A modest, well-tolerated supplement that earns its keep in two narrow windows: older adults whose memory has started to slip, and people who need to flatten a predictable cortisol spike β a hard training session, a high-stakes presentation. Not a daily nootropic for healthy young adults, and not a general energy or mood pill. Honestly told: real but quiet wins for the right person, and the buying decision matters more than most supplements.
Phosphatidylserine β PS β is a fat molecule that sits in the membrane of every cell in your body, concentrated on the inside surface of brain cells where their signals get sent. It makes up roughly an eighth of all the fat in your brain. As you age, that proportion drifts down; cholinergic signalling (the messenger system most tied to memory) drifts down with it, and the supplement was originally pitched as a way to top up the membrane and the signal together.
The honest version of the mechanism is less direct. PS you swallow doesn't reach the brain as PS β your gut breaks it down into lysoPS, glycerophosphoserine, and free serine. Your liver and brain then use those building blocks, plus omega-3 you've eaten and choline from food, to rebuild PS in the membrane where it's actually needed Glade & Smith 2015. So a PS supplement is more accurately a phospholipid-precursor supplement; the brand name happens to be one of the precursors.
For the stress effect, the story is cleaner. PS dampens how loudly the pituitary gland yells "release cortisol" in response to a stressor β an exercise bout, a public-speaking situation, a maths test under time pressure. Resting cortisol doesn't move much; the peak of the stress response gets flatter Monteleone et al. 1992.
What the trials actually show
Two distinct effect classes, two distinct trial literatures, and they use different doses.
The memory evidence is older and best in the bovine-source trials from the early 90s. In 149 adults aged 50β75 with the kind of memory complaint that gets called "age-associated", twelve weeks at 300 mg/day of bovine-cortex PS produced meaningful improvements on name-and-face recall and several other memory tasks β the strongest gains in the people who had the most to gain.
The catch: the cow-brain PS those trials used is not what's on the shelf today. Mad cow concerns in the late 90s pushed manufacturers onto soy-derived PS, and the most carefully done trial of the soy version at the same dose for the same memory problem β Jorissen et al. 2001, 120 elderly with the same complaint β found nothing. The trials that have salvaged the soy story used a special preparation where the soy PS is bound to omega-3 (DHA), reverse-engineering the fatty-acid composition of the bovine version Vakhapova et al. 2010. The Japanese replication at 300 mg/day for six months did show gains in older adults with memory complaints, again concentrated in the lower-baseline subgroup Kato-Kataoka et al. 2010. The FDA looked at all this in 2003 and authorised the carefully hedged claim that PS "may reduce the risk of cognitive dysfunction in the elderly", with a mandatory disclaimer that the evidence is limited FDA 2003.
The cortisol evidence is smaller in sample sizes but cleaner in direction. Across four trials in healthy men under either a cycling stressor or a public-speaking-and-arithmetic battery, PS at 400β800 mg/day flattens the cortisol peak by roughly a third.
Across both literatures, the lower memory dose doesn't reliably move cortisol, and the higher cortisol dose isn't what the cognitive trials measured. Different jobs, different doses.
Three things most bottles get wrong
"Soy PS works like cow-brain PS." The trial data don't back this up. Most product copy on supplement bottles cites the Crook and Cenacchi cow-brain trials β the strongest cognitive evidence in the file β and ships you the cheap plain soy version, which is the one that came up empty in Jorissen et al. 2001. The closest the soy literature comes to matching the bovine results is the version conjugated to DHA (omega-3), and even there the wins are concentrated in subgroups, not headline endpoints.
"PS lowers cortisol." Only the spike, not the baseline. If your cortisol is already humming along at a normal resting level and your day doesn't include a real stressor, PS won't move the needle Starks 2008. The supplement works by flattening the peak when something actually demands a stress response β the hard set, the speech, the cold conversation. For someone whose stress is diffuse and low-grade rather than spiky, it's the wrong tool.
"Take a PS supplement and the PS reaches your brain." No. The PS in the softgel gets broken down in your gut into smaller building blocks (serine and fatty acids), and your brain uses those, plus omega-3 and choline from food, to make its own PS in place Glade 2015. That's still useful β you're feeding the assembly line β but it's why adequate dietary omega-3 matters as much as the PS dose for the cognitive use case.
