Why anxiety spikes in the week before your period

There is a specific and disorientating experience that a lot of people have every month: for a few days, everything feels more threatening. The unanswered message means something. The conversation you had on Tuesday is now evidence. Your chest is tight for no reason you can name.

And then your period starts and it lifts, often within a day, which is somehow the strangest part — because if it were really about your job or your relationship, it would not resolve on schedule.

The molecule in the middle

After ovulation, progesterone rises. Some of it is converted in the brain to allopregnanolone, a neurosteroid that binds to GABA-A receptors — the brain's main inhibitory system, and the same receptor family targeted by benzodiazepines and alcohol.

For most people, allopregnanolone is calming, in the way a small drink is calming.

For a subset, the effect is paradoxical: the same molecule produces anxiety, irritability and agitation. This paradoxical response is well documented and appears to relate to how an individual's GABA-A receptors adapt to fluctuating neurosteroid concentrations. Some brains adjust smoothly. Some do not, and the adjustment period itself is where the symptoms live.

This matters because it explains a fact that otherwise makes no sense: people with severe premenstrual anxiety usually have entirely normal hormone levels. The difference is not the dose. It is the response.

The withdrawal is as important as the peak

Allopregnanolone rises through the luteal phase and then falls sharply in the last few days before bleeding, as the corpus luteum breaks down.

For many people the worst days are those final two or three — a withdrawal effect, structurally similar to what happens when any GABA-active substance is removed. That is consistent with the timing most people report: symptoms peaking immediately before the period, then resolving with startling speed once bleeding starts. The luteal phase covers the wider set of changes happening at the same time.

Everything else is running in the same direction

The neurosteroid story is not the whole picture. Several other things stack up in the same window.

Serotonin activity falls in the late luteal phase, which is the leading explanation for why SSRIs work for premenstrual mood symptoms — and work unusually fast, often within a day or two rather than the weeks required in depression.

Sleep is worse. Progesterone raises core body temperature, which makes falling asleep harder and sleep more fragmented. Poor sleep independently amplifies anxiety, so the two compound. Sleep in the luteal phase covers what helps.

Alcohol hits differently. Drinking to take the edge off a premenstrual evening reliably produces worse anxiety the next day, and the effect appears larger in the luteal phase — you are adding a GABA-active substance to a system already destabilised by one. Alcohol, sleep and next-day anxiety explains the rebound.

Blood sugar swings more, and hunger is higher, both of which affect mood in people already primed for it.

Where the line sits

Feeling more anxious for a few days a month is common. What separates it from something with a diagnosis and a treatment pathway is functional impact and remission.

PMDD requires marked symptoms in the final week before bleeding, significant interference with work or relationships, and — critically — absence of symptoms in the week after the period. If anxiety is present all month and simply gets worse premenstrually, that is premenstrual exacerbation of an underlying anxiety disorder, which is common and needs a different treatment approach: continuous rather than cycle-timed. PMS and PMDD are not the same thing sets out the criteria.

Getting that distinction right is the single most consequential thing in this area, and it cannot be done from memory.

What actually helps

SSRIs, either continuously or dosed only during the luteal phase. The luteal-only option is genuinely effective for PMDD and is under-offered.

Ovulation suppression with continuous combined hormonal contraception — no hormone-free week, no cyclical fluctuation. It helps some people substantially and makes others worse; birth control and mood is worth reading first.

Protecting sleep in the second half of the cycle, deliberately, as a scheduled thing rather than a reaction.

Cutting alcohol in the premenstrual week specifically, which many people find is the single highest-yield change they make.

CBT, which has evidence here, particularly for the interpersonal damage that premenstrual irritability causes.

Exercise helps, though the evidence is more modest than it is often presented as. Most supplements marketed for this have thin evidence; calcium has some.

Two cycles of daily records changes the appointment

Because the diagnosis rests on timing and remission rather than any test, prospective daily tracking across two cycles is not a formality — it is the diagnostic instrument. And the days you rate as fine matter as much as the bad ones, because remission is the criterion.

There is also a personal benefit that arrives before any doctor sees it. Knowing on Tuesday that you are on day 25 and that this is what day 25 does reframes the experience entirely. It does not make the anxiety smaller. It makes it explicable, which is enough to stop you acting on it.

Naked is built to make that record accumulate without effort — daily mood and anxiety logged against your cycle, sleep and everything else in your life, so the shape of it becomes something you can see rather than something you suspect and then forget by the following month.