Cravings & Their Doings
Your Body Is Not Broken.
It Is Asking.
What science now understands about cravings — and why the language of willpower has never been the right translation.
There is a particular kind of afternoon I know very well. It is not quite hunger. It is more specific than that — a pull toward something sweet, something salty, something warm. It arrives around 3 pm, or in the quiet hour after dinner, or with a particular insistence in the days before my period. And for most of my life, I named it by the only word anyone had given me: weakness.
That was the frame I had inherited — from diet culture, from medical appointments where no one asked follow-up questions, from the general cultural assumption that a craving is a failure of discipline, an emotional shortcut, proof that you cannot be trusted around your own appetites. The solution offered was always the same: resist more. Want less. Override the signal and choose something virtuous instead.
I did this for years. And what I noticed, with increasing confusion, was that the signal never went away. It shifted, sometimes. It intensified, reliably, at certain points in the month. It responded poorly to being ignored, and responded well — briefly, imperfectly — to being met. But it never disappeared. It returned, as it always had, punctual as a tide.
It took reading a significant amount of nutritional neuroscience before I understood why. The signal was not the problem. The translation was.
A craving is not a character flaw wearing the disguise of physiology. It is physiology — specific, layered, driven by hormone fluctuations and neural reward circuits and learned associations so deeply embedded that they operate below the level of conscious choice. It carries information. And what it has to say, when you stop trying to silence it and start trying to read it, is often genuinely useful.
To understand cravings you need two maps: one of the brain, and one of your hormones across the month. Together they explain almost everything.
The brain map starts in the midbrain, in a region called the ventral tegmental area, where dopamine neurons originate. From there, a pathway called the mesolimbic circuit projects to the nucleus accumbens — the reward hub — and on to the prefrontal cortex. This is the neural architecture of wanting. When you catch the smell of something baking, or find yourself in the kitchen at ten o'clock the night before your period, this circuit lights up. Not because you are weak. Because you are human, with a reward system that evolved to motivate the pursuit of energy-dense food, and that does its job extremely well [1].
What the research has established is that the key hormones governing hunger — ghrelin, which rises to signal hunger and activate the dopamine system; leptin, which signals satiety — are not operating independently. They speak directly to the midbrain. Ghrelin activates dopamine neurons in the VTA. Leptin suppresses that same pathway. When you are sleep-deprived, chronically stressed, or in the luteal phase of your cycle, ghrelin rises and leptin signalling is blunted [1]. This is not weakness. This is a measurable, reproducible neurochemical state.
Anatomy of a craving — four layers
A smell, a time of day, an emotional state, a drop in blood sugar — any of these can act as a conditioned cue, activating ghrelin secretion even before food is present.
The VTA–nucleus accumbens circuit activates. Dopamine here encodes anticipation of reward. The craving is the wanting, not the having.
Ghrelin, leptin, estrogen, and progesterone all interact with this circuit. In the luteal phase, the hormonal environment turns the amplifier up — the same cue produces a stronger signal.
By the time it reaches conscious awareness, the neurochemical cascade is already well underway. This is the signal asking to be read, not resisted.
Cravings as learned, conditioned responses. A landmark review by Sun and Kober published in Physiology & Behavior established that food cravings are best understood as conditioned responses, in the Pavlovian sense. Cues that have been repeatedly paired with food — times of day, emotional states, hormonal fluctuations — come to reliably evoke the physiological and neural response of craving independently of actual hunger. Stress, negative emotion, and hormonal changes function as internal conditioned cues, reliably activating the ventral striatum and generating the subjective experience of craving [2].
Dopamine and the midbrain feeding circuit. A comprehensive review published in Endocrinology and Metabolism mapped the specific pathways through which metabolic hormones interact with the reward system. Ghrelin directly activates VTA dopamine neurons, increasing dopamine release in the nucleus accumbens and amplifying the motivational drive toward palatable food. Leptin suppresses dopaminergic firing in the same pathway [1].
The menstrual cycle as a craving amplifier. Research published in Nutrients found a significant inverse association between mid-luteal progesterone levels and the intensity of food cravings in the final days of the cycle. As progesterone falls sharply in the premenstrual window, the attenuation of reward circuit activity it provided is withdrawn, and craving intensity increases [3].
Serotonin, the luteal phase, and carbohydrate cravings. Research on neurotransmitter changes across the menstrual cycle has established that reduced serotonin availability in the premenstrual window drives cravings for carbohydrate-dense foods — because carbohydrates drive insulin secretion, which allows tryptophan to cross the blood-brain barrier more readily, temporarily increasing serotonin synthesis [4]. The craving for bread, chocolate, and sweetness in the premenstrual week is, in part, the brain requesting its serotonin precursor.
What actually quiets a craving: protein. A systematic review and meta-analysis of 49 randomised controlled trials published in Physiology & Behavior found that acute protein intake reduces ghrelin significantly while substantially increasing cholecystokinin and GLP-1 — the hormones that most reliably signal satiety and reduce the motivational pull toward further eating [5]. Protein modifies the hormonal environment that speaks to dopaminergic reward circuitry — reducing the volume of the craving signal rather than just overriding it with willpower.
Craving intensity across the menstrual cycle
Schematic based on Hamidović et al. (2023) and Dikshit & Bhatt (2024). Individual patterns vary with chronotype, cycle length, and overall hormonal balance.
The craving for chocolate the week before your period is not a failure of discipline. It is the brain requesting its serotonin precursor through the most direct neurochemical route it has available to it.