How to actually take it
Two protocols, one for each use case. Pick the one that matches your goal; don't try to split the difference.
The two protocols are not interchangeable. A reader hoping to cover both with one bottle is overpaying for the memory case or under-dosing the cortisol case, and shouldn't expect either effect to land cleanly.
Otherwise the safety profile is clean. A 30-week extension trial at the memory dose in older adults found no adverse events that differed from placebo Vakhapova et al. 2014. Mild stomach upset at the higher cortisol doses is the only side effect that turns up consistently.
Why people try this and conclude it does nothing
Wrong source for the goal. The most common failure is buying the cheapest soy PS for a memory goal. That's the version the largest dedicated trial said does nothing Jorissen 2001. The version that has the cognitive evidence in the soy era is the DHA-conjugated one, and it costs three to four times as much per milligram. Skipping that step is a near-guaranteed null result.
Wrong dose for the goal. The memory dose (100β300 mg) doesn't reliably blunt cortisol; the cortisol dose (400β800 mg) is expensive overkill if all you want is a daily membrane-support softgel. A reader who takes 100 mg and expects the Trier Stress Test cortisol effect is testing the wrong protocol.
Wrong timescale. The cognitive trials run twelve weeks at minimum to detect modest effects. A reader who tries it for two weeks, feels nothing, and stops has not actually run the experiment the literature is describing. The cortisol effect, by contrast, can be tested in a single session β but the trained-down baseline takes a week of daily dosing to set in.
Stacking that breaks attribution. Most people start PS alongside three other supplements they read about in the same article β omega-3, magnesium, ashwagandha. When something changes or doesn't, attribution collapses; PS gets credit or blame for the wrong reason. If you genuinely want to know whether PS works for you, change one thing at a time.
What it costs and what to buy
The cost depends entirely on which protocol you're running. Plain soy PS at the memory dose runs $60β100 a year β cheap, but, as covered above, probably the wrong product for that goal. The DHA-conjugated version at 300 mg/day runs $200β400 a year. The cortisol-dose protocol at 600β800 mg/day pushes $400β800 a year if you take it daily, considerably less if you treat it as a course around specific high-load periods.
Things worth checking on a label:
- Raw-material source. The two big trial-grade suppliers are Lipogen (Israeli, Sharp-PS brand) and Chemi Nutra (American). If the bottle doesn't tell you, that's a small red flag.
- For the memory case: PS-DHA or Sharp-PS Gold. Plain soy PS at the same dose has the null-trial problem.
- Third-party testing. USP or NSF certification means the bottle contains what the label claims. For competitive athletes, Informed-Sport tested matters β PS isn't banned, but supplement cross-contamination is a real cause of failed tests.
- Soy allergy escape hatch: sunflower-derived PS exists, costs a bit more, and works the same way.
What you should actually notice
Pitched honestly to the right person, the wins are real but quiet.
For the older adult with memory slipping. Three months in on the DHA-conjugated version, the name that wouldn't come back at the dinner party comes back. The bit of the morning routine that used to require a list now runs without one. Your spouse, who has heard you reach for the same name twice in the same conversation more often than they used to, stops reaching for it on your behalf. The effect is modest β you are not the version of yourself you were at forty β but the slope of the slide has flattened a little, and the people who see you most days are the ones who pick that up first Kato-Kataoka 2010, Crook 1991.
For the athlete in a hard training block. Ten days into the cortisol dose, the morning after the hard session is different. The wrung-out feeling that used to colour the whole next day is shorter; your second session of the week lands closer to where you wanted it. You haven't gotten stronger; you've stopped paying the recovery tax twice Starks 2008.
For the high-stakes-speech week. The day of the talk, you take the dose with breakfast. You still get nervous β PS is not a beta-blocker β but the wave that usually arrives ten minutes before you go on stage is smaller. You finish the talk closer to baseline than you usually do, and the evening that always used to be a write-off after that kind of day is not a write-off this time Hellhammer 2004, Benton 2001.
