after it. With hourly grazing, it sat in the middle all day, and overall they were hungrier and less satisfied (Munsters and Saris, 2012). A US trial comparing three and six meals a day found the same direction: six meals didn't burn more fat, and left people hungrier and keener to eat (Ohkawara et al., 2013). A review of controlled feeding studies concluded that eating more than three times a day does little for appetite, while eating fewer than three times makes appetite control worse (Leidy and Campbell, 2011).

This clears up a common mix-up. Regular eating often gets described as "eat little and often to keep hunger away". The trials suggest grazing all day doesn't calm hunger. What the CBT-E pattern gives you is predictability and no long gaps, with clear stops in between, which is why the manual also asks people to keep their eating to the planned times (Murphy et al., 2010).

What a long gap does, and what it doesn't

If you've ever gone most of a day without eating and then eaten everything in sight in the evening, you'll want the research to say that long gaps cause overeating. It's more nuanced than that, and the nuance is useful.

In a University of Bath trial, 35 lean adults came into the lab on two mornings. On one they ate a typical breakfast, and on the other they didn't. By noon, appetite on the no-breakfast morning was well above where it was on the breakfast morning. They ate more at lunch, but not enough to make up for the missed breakfast, and by mid-afternoon their appetite was the same on both days (Chowdhury et al., 2015).

That's the usual finding in healthy volunteers. Skipping a meal makes you hungrier and makes the next meal bigger, but people don't usually eat back everything they missed (Levitsky and Pacanowski, 2013); (Clayton and James, 2016). And for weight loss specifically, a meta-analysis in the BMJ found that adding breakfast doesn't help (Sievert et al., 2019).

So why does regular eating help people who binge? Because the lab studies measure healthy volunteers on a single day, with no rules riding on the outcome. The real-time studies earlier in this piece measured people who restrict again and again, hold rigid rules and already binge, and in them, restriction and long gaps do come before binges (Zunker et al., 2011); (Holmes et al., 2014). Regular eating is aimed at that second group. For someone without a binge-restrict pattern, a late lunch is just a late lunch.

Does meal timing change your metabolism?

A small set of controlled trials, most of them from the University of Nottingham, has tested regularity itself: eating the same number of times each day compared with letting it vary.

In the first, nine lean women ate their normal food for two weeks either six times a day or anywhere from three to nine times a day, then swapped. After the irregular fortnight, their insulin response to a test meal was higher, and so were their fasting total and LDL cholesterol (Farshchi et al., 2004). The same design in ten women with obesity found that the regular pattern came with a bigger rise in the energy used to digest food (the thermic effect), lower cholesterol and a lower insulin response, and the women reported eating less (Farshchi et al., 2005). Later trials, where all the food was provided so intake was identical, again found a smaller thermic effect after the irregular pattern, in 11 normal-weight women (Alhussain et al., 2016) and in nine women with obesity and insulin resistance (Alhussain et al., 2022). In the first of those, the women also felt less hungry and fuller around meals during their regular fortnight.

In the other direction, a US trial had healthy, normal-weight, middle-aged adults eat all their food in one evening meal instead of three meals, with the same total amount, for eight weeks. On one meal a day they were hungrier and had higher blood pressure and cholesterol, though they also lost a little body fat (Stote et al., 2007). Their morning blood sugar control got worse too (Carlson et al., 2007). The American Heart Association reviewed the area in 2017 and concluded that irregular eating patterns "appear less favorable" for heart and metabolic health (St-Onge et al., 2017).

How much should you read into this? Not a lot on its own. The trials are tiny, mostly in women, and last two weeks. The differences in the thermic effect are small, and none of these trials followed weight over months. They're reasonable evidence that the body handles food a little better on a predictable schedule, and weak evidence for anything bigger.

These trials tested regularity. Eating more often is a separate question, and the answer there is clearer: it doesn't speed up your metabolism or help with fat loss (Schoenfeld et al., 2015); (Ohkawara et al., 2013).

"Starvation mode": what's real and what isn't

You'll often hear that if you don't eat enough, your body goes into "starvation mode": your metabolism shuts down and you stop losing weight, or even gain. Some patient handouts use the phrase too. The reality is more interesting, and more useful.

What's real

When people lose weight and hold it, their bodies burn somewhat less energy than their new size would predict. This was measured carefully in the 1990s (Leibel et al., 1995), and it can last more than a year (Rosenbaum et al., 2008). Appetite hormones shift as well. A year after a diet, people in an Australian study still had higher ghrelin, lower levels of several fullness hormones and more hunger than before they started (Sumithran et al., 2011).

