Educational article. Not medical advice. We make a sleep formula that does not contain phosphatidylserine — our conflict of interest is spelled out at the end. Everything counted here is from public sources with the queries printed, so you can rerun it.
Ask the internet what to take for the 3 a.m. wake-up and one answer rises above the rest. Not melatonin — people have learned that one knocks you out and then leaves you wide awake at three. Not magnesium — "helped a little." The answer with the most upvotes is a phospholipid most people have never heard of: phosphatidylserine, usually bought as "PS-100." The logic is tidy. The 3 a.m. wake-up is a cortisol spike; phosphatidylserine blunts cortisol; therefore PS-100 before bed. In the most-read thread on the subject this year, the top comment calls it the only supplement that helps 3 a.m. wake-ups specifically, and the person taking it mentions, in passing, that they think 200 mg is the dose that's been studied.
We make a melatonin-free sleep supplement. Phosphatidylserine is not in it — and until this month it wasn't even on our published list of ingredients we considered and rejected. So we owe the internet's favorite 3 a.m. pill the same treatment we've given everything else: two counts, one of the research and one of the shelf.
Count one: the research
On 16 September 2026, PubMed returns four randomized trials for phosphatidylserine AND sleep. Here is what those four actually are:
- A Ginkgo biloba extract combined with phosphatidylserine, tested on mental stress and cognition (PMID 42294065). Not sleep, not PS alone.
- 100 mg of sunflower PS in gummies, given to children aged 8 to 12 for 12 weeks, measuring cognition, mood and sleep — "no differences in the primary or secondary outcomes in the total cohort" (PMID 41318468).
- PS conjugated to omega-3, in children with epilepsy and ADHD, stopped early for lack of participants (PMID 38173190).
- A curcumin trial for migraine, where phosphatidylserine appears only as the delivery vehicle for curcumin (PMID 36698737).
Randomized trials of phosphatidylserine for sleep in adults: zero. Not "weak." Not "small." None.
The cortisol story is more real, and it is worth being precise about, because it is the whole basis of the 3 a.m. theory. Phosphatidylserine AND cortisol AND randomized returns eight results; strip out the ones about amino-acid mixtures, curcumin, PMS and exercise capacity, and the evidence that PS dampens the stress hormone rests on two trials from the same German group — and both tested a complex, not PS alone: in 2004, a soy-derived PS-plus-phosphatidic-acid complex (PAS) blunted the cortisol and ACTH response to a standardized laboratory stress test (PMID 15512856); in 2014, 75 chronically stressed men took a PS-plus-phosphatidic-acid complex for 42 days, and the 400 mg dose normalized their cortisol response to the same test (PMID 25081826). Both trials are daytime. Both measure cortisol under an induced acute stressor — a mock job interview, in effect. Neither measured cortisol at night, sleep continuity, or awakenings. The molecule the internet takes for 3 a.m. has never been tested at 3 a.m.
Count two: the shelf
The NIH label database holds well over a thousand on-market products that list phosphatidylserine somewhere; we took the 1,500 most relevant records and, within them, the 190 products that are phosphatidylserine by name — the "PS-100" bottles, not the multivitamins that sprinkle it in. We read every label.
Then the dose. Of 159 labels where the phosphatidylserine amount can be read directly, the median is 100 mg. Fifty-seven percent sell exactly 100 mg; seventy percent sell 100 mg or less. The dose the cortisol trials used — 400 mg — appears on 6% of labels. So the typical person testing the internet's theory is testing it at a quarter of the only dose with a cortisol result behind it, for a use nobody has trialed at any dose.
Why the internet might still be onto something
Our reading — flagged as interpretation, not counted fact.
The theory is not silly. Cortisol does begin its morning climb in the early hours. A molecule that measurably softens the cortisol response to a stressor is a reasonable thing to be curious about at 3 a.m. The problem is not the idea; it is that the idea has been running on the shelf for years without anyone checking it, at a dose that the only relevant trials didn't use, in a bottle that says "memory" on the front. If phosphatidylserine works for night waking, the world will find out from a trial, not from a comment thread — and the trial would need 400 mg, an evening dose, and a tracker or a lab, not a mood questionnaire in children.
Where we stand
Phosphatidylserine was not among the ingredients we considered for Small Hours, and we are not going to pretend it was. As of today it is on our published "What we left out" list, dated, with this reason: zero sleep trials in adults, and a cortisol effect shown only in the daytime at a dose the shelf rarely sells. That is not a verdict on the molecule. It is a verdict on the evidence — and it is why the ashwagandha we did choose comes at the daily dose used in its published randomized trials (details on the evidence page).
If you know of a randomized trial of phosphatidylserine for sleep maintenance in adults that we missed, write to us. We will update this article and say so at the top. We have done it before.
How we counted, so you can check us
Literature. PubMed, 16 September 2026. phosphatidylserine AND sleep AND randomized — 4 results (PMIDs 42294065, 41318468, 38173190, 36698737), classified by reading each abstract. phosphatidylserine AND cortisol AND randomized — 8 results; the two trials of a PS complex (PS plus phosphatidic acid, PAS) with a cortisol outcome are PMID 15512856 and PMID 25081826. Labels. NIH Dietary Supplement Label Database, on-market records, query phosphatidylserine, the 1,500 most relevant retrieved on 16 September 2026; 190 products with phosphatidylserine in the product name form the label sample. Claim counts come from the label's own statements, matched on the words memory/brain/cognitive/mental/focus, stress, sleep/night/insomnia, and cortisol. Dose is the phosphatidylserine line in the supplement facts panel, per serving, converted to milligrams; 159 readable, 18 "complex" weights excluded, 13 without a readable amount.
Disclosure, as every time: we sell a sleep formula, and it does not contain phosphatidylserine, so a count that finds the evidence thin flatters our formula. Weigh it accordingly. We have tried to make the count reproducible precisely so you don't have to take our word for it.
This article is general education, not medical advice — decisions about supplements and medication belong with you and your doctor.