None of this is transformative. PS is not a supplement people quietly become unrecognisable on. It is a useful modulator on two specific dials, in the people whose dials are actually being pushed.
If PS interests you, three adjacent entries are worth knowing about. Omega-3 (EPA/DHA) has stronger trial data for cognitive maintenance than soy PS does on its own β and the DHA-conjugated PS preparation is essentially a combination product, so adequate dietary omega-3 sits underneath any honest version of the memory case. Ashwagandha is the alternative to consider for the stress / cortisol use case if you want a lower baseline rather than a flatter peak. And citicoline (CDP-choline) is the other phospholipid-precursor supplement with cognitive evidence in older adults; it acts on a different part of the same membrane-building pathway and the choice between them is rarely framed honestly.
Substance + claimed effects
Phosphatidylserine (PS) is an anionic phospholipid present in every cell membrane, enriched on the inner leaflet of neuronal membranes where it concentrates at synaptic terminals. As a supplement it is sold in 100β300 mg/day softgels for memory and cognition, and at 400β800 mg/day for cortisol blunting and exercise recovery. Four claims travel with it: (1) slows or partially reverses age-associated cognitive decline; (2) blunts the cortisol response to physical and psychological stress; (3) sharpens focus / attention, including in ADHD-pattern symptoms; (4) reduces perceived exertion, muscle damage, and recovery time after hard exercise. This entry covers all four. Source matters: early trials used bovine-cortex PS (BC-PS), which carries DHA and other long-chain polyunsaturated fatty acids in its sn-2 position; after BSE concerns in the late 1990s, the market switched to soy-derived (S-PS) and later sunflower-derived PS, which carry mostly linoleic and palmitic acid. The headline question hanging over the whole literature is whether soy/sunflower PS reproduces what BC-PS did Glade & Smith 2015.
Evidence by addressing question
mechanism
Membrane biology. PS makes up ~13β15% of brain phospholipid by mass and is held on the cytoplasmic leaflet by flippase enzymes; its negative charge anchors and activates membrane-associated signalling proteins including PKC, Raf-1, and Akt, all relevant to synaptic plasticity. PS also serves as the cofactor for Na+/K+-ATPase activity and modulates acetylcholine release; aged neurons show falling membrane PS content and falling cholinergic output in parallel, the mechanistic basis offered for memory-rescue claims Glade & Smith 2015. Oral PS is digested to lysoPS, glycerophosphoserine, and free serine in the gut; intact PS does not cross the intestinal wall. Reconstitution into neuronal membranes is therefore indirect: serine + DHA + choline pools feed phospholipid biosynthesis at the endoplasmic reticulum.
HPA axis. The cortisol-blunting effect is hypothesised to act via dampened ACTH release from the anterior pituitary in response to CRH, possibly through PS-mediated changes in pituitary membrane signalling or central cholinergic tone Monteleone et al. 1992. The effect is reproducible for physical stressors (cycling at 65β85% VO2max) but the magnitude is set-point-dependent β PS lowers stress-induced cortisol peaks more than resting cortisol.
BC-PS vs S-PS mechanism gap. BC-PS's fatty-acid sidechains are ~50% polyunsaturated, including DHA; S-PS is ~60% linoleic / palmitic. Brain PS is heavily enriched in DHA at sn-2. Whether the cognitive benefits of BC-PS rode on PS-as-a-DHA-delivery-vehicle or PS-as-a-headgroup-effect is unresolved; the soy-PS literature is weaker than bovine-PS on cognitive endpoints, consistent with the former Glade & Smith 2015. The Vakhapova trials deliberately used an S-PS preparation conjugated to omega-3 (PS-DHA), reverse-engineering the BC-PS structure.
evidence
Memory in age-associated cognitive decline (AACD / AAMI) β bovine source. The strongest data are from BC-PS. Crook et al. 1991 randomised 149 adults aged 50β75 with AAMI to 300 mg/day BC-PS or placebo for 12 weeks; the PS group improved on name-face recall and several telephone-based memory tasks, with the largest effect (β baseline rolled back ~12 years) in the lowest-performing subgroup. Cenacchi et al. 1993 randomised 494 elderly Italians with mild-to-moderate cognitive impairment to 300 mg/day BC-PS or placebo for 6 months; the PS arm beat placebo on the Plutchik Geriatric Rating Scale and the Buschke Selective Reminding test β modest but statistically clean. Crook et al. 1992 ran BC-PS at the same dose in 51 mild-Alzheimer patients for 12 weeks; cognitive improvement was significant but small and limited to milder cases.
