At research-chem doses (40-80 mg/day oral