Tuesday, about nine o'clock
He's on the couch with his phone and you are standing in the kitchen doorway holding two mugs, and you know — not suspect, know — that he doesn't love you anymore.
It isn't a thought. It arrived already finished, the way a fact does. You run the evidence while the tea goes cold: he didn't ask about the thing you told him on Sunday, he said "sure" in that flat way, he's been on his phone since eight. None of it is enormous. All of it lines up.
By Thursday the certainty has turned into something with more edges. He loads the dishwasher wrong and you hear yourself say something sharper than the dishwasher deserves, in a voice pitched slightly too high, and there's a half-second where you're watching yourself from the doorway of your own body and you cannot stop.
He goes quiet. That's worse.
Saturday your period starts. By Sunday afternoon the whole thing has lifted like weather, and you look back at Tuesday and it's like reading someone else's diary. You feel a flicker of something — not quite relief, more like vertigo.
And here's the part I want to name, because it is the part almost nobody says out loud: you have noticed this pattern before. You have also explained it away every single time. It really was a bad week. He really was being distant. You're just less patient right now, and anyone would be.
That explanation has held for years. It has the advantage of being partly true, which is what makes it so hard to put down.
What you were actually tracking
Let me be careful here, because the cheap version of this is insulting, and you have probably already been handed it by someone asking whether it's your time of the month in a voice you wanted to slap.
The feelings are not fake. He might genuinely have been distant on Tuesday. You are not hallucinating a marriage.
But the size of the feeling, and the certainty attached to it, and the fact that both drain out of you within a day or two of bleeding — that is not an opinion you arrived at. That is a chemical shift setting the volume, and you have been arguing with the music instead of the dial.
It also isn't who you are now. That distinction matters more than anything else in this post, so hold onto it while I show you the mechanism: a personality is permanent, and a signal is treatable.
The molecule nobody mentioned
After you ovulate, in the second half of your cycle, your ovaries produce progesterone. Your body converts some of it into a calming molecule called allopregnanolone — think of it as your own home-made sedative. It acts on GABA, the brain's main calm-down system, the same system alcohol and anti-anxiety medication act on.
It climbs through the second half of the cycle, then drops off sharply in the last few days before your period.
Here is the finding that reframes the whole thing. In premenstrual dysphoric disorder — PMDD, the severe form of what gets waved off as PMS — hormone levels are usually normal. What differs is the brain's sensitivity to the change. The same rise and fall that another woman barely registers lands on a nervous system that responds to it differently.
Read that again if you have ever had blood drawn and been told everything looked fine. Nothing was going to show up on that test, because the problem was never the level. It was the response to the level. You were not failing to be found. You were being looked for in the wrong place.
PMDD proper affects somewhere around 2 to 5 percent of women depending on how it's measured — and roughly another one in ten has a cycle-linked mood disorder severe enough to need treatment while falling short of the full criteria. That second group almost never gets named, and it is very large.
Then there is what PCOS adds. Women with PCOS have more than double the odds of depression and anxiety, which is why the 2023 international guideline says every woman with PCOS should be screened for both. Higher androgens track with higher depression scores, though that link softens once age and weight are accounted for. And the blood sugar swings I wrote about in the post on crashes and cravings bring their own irritability, on their own schedule, layered over this one.
One honest complication: if you're not ovulating regularly, you may not have a reliable second half of the cycle at all, which makes the pattern harder to see and harder to prove. The research on premenstrual mood flares specifically in PCOS is still thin. That is a gap in the literature, not a gap in you.
The people standing next to it
This part is hard to read, so I'll be quick and I won't moralise.
Partners of women with cyclical mood shifts describe it consistently: not knowing which version will be in the kitchen, monitoring their own tone, going quieter to avoid setting something off. Some stop bringing things up at all.
Sit with that loop for a second. The distance you were certain about on Tuesday is partly real — and partly something the pattern itself manufactured, because he has learned that some weeks it is safer to say less.
You are not a bad person having a bad decade. You are also not required to fix this by apologising more or trying harder, which is what everyone will suggest and which has never once worked, because you cannot out-discipline a signal. The loop breaks when the thing driving it gets treated, not when you get better at white-knuckling the week.