Memory β soy source. Jorissen et al. 2001 randomised 120 elderly with AAMI to S-PS 300 mg/day, 600 mg/day, or placebo for 12 weeks β null on every cognitive endpoint. This was the trial that punctured the assumption BC-PS data would transfer. Vakhapova et al. 2010 tried a different fix: 300 mg/day S-PS bound to omega-3 (PS-DHA) for 15 weeks in 157 non-demented elderly with memory complaints. Primary endpoint (composite memory score) was null overall but positive in the subgroup with higher baseline cognition. Vakhapova et al. 2014 extended this to 30 weeks (open-label arm) without safety signals. Kato-Kataoka et al. 2010 randomised 78 elderly Japanese with mild cognitive complaints to S-PS at 100, 300, or 0 mg/day for 6 months; the 300-mg arm improved on delayed-recall measures, with the lower-baseline subgroup benefitting most.
FDA qualified health claim (2003). The FDA reviewed the PS-and-cognition literature in response to a petition from Lipogen Ltd. and concluded that the evidence was "limited and not conclusive" but allowed the qualified claim that PS "may reduce the risk of cognitive dysfunction in the elderly" with a mandatory disclaimer. This is the regulatory ceiling: enough signal for a hedged claim, not enough for an unhedged one FDA 2003.
Cortisol / stress. Monteleone et al. 1990 infused 50β75 mg BC-PS IV in 8 men before a cycling stressor; ACTH and cortisol peaks fell ~30%. Monteleone et al. 1992 followed with oral BC-PS at 400 and 800 mg/day for 10 days in 9 men; the 800-mg arm blunted ACTH and cortisol response to the same cycling protocol by ~30%, the 400-mg arm did not. Hellhammer et al. 2004 ran a soy-PS preparation (PAS, 400 mg) on the Trier Social Stress Test in 80 men; cortisol response to the speech / arithmetic battery fell sharply at the 400-mg dose, with self-report distress also lower. Starks et al. 2008 randomised 10 trained men to 600 mg/day S-PS for 10 days; serum cortisol after moderate-intensity cycling fell ~39% relative to placebo, with no change in resting cortisol.
Cortisol-response dose curve. Across the four cortisol trials, the consistent pattern is: 400β800 mg/day blunts stress-induced cortisol by ~30β40%; 100β300 mg/day (the memory dose) does not reliably move cortisol. The cortisol effect requires the higher dose Monteleone et al. 1992, Hellhammer et al. 2004, Starks et al. 2008.
Focus / attention. Direct cognitive-performance trials in healthy young adults are sparse. Hirayama et al. 2014 randomised 36 children aged 4β14 with ADHD-like attentional symptoms to 200 mg/day S-PS or placebo for 2 months; PS improved auditory and short-term auditory memory scores and parent-rated inattention vs placebo. Benton et al. 2001 randomised 54 young adults to S-PS at 300 mg/day for 1 month; under a mental-arithmetic stressor, the PS group reported feeling less stressed and had a smaller heart-rate rise, but cognitive performance per se was unchanged. The focus claim therefore mostly rides on (a) the cortisol-blunting effect translating to perceived clarity under load, and (b) the indirect mechanism via membrane phospholipid maintenance β neither is settled as a direct attention-boost in unstressed healthy adults.