What's argued about

How big that slowdown is, and how long it lasts. One study found it roughly halved once people's weight had stabilised, had gone by a year, and didn't predict who regained (Martins et al., 2020). The well-known "Biggest Loser" study found a large slowdown six years after the competition (Fothergill et al., 2016), but one of its authors has since argued that the contestants' very high activity levels explain much of it (Hall, 2022).

What doesn't hold up

The idea that eating too little stops weight loss altogether, or that one skipped meal turns your metabolism down. In the Bath trial, six weeks of skipping breakfast didn't change resting metabolism (Betts et al., 2014).

The bigger force is appetite

In a study of people losing weight on a diabetes drug that makes the body pass sugar in the urine, so they weren't aware of the energy deficit, appetite rose in proportion to the weight they lost, and that pull was more than three times bigger than the drop in energy burned (Polidori et al., 2016).

So here's the useful version of "starvation mode": your body defends against weight loss mostly by making you hungrier. That's the same pressure that can tip people into the binge-restrict cycle, and part of the reason a steady pattern is worth having.

Energy, mood and concentration

This is where popular claims run furthest ahead of the evidence. You'll read that regular eating keeps your blood sugar steady, so you don't get tired, irritable and foggy. Here's what's actually known.

Hunger and mood are linked. When 64 adults in Central Europe rated their hunger and feelings five times a day for three weeks, hungrier moments came with more anger and irritability and less pleasure, even after taking account of things like age, eating habits and how angry a person tends to be in general (Swami et al., 2022). "Hangry" holds up.

Breakfast gives a small lift to memory. Across 38 studies in adults, eating breakfast gave a small but consistent advantage for memory, while the effects on attention and other kinds of thinking were mixed (Galioto and Spitznagel, 2016).

People who eat breakfast move more in the morning. In the Bath trials, people eating breakfast were more physically active during the morning than people fasting (Betts et al., 2014); (Chowdhury et al., 2016). It's an indirect sign of energy, but it was measured rather than asked about.

Blood sugar is steadier than the "crash" story suggests. In people without diabetes it stays in a fairly narrow range whatever the pattern. In the Dutch trial, three meals produced bigger swings in blood glucose than hourly eating, but on a continuous monitor the average daytime lows and highs were almost the same for both, and within the normal range (Munsters and Saris, 2012). What may matter more is the dip a few hours after eating. In a study of more than a thousand people in the UK and US, those whose glucose dipped further two to three hours after a standard meal felt hungrier sooner and ate more over the next day. The links were real but modest, and several of the authors work with the nutrition company Zoe (Wyatt et al., 2021).

The afternoon slump is partly your body clock. The post-lunch dip in alertness happens even when people haven't eaten lunch, though a big, high-carbohydrate lunch can make it worse (Monk, 2005).

So the fair summary is this. Regular eating probably helps mood mostly by keeping you out of the very hungry zone. The blood sugar crash story is overstated for most people, and the evidence for better energy is real but modest.

On a GLP-1: when you're barely hungry

Regular eating was designed for people whose appetite pushes them to eat. On a GLP-1 medicine, the problem is often the other way round.

These medicines cut appetite hard. In a trial funded by the manufacturer, people on semaglutide 2.4 mg ate 35% less at a help-yourself lunch than people on placebo, and reported less hunger and fewer cravings (Friedrichsen et al., 2021). In an earlier trial, also industry-funded, intake across a whole day of free eating was 24% lower (Blundell et al., 2017). Tirzepatide reduced appetite in a similar way in people with type 2 diabetes (Heise et al., 2023). A 2026 review of diet studies in people taking semaglutide or tirzepatide found that intake fell by between about a quarter and two-fifths, that lean tissue made up as much as 40% of the weight lost, and that hardly any of the studies checked whether people were getting enough protein, vitamins and minerals (Spreckley et al., 2026).

A joint 2025 advisory from four US obesity and nutrition organisations lists nutrient shortfalls from eating less, and loss of muscle and bone, among the main problems these medicines bring, and recommends screening for disordered eating before starting (Mozaffarian et al., 2025). In the UK, the British Dietetic Association and the British Nutrition Foundation welcome prescribing them "alongside dietary support and physical activity" (BDA and BNF, 2024).

Eating disorder researchers have raised a specific worry. The same effect that reduces appetite and food preoccupation may, in vulnerable people, reinforce restriction and "avoidance of regular eating" (Škudar, 2026). A 2026 UCL-led meta-analysis adds a twist. Across trials, GLP-1 medicines reduced binge eating on average, which is good news for many people. They also raised dietary restraint, and the authors say it isn't clear whether that's healthy control or the rigid kind. All of the trials had some risk of bias (Emptage et al., 2026).

Structure helps here in a practical way. Planned eating occasions spread through the day make it easier to eat enough, and especially enough protein, when you can only manage a little at a time. A small study found that when protein was spread evenly across three meals, the muscles made more new protein over 24 hours than when most of it came at dinner (Mamerow et al., 2014). Our guide to protecting muscle on a GLP-1 covers protein and strength training in detail.

