Phosphatidylserine (PS): FDA Claim vs. Discontinued Trial Ingredient

By Erin Rose · Updated July 18, 2026 · Methodology

Educational overview — not medical advice. Phosphatidylserine (PS) is not a treatment for dementia, Alzheimer's disease, or any diagnosed condition. This statement has not been evaluated by the FDA. This product is not intended to diagnose, treat, cure, or prevent any disease.

As an Amazon Associate we earn from qualifying purchases. On best PS, products are ranked by disclosed source and branded-vs-generic raw material first, cost per mg second — not by commissions.

Start here

What phosphatidylserine is, and the source-shift that changes everything

Phosphatidylserine (PS) is a phospholipid concentrated in cell membranes, especially in brain tissue, where it's involved in membrane signaling and cell-to-cell communication. It is not an essential nutrient — there's no deficiency state, no RDA, and no government-set upper limit — so any claimed benefit of supplementing has to stand on its own trial evidence, not a "correcting a shortfall" story. The single most important thing to understand before reading any PS marketing claim is that the ingredient sold today is not the ingredient most of the classic evidence was built on. The trials that underpin both the FDA's qualified claim and most PS marketing copy — Crook 1991 (n=149, age-associated memory impairment) and Crook 1992 (n=51, probable Alzheimer's disease) — both used bovine-cortex PS (BC-PS) at 100 mg three times daily (300 mg/day), and both found improvement vs. placebo on cognitive test scores. BC-PS was discontinued industry-wide in the 1990s over BSE (mad-cow disease) contamination risk. It has not been sold in decades. Every capsule on shelves today is derived from soy lecithin or, in a smaller number of newer products, sunflower lecithin — and EFSA's expert panel has formally concluded that bovine-cortex- and soy-based phosphatidylserine "are different substances" that "might...have different biological activities," and that a cause-and-effect relationship between soy PS and cognition "cannot be established."

Does it work? (the honest short version)

Read the modern evidence on its own, and it's thin. The one trial built as a direct soy-PS replication of the Crook design — same 300 mg/day dose, same age-associated-memory-impairment population — is Jorissen 2001 (n=120, 12 weeks plus a 3-week washout), and it found no significant differences between treatment groups on any outcome, including the delayed recall and recognition primary endpoints; the authors concluded soy-PS "does not affect memory or other cognitive functions." A combination soy-PS-plus-omega-3 product (Vakhapova 2010, n=157 randomized) did find a significant improvement in immediate verbal recall, but the effect was concentrated in a subgroup with higher baseline cognitive performance, not the full sample, and the combination design means PS's independent contribution can't be isolated. The newest evidence on the actual sunflower-derived ingredient sold today — Friling 2025, testing Sharp-PS® Green at 100 mg/day in children — was null on its primary and secondary outcomes across the full cohort, with a benefit only in a pre-defined below-median-baseline subgroup; four co-authors are employees of IFF Health, which commercializes that ingredient. Nothing sold today has cleanly replicated the old bovine result. See the full trial-by-trial breakdown on the memory page.

Beyond memory: cortisol, exercise, and ADHD (all small, all separate threads)

PS has been tested outside the cognition-dose lane too, and none of these threads should be merged into one unified story. Four small trials looked at cortisol and exercise stress: Starks 2008 (n=10, 600 mg/day; two of its authors are sports-supplement-industry researchers) and Monteleone 1992 (n=9, oral 800 mg/day) both found blunted cortisol/ACTH response to exercise, without a measured performance outcome; Kingsley 2006 (n=14, 750 mg/day) found an exercise-capacity benefit but no cortisol effect — a genuinely inconsistent picture across trials, not one unified "PS lowers cortisol and improves performance" claim. A third Monteleone trial (Monteleone 1990) used PS intravenously, not orally, so it's mechanistic evidence only, never oral-dose evidence. Separately, three pediatric ADHD trials show a "preliminary evidence, low-quality" pattern: Hirayama 2014 and Manor 2012 found symptom improvements, but a 2021 meta-analysis (Bruton 2021, 3 RCTs/n=216 pooled) found overall ADHD symptoms and hyperactivity-impulsivity were not statistically significant in the pooled analysis — the essential honest counterweight, and this is pediatric-specific evidence, never a general "focus" claim for adults.

The buying problem: source and dose, not purity fraud

Unlike some categories on this site, PS doesn't have a documented industry-wide label-fraud problem — the six products tracked here all disclose actual PS mg on the label. The real buying problem is different: source and dose. Most retail servings are 100 mg, meaning three capsules are needed to reach the classic 300 mg cognition-trial dose, and one product in this comparison labels itself "soy-free" without disclosing what the alternative source actually is — a real, honest example of the disclosure gap this cluster exists to flag, not an accusation. See best PS for the full ranking and the dosage guide for the cognition-vs-cortisol dose gap.

Frequently asked questions

Related guides

• Omega-3 — often combined with PS in the one positive soy-PS memory trial (Vakhapova 2010)
• NMN — the site's other regulatory-honesty structural twin, with its own resolved FDA whiplash story
• CoQ10 — another mitochondrial/cellular-membrane-adjacent supplement with an absorption-form distinction
• Magnesium — the site's most-built cluster, for a comparison of how form and dose evidence get separated honestly

Sources