Exercise recovery. Kingsley et al. 2006 randomised 16 active men to 750 mg/day S-PS or placebo for 10 days followed by 60 min intermittent running; PS attenuated post-exercise lipid peroxidation markers but did not change creatine kinase or perceived soreness. JΓ€ger et al. 2007 randomised 20 golfers to 200 mg/day S-PS for 6 weeks; tee-shot accuracy improved (effect size ~0.4 SD), interpreted as a stress / cortisol effect on fine-motor control. Starks et al. 2008 additionally found PS reduced testosterone-to-cortisol ratio collapse after exercise β relevant to recovery and anabolic state.
protocol
Memory dose. 100β300 mg/day, divided across 2β3 doses or taken once with the largest meal (fat aids absorption). Most positive trials ran 12β24 weeks; effects appear gradually, not acutely Crook 1991, Cenacchi 1993, Kato-Kataoka 2010.
Cortisol / pre-stressor dose. 400β800 mg/day, taken 30β60 min before the stressor for acute use, or sustained for 7β10 days for the trained-down baseline response Monteleone 1992, Hellhammer 2004, Starks 2008.
Source choice. BC-PS is no longer commercially available outside historical research stocks. The S-PS supplement market segments into plain S-PS (cheapest), S-PS conjugated to omega-3 / DHA (Vayarin, Sharp-PS Gold β the Vakhapova-trial preparation), and sunflower-PS (for soy-allergic users). The DHA-conjugated preparation is the one with the cognitive evidence; plain S-PS at memory doses has more null data than positive.
contraindications
Anticoagulants. PS has weak antiplatelet activity in vitro at supraphysiological doses; clinical anticoagulation interactions have been theorised but not demonstrated at supplement doses. Conventional caution applies for users on warfarin, DOACs, or dual antiplatelet therapy β flag, don't necessarily contraindicate.
Acetylcholinesterase inhibitors. Memory-clinic patients on donepezil / rivastigmine / galantamine have additive cholinergic-tone effects with PS; coordinate with the prescriber.
Pregnancy / breastfeeding. No safety trials. Default to avoid.
Soy allergy. Plain S-PS carries trace soy protein; sunflower-PS or krill-PS alternatives exist.
Safety profile is otherwise clean. The 30-week Vakhapova extension found no adverse events differing from placebo at 300 mg/day in elderly users Vakhapova et al. 2014. Mild GI upset at the high (600β800 mg/day) doses is the only consistently reported side effect.
misconceptions
"Bovine and soy PS are equivalent." The trial data say otherwise β the cognitive-rescue effect is strongest in the bovine literature and weakens in soy, with the DHA-conjugated soy preparation closest to bridging the gap Jorissen 2001, Glade 2015. Marketing materials routinely cite the Crook and Cenacchi BC-PS trials to sell soy-PS softgels; the substitution is not justified by the trial data.
"PS crosses the blood-brain barrier intact." No. Oral PS is hydrolysed in the gut; the brain gets serine + fatty acids and rebuilds PS in situ. The mechanism for cognitive effects is via substrate supply for brain membrane phospholipid biosynthesis, not membrane PS delivery Glade 2015.
"PS lowers cortisol generally." PS does not reliably lower resting cortisol β it blunts the response to a stressor. A user supplementing PS who isn't under measurable load won't see basal cortisol move Starks 2008.
audience
Older adults with subjective memory complaint. The cognitive evidence is strongest in age-50+ with AACD / AAMI β the population the FDA qualified claim is scoped to. Healthy young adults wanting a nootropic do not have direct trial backing.
Endurance athletes / hard trainers. The cortisol-blunting and recovery evidence is in trained men under structured cycling / running protocols. Translation to recreational lifters is mechanistic-plausible but not directly tested.
High-stress knowledge workers. The Hellhammer TSST trial used a psychological stressor (public-speaking battery) and saw cortisol blunting at 400 mg/day. The intermediate population β someone running a high-cortisol meeting day, not a 90-min cycling session β sits between the two trial populations and gets the indirect read.
Children with attentional difficulties. One small Japanese trial supports a parent-rated benefit at 200 mg/day over 2 months; this entry doesn't recommend pediatric use without clinician input.
alternatives
Adjacent options the reader weighing PS may have on the table:
- For memory / age-related cognition: omega-3 (EPA + DHA) on its own has stronger trial evidence than S-PS, and the DHA-conjugated PS preparation is essentially the combination. Citicoline (CDP-choline) is another phospholipid-precursor supplement with cleaner cognitive data in older adults.