If you have a history of an eating disorder and you're on one of these medicines, or thinking about it, tell your prescriber. Screening is exactly what specialists are asking for (Mozaffarian et al., 2025); (Škudar, 2026).

Coming off: when appetite comes back

Stopping is where regular eating matters most for a lot of our readers. Semaglutide has a half-life of about a week and is still in the circulation for around seven weeks after the last 2.4 mg dose. Tirzepatide's half-life is about five days (Wegovy SmPC); (Mounjaro SmPC). So appetite doesn't switch back on overnight. But it does come back.

The stopping trials show what happens to weight. In the STEP 1 extension, people who stopped semaglutide, and the lifestyle programme that came with it, regained about two-thirds of the weight they had lost within a year, on average (Wilding et al., 2022). In SURMOUNT-4, people switched from tirzepatide to a placebo regained much of theirs, while people who stayed on it kept losing (Aronne et al., 2024). A 2026 Oxford-led meta-analysis of 37 studies found that regain after weight-loss medicines was faster than after behavioural weight-loss programmes (West et al., 2026).

Those trials measured weight, not hunger, so the appetite part is an inference. It's a well-grounded one. After weight loss by diet, appetite hormones stay tilted towards eating more for at least a year (Sumithran et al., 2011), and appetite rises in proportion to the weight lost (Polidori et al., 2016). Take away a medicine that was holding appetite down, and you'd expect to feel that pull.

This is a point where a binge-restrict cycle can start. For someone who's got used to eating very little, returning hunger can feel alarming, and the instinct is to clamp down harder. That's the first step of the loop described earlier, and regular eating is the first step of the first-line treatment for that loop in people who binge.

We want to be careful here. No trial has tested regular eating as a way of coming off a GLP-1, and we're not claiming it prevents regain. What we can say is narrower. The binge-restrict cycle is well documented, returning appetite is predictable, and regular eating is where treatment for the cycle begins. Setting the pattern up while the medicine is still doing some of the work means it's already in place when your appetite returns. Our coming-off playbook and our piece on what actually happens when you stop cover the rest, and any change to your dose is a conversation to have with your prescriber.

How to start

If you're in treatment, your therapist or dietitian will shape this with you, and their plan comes first. If you're not, and nothing in the note at the top applies to you, this is how the approach usually works.

Plan tomorrow the night before. Write down roughly when you'll eat, starting with breakfast and then something every three to four hours until bed. The times can move from day to day to fit your life. The gaps are what matter (Murphy et al., 2010).

Keep the food itself simple. In CBT-E, you choose what you eat (Murphy et al., 2010). This isn't the moment to cut foods out or shrink portions.

Eat at the planned time, even if you're not very hungry. Clinical handouts suggest eating "by the clock" at first (CCI). On a GLP-1, each eating occasion can be small.

Think about where tomorrow's long gaps might be: a long meeting, a commute, a day out. A snack in your bag covers most of them.

Make it a priority. The CBT-E manual says the eating pattern should come before other activities, while still bending to fit your commitments (Murphy et al., 2010). In practice that might mean moving a call, or eating at your desk for a few weeks.

If you miss one, have the next one. Don't skip the next meal to make up for a binge, and don't squeeze in extra to make up for a missed snack. Getting back on track at the next planned time is the skill (CCI).

Keep a simple record. CBT-E uses real-time self-monitoring: writing down what you ate, when, and what was going on, as it happens (Murphy et al., 2010). You don't need any numbers for it. A notes app works, and so does a food diary.

Plan, prepare, prioritise: the three words from the Centre for Clinical Interventions handout are a good summary of all of this. The handout also makes the point that the strict version doesn't have to last forever. As eating settles, it can become more flexible (CCI).

An example day, for someone who's up at about 7am and in bed by about 11pm:

Shift the whole thing to fit your day. There are no amounts on purpose.

What's solid, what's likely and what's contested

When to get more help

Regular eating is a clinical technique, and this article isn't a substitute for treatment. If you think you might have an eating disorder, talk to a doctor as soon as you can. That is the advice the UK's NHS gives (NHS), and the bar for asking is lower than most people assume: NICE tells clinicians not to use a single measure such as BMI to decide whether someone gets treatment (NICE NG69, recommendation 1.2.8), so you don't need to be underweight, or certain, to ask.

If you're underweight or have been eating very little for a long time, increasing your food needs medical monitoring. Guidance from the UK's Royal College of Psychiatrists on medical emergencies in eating disorders covers the risks of refeeding too fast and of refeeding too cautiously (RCPsych, 2022).

This is information, not medical advice. If you're being treated for an eating disorder, your care team's plan comes first.