The research on cravings carries the usual limitation of most nutrition and neuroscience: the female-specific picture is still being assembled. Much of what we know about dopaminergic reward circuitry comes from studies conducted primarily in men or in rodents. The hormonal dimension of craving — the way estrogen and progesterone directly modulate the reward system's sensitivity — is a newer and still-developing field.
What it tells us, clearly, is this: women's craving experience is not a deficiency of the male default. It is a separate and hormonally structured phenomenon. In the follicular phase, rising estrogen supports serotonin synthesis, dopamine sensitivity, and insulin sensitivity — cravings are generally milder and easier to navigate. In the luteal phase, the neurochemical landscape shifts meaningfully. Reward circuit attenuation decreases. Serotonin availability drops. The craving that arrives in this window is amplified by overlapping neurochemical mechanisms, at the precise moment when the biology makes it hardest to dismiss.
Understanding this does not make the craving disappear. But it changes the relationship to it entirely. You are not fighting yourself. You are navigating a known biological state, with specific features, at a predictable time in the month — and there are evidence-based approaches that work with that state rather than against it.
What protein does to craving hormones — from the meta-analysis of 49 RCTs
Click any stat card or bar to learn more
Data from Kohanmoo et al. (2020), meta-analysis of 49 RCTs. VAS = visual analogue scale (0–100mm). All outcomes vs. isocaloric non-protein comparator. Acute effects.
The goal is not to eliminate cravings. It is to develop enough fluency with your own body's signals to know what is being asked, and to respond in a way that actually meets the need — rather than ignoring it until it grows louder, or silencing it in ways that create other problems.
This is, in part, the serotonin pathway asking for tryptophan. Complex carbohydrates — oats, sweet potato, lentils, whole grain bread — support tryptophan transport and serotonin synthesis more sustainably than refined sugar, which produces a sharper insulin spike and a faster comedown. Meeting this craving with quality carbohydrate rather than suppressing or surrendering entirely to it is one of the more direct ways to address the underlying signal.
This pattern frequently maps to insufficient protein earlier in the day. Ghrelin rises in the absence of adequate satiety signals; dopaminergic reward motivation amplifies in the window before the next significant meal. A protein-anchored breakfast and lunch do not just reduce hunger — they modulate the hormonal environment that produces cravings in the late afternoon and after dinner. The craving at nine o'clock may have been preventable at seven in the morning.
Stress elevates cortisol, which directly upregulates ghrelin and increases the sensitivity of the reward system to food cues. Sleep deprivation produces the same effect: ghrelin rises, leptin falls, and the dopamine system's response to palatable food cues intensifies. This is not emotional eating in the pejorative sense. It is the neural reward system responding predictably to a compromised hormonal state.
This often signals that the meal did not hit the neurochemical targets that would reduce motivational drive toward food. A high-carbohydrate, low-protein meal may satisfy caloric need without meaningfully reducing ghrelin or triggering the CCK and GLP-1 release that quiets the reward circuit. If you eat and the signal persists, the composition of what you ate is worth examining before the willpower narrative takes over.
Track them for two or three cycles. The pattern will clarify. Once you know that your body reliably intensifies its pull toward certain foods in days 22–26 of your cycle, you are working with information rather than against a mystery. You can build that week into your food planning — with intentional protein loading, quality carbohydrates already stocked, and a reduced expectation that willpower alone will feel sufficient. The neuroscience does not support that expectation.
I think about the afternoons I spent in quiet self-recrimination for wanting something sweet. The premenstrual evenings where I read the insistence of a craving as proof that something was fundamentally wrong with me. The years of treating a physiological signal as a character flaw in need of correction.
What the research asks you to consider — gently but clearly — is that the signal was never the problem. The translation was. A craving that arrives in week three of your cycle is not weakness. It is a predictable neurochemical event, driven by progesterone's withdrawal from the reward circuit, reduced serotonin availability, and a dopamine system doing exactly what it was built to do. It is your body, legibly, in its own language, asking for something specific.
You do not have to surrender to every craving, or treat them as commands. But you can stop treating them as enemies. You can learn to read them — to notice when the afternoon signal is about protein distribution, or when the late-luteal pull is about serotonin, or when the evening intensity is about sleep debt caught up in ghrelin. You can respond with something that actually addresses the underlying signal.
The body is not broken. It is articulate, in ways that take some learning to hear. But it is asking something specific. And it has been asking the same thing, in the same ways, all along.
Start with the signal. The translation becomes easier from there. ❤References
- Baik, J.-H. (2021). Dopaminergic control of the feeding circuit. Endocrinology and Metabolism, 36(2), 229–239. https://doi.org/10.3803/EnM.2021.979
- Sun, W., & Kober, H. (2020). Regulating food craving: From mechanisms to interventions. Physiology & Behavior, 222, 112878. https://doi.org/10.1016/j.physbeh.2020.112878
- Hamidovic, A., Soumare, F., Naveed, A., & Davis, J. (2023). Mid-luteal progesterone is inversely associated with premenstrual food cravings. Nutrients, 15(5), 1097. https://doi.org/10.3390/nu15051097
- Dikshit, B., & Bhatt, T. (2024). Unveiling the neurotransmitter symphony: Dynamic shifts in neurotransmitter levels during menstruation. PMID: 39562466. https://pubmed.ncbi.nlm.nih.gov/39562466/
- Kohanmoo, A., Faghih, S., & Akhlaghi, M. (2020). Effect of short- and long-term protein consumption on appetite and appetite-regulating gastrointestinal hormones. Physiology & Behavior, 226, 113123. https://doi.org/10.1016/j.physbeh.2020.113123