- For cortisol / stress: ashwagandha (KSM-66) has the largest reported cortisol-reduction effect size at supplement scale; rhodiola for perceived fatigue; mindfulness practice for HPA-axis recalibration with the cleanest non-pharmacological evidence. PS distinguishes itself by acting on the stress-response peak rather than baseline.
- For exercise recovery: creatine, omega-3, tart cherry. PS's niche is the cortisol arm of recovery, not muscle-protein synthesis or inflammation.
failure-modes
Wrong source. A user buys cheap plain S-PS, dosing the Crook BC-PS dose, expecting the Crook BC-PS effect. The trial that matters is Jorissen 2001 β null. Buy the DHA-conjugated preparation if memory is the goal.
Wrong dose for the goal. Memory dose (100β300 mg/day) won't blunt cortisol; cortisol dose (400β800 mg/day) is expensive overkill if the only goal is membrane support over years.
Wrong timescale expectation. Memory trials run 12β24 weeks to detect modest effects. A 2-week trial-then-quit cycle catches nothing. Cortisol trials see the effect acutely (single dose, single stressor) but the chronic-blunting protocol runs 7β10 days minimum.
Stacking that breaks the test. PS is rarely the sole change in a stack; users often start it alongside omega-3, magnesium, ashwagandha. Attribution then collapses. If the question is "does PS work for me," isolate it.
practicalities
Cost: plain S-PS at 100 mg/day runs ~$60β100/year; DHA-conjugated PS (Sharp-PS Gold / Vayarin generics) at 300 mg/day runs ~$200β400/year; cortisol-dose protocols at 600β800 mg/day push $400β800/year. Daily compliance is one to two softgels with food. Quality bar: look for Sharp-PS or Lipogen as the raw-material supplier (most trial preparations source from one of these); third-party testing (USP, NSF, Informed-Sport for athletes β PS is not banned but cross-contamination matters).
stakes
The honest stakes are modest. Skipping PS in midlife is not the equivalent of skipping sleep or skipping strength training β this is a supplement with a hedged FDA claim, not a cornerstone intervention. The forecast: someone with mild age-related memory complaint who never tries PS is most likely to have whatever cognitive trajectory their lifestyle and genetics set anyway; the supplement at best shifts that trajectory at the margin. The cortisol-blunting use case is where a real opportunity-cost shows up β a trained athlete or a high-load knowledge worker who is durably elevated on the stress-response peak and never tries the modulator that the trial data say works at their dose.
payoff
Realistic best case β from the trial data β reads as: at 6 months on the DHA-conjugated PS at memory dose, modest improvements on delayed-recall tasks for the older user with baseline complaint, more pronounced in the lower-baseline subgroup Kato-Kataoka 2010; at 7β10 days on the cortisol dose for the athlete, a flatter post-session cortisol peak, less perceived recovery debt the next day Starks 2008; for the public-speaker, a less catastrophic stress response to a known-bad situation Hellhammer 2004. Not transformative on any axis β a useful modulator on the stress and cognition systems with safety on its side and trial backing for narrowly-scoped use cases.
The credibility range
Optimist case. PS has one of the more credible mechanistic stories among popular cognition supplements: it's a real and abundant brain phospholipid, it concentrates at synapses where it gates membrane-coupled signalling, and its content drops with age in parallel with cognitive decline. The bovine-source trials (Crook, Cenacchi) are real RCTs with clinically meaningful endpoints and consistent direction. The cortisol trials are mechanistically plausible (HPA axis is a well-mapped circuit), replicated across labs, and show effect sizes (~30β40% peak blunting) that are large for any non-prescription intervention. The FDA qualified claim is regulatory acknowledgement that the signal exists. Safety at supplement doses across 30-week extension trials is clean. The right read: a substance with a coherent mechanism, multiple positive RCTs in two distinct effect classes, no safety overhang β for an older adult with memory complaint or a high-cortisol athlete, the expected-value calculus favours trying it.