What to actually do
Start tracking, today, before the appointment. Not from memory — memory is exactly what's unreliable here, which is why a PMDD diagnosis formally requires daily ratings across two cycles rather than recall. Rate your mood one to five each day and mark day one of bleeding. Two months turns "I think I might be losing my mind" into a chart, and a chart is very hard for a clinician to wave off.
Ask for the rule-outs too: thyroid, sleep, whether this is depression or anxiety in its own right rather than only cyclical, and whether anything you take — including hormonal birth control — is part of it.
For PMDD, SSRIs are first-line, and they behave unusually here: roughly 60 to 70 percent of women respond versus about 30 percent on placebo, the response comes in days rather than weeks, and because of that they can be taken only during the second half of the cycle rather than every day. Most women have never been told that last part.
Metformin is worth asking about for the PCOS underneath, and I'll hedge it properly. A pilot trial found depression and anxiety scores fell after three months, and metformin plus lifestyle beat lifestyle alone on a depression score — but another randomised trial found no effect at six months, and in one comparison the drug that helped mood more didn't change insulin resistance more. So: a reasonable thing to ask about, not a mood drug. It is off-label for PCOS — approved for type 2 diabetes, used here on guideline recommendation — and it commonly causes stomach upset early, better on the slow-release form, with B12 worth checking long term.
GLP-1 medications I'd raise differently here than in my other posts. The mental-health evidence is genuinely mixed, and some of it points the wrong way: large studies have reported higher rates of anxiety and depression in some people taking them. They are not a treatment for this. If you're already on one for metabolic reasons and your mood got worse, that is worth saying out loud at your next appointment rather than filing under your own failings.
Bring the sentence:
My mood changes follow my cycle — I've tracked it for two months and here it is. I'd like to talk about PMDD, and about whether the PCOS underneath is making it worse.
Two doors, and then it's yours
First door: your PCP, OB-GYN, or endocrinologist. Bring the chart. Use the word PMDD, because naming it changes what gets considered.
Second thing to know before you go: this gets missed constantly. PCOS affects roughly one in ten women and is still routinely answered with a birth control prescription and nothing else, and cyclical mood symptoms get written off as stress or personality more often than they get worked up. If your appointment goes that way, you've learned something about that clinician's training, not about whether this is real. Ask again, or ask someone else.
Second door: telehealth built for this. Allara Health is a virtual practice for PCOS and women's hormonal and metabolic health that pairs clinicians with dietitians and works with many major insurance plans, though availability varies by state. Sesame is a direct-pay marketplace where you book a named clinician at a listed price without insurance or a referral, useful if you're uninsured or facing a long wait.
I want to be clear that I have no financial relationship with either of these. I'm not an affiliate and I don't get paid if you book. I mention them because "find a doctor who actually knows PCOS" is useless advice without somewhere to start.
PCOS is managed, not cured, and cyclical mood symptoms don't vanish on a schedule anyone can promise you. But a week that reliably arrives and reliably leaves is the most treatable shape a problem can have, and there are women whose Tuesdays stopped feeling like that. You were never someone who became difficult in her thirties. You were someone carrying a signal nobody measured — and it is measurable, starting with two months and a piece of paper.
This is my experience and my reading of the research, not medical advice. Nothing here should replace a conversation with a clinician who knows your history.
Sources
- Role of allopregnanolone-mediated GABA-A receptor sensitivity in the pathogenesis of premenstrual dysphoric disorder — Journal of Zhejiang University Science B, 2023
- Premenstrual Disorders — StatPearls, NIH National Library of Medicine
- Mental Health Across the Menstrual Cycle in Polycystic Ovary Syndrome: Insights and Implications — Phimphasone-Brady et al., Current Psychiatry Reports, 2024
- Toward the Reliable Diagnosis of DSM-5 Premenstrual Dysphoric Disorder: The Carolina Premenstrual Assessment Scoring System — Eisenlohr-Moul et al., American Journal of Psychiatry, 2017
- Expert Guidelines for the Treatment of PMDD — MGH Center for Women's Mental Health
- Recommendations from the 2023 International Evidence-based Guideline for the Assessment and Management of PCOS — Teede et al., Journal of Clinical Endocrinology and Metabolism, 2023