Skeptic case. The strongest trials are old (1991β1993) and used BC-PS, which no longer exists commercially after BSE; whether anything from those trials transfers to soy-PS is contested, with the largest dedicated S-PS-in-AAMI trial (Jorissen 2001, n=120) reporting null on every endpoint. The trials that bridge the gap β Vakhapova, Kato-Kataoka β show effects in subgroups, not primary endpoints, and run on small samples. The cortisol trials are small (n=8 to n=20), industry-funded in several cases (Lipogen, Chemi Nutra raw-material suppliers), and show no effect on resting cortisol β only on the stressor response, which has a substantial placebo-and-expectation component. The FDA qualified claim explicitly says "evidence is limited and not conclusive." The market is saturated with cheap S-PS sold against trial data that doesn't apply to it. Independent meta-analytic synthesis of S-PS in healthy or AAMI cognition is sparse; no Cochrane review exists. Skeptic call: real molecule, real mechanism, real bovine-source signal β but the supplement most users will actually buy isn't what the trials measured, the effects in the trials that match the available product are modest and inconsistent, and the cost is non-trivial.
Author's call. PS is a real-but-modest modulator with two legitimate use cases at two different doses. The cognitive case is best supported by BC-PS data that doesn't quite transfer to the available soy product, with the DHA-conjugated S-PS preparation as the partial bridge; an older adult with subjective memory complaint trying it for 12β24 weeks is making a reasonable bet but not a high-confidence one. The cortisol case is more solid β smaller trials but consistent direction and a coherent mechanism β and is the use case where PS earns its keep: a trained athlete or high-cortisol professional dosing 400β800 mg/day around stressors. evidence: 3 β small trials with a plausible mechanism, worth trying, monitor for updates. controversy: 2 β minor pushback at the margins, mostly around the BC-vs-S source question.
Stakeholder + incentive map
- Raw-material suppliers (Lipogen, Chemi Nutra, Enzymotec/Frutarom). Sell branded S-PS / PS-DHA / Sharp-PS to consumer brands; fund or co-fund most of the post-2000 trials. Trials are real but the publication mix skews toward positive findings; null results (Jorissen) are visible but underweighted in marketing materials.
- Consumer supplement brands. Repackage at 100β300 mg per softgel; routinely cite bovine-source trials in product copy without the source caveat.
- Sports nutrition channel. Pushes the cortisol / recovery use case at the higher dose; this is where the trial-to-product match is cleanest (cortisol trials used S-PS).
- Clinical / academic. No professional society guideline recommends PS for memory or stress. The FDA qualified claim is the highest formal recognition; geriatric and sports-medicine journals treat PS as a plausible but weakly-evidenced option.
- Skeptic / counter-incentive. The nootropic-skeptic community (Examine.com, SBM-adjacent reviewers) treats S-PS for cognition as a low-confidence bet and notes the source-substitution issue.
Population variability
- Age and baseline cognitive status. Effects in the memory trials concentrate in the older / lower-performing subgroups. Healthy under-50s with no baseline complaint do not have a trial-grounded reason to expect cognitive benefit.
- Baseline stress / cortisol load. Cortisol-blunting effects scale with the stressor β the bigger the cortisol peak being blunted, the more PS has to work with. Low-stress users see little movement.
- Training status. Exercise-recovery trials enrolled trained men; sedentary populations have not been studied at the high dose.
- Sex. Cortisol trials are male-only. HPA-axis sex differences (menstrual-cycle modulation, oral-contraceptive effects on cortisol-binding globulin) are not addressed.
- Source sensitivity. Soy-allergic users should pick sunflower- or krill-derived PS. Vegan users have only the S-PS / sunflower options (BC-PS is unavailable; krill-PS isn't).
- Medication interactions. Anticoagulants, AChE inhibitors (see contraindications) are the populations where the consult-clinician bar is real.
Knowledge gaps
What hasn't been settled: (1) whether S-PS at any dose reproduces BC-PS's cognitive effect, with or without DHA conjugation β the question deserves a properly powered (n > 300) RCT and hasn't gotten one; (2) whether the cortisol-blunting effect generalises to women under the same protocols; (3) whether chronic dosing (years, not months) at memory dose produces any longevity, neurodegeneration-prevention, or beauty-cumulative signal β the longest trial is 30 weeks; (4) whether PS has any effect on basal cortisol with very long dosing, or remains response-only; (5) optimal timing of acute dosing relative to stressor (30 vs 60 vs 90 min pre); (6) interaction with daily caffeine, which is itself a cortisol modulator and was unaccounted-for in most trials. Evidence that would change the author's call: a large (n > 500) replication of Crook 1991 using PS-DHA, or a properly-controlled S-PS-vs-PS-DHA head-to-head in AAMI.
Brief vs scope. The input brief named four consequences β memory, stress/cortisol, focus, exercise recovery. The article covers all four, though focus rides indirectly on the memory and cortisol evidence rather than getting its own dedicated trial backing in healthy young adults. No silent narrowing.
The bovine-vs-soy source question. This is the editorial centre of gravity. The strongest cognitive evidence is in cow-brain PS (Crook 1991, Cenacchi 1993), which no longer exists commercially after BSE. The dedicated soy-source replication (Jorissen 2001) was null. The DHA-conjugated soy preparation (Vakhapova, Kato-Kataoka) recovers some signal but only in subgroups. Chose to surface this aggressively in misconceptions and failure-modes rather than burying it β otherwise the article would inherit the supplement industry's source-substitution sleight-of-hand. The cost is a less excited cognitive pitch than other PS write-ups.
Why sleep is 0, not 1. Initially scored 1 on the mechanistic argument (blunted evening cortisol β easier sleep onset for chronically stressed users). On review, no direct trial measures sleep as an endpoint and the chronic-blunting evidence is itself response-only, not baseline. Dropped to 0 to match the dossier's honesty bar rather than score from priors.
Why no dream narrative. Overall score ~13 (well below 40); honest hook is calibration/clarity (right tool, right dose, right product) rather than aspiration or relief. A dream narrative would have rung false at this tier.
Cortisol score sits on mood and health_short_term, not its own dimension. The catalogue has no dedicated stress dimension. Routed the felt stress-resilience effect through mood (peakier wellbeing under load) and the day-after-stress lift through health_short_term. Both 1β2, neither dominant.
Sex bias in the cortisol trials. All cortisol trials are male-only. Did not gender-scope the entry (the memory evidence is mixed-sex and the cortisol effect is mechanistically expected in women too) but flagged it in research population variability. A future replication in women would tighten this.
Future-link candidates. Entries to wire when they exist:
omega-3β underpins the cognitive case, sits next to PS in any honest stackashwagandhaβ the better-evidenced cortisol alternative (lowers baseline rather than flattens peak)citicolineβ the other phospholipid-precursor supplement for cognition in older adultstrier-social-stress-testor a broader HPA axis entry β the underlying stress-physiology model PS acts on
Pediatric-ADHD finding excluded from the body. Hirayama 2014 is one small Japanese trial in children aged 4β14; not a strong enough basis to recommend, and pediatric supplementation deserves its own clinician-mediated entry. Mentioned in research audience; left out of the article.
Cost-burden call. Scored 2, not 1, because the version that actually has cognitive evidence (DHA-conjugated) and the cortisol-dose protocol both push past the trivial-cost band into the $200β800/year range. A reader running the cheap plain-soy version at $60/year is taking the supplement the trials say doesn't work for memory; honest cost-burden has to reflect the effective protocol, not the cheapest one available.
Phosphatidylserine
A softgel or two with food. Easier than remembering most things.
Cheap as a daily memory dose ($60–100/year), pricier as a daily cortisol dose ($400–800/year).
Multiple small trials in two areas, plausible biology, an FDA-allowed hedged claim — not a slam dunk, but real signal.
The clearest cognitive lift is in older adults whose memory has started slipping; for everyone else, the effect runs through stress-blunting rather than a direct attention boost.
Less rattled by the things that rattle you — the meeting, the speech, the session — rather than a general mood lift.
For people who actually face cortisol-spiking situations (a hard workout, a public-speaking week), it takes the edge off how wrung-out you feel after.
Mostly indirect, mostly in athletes — the day after a hard session, recovery debt is a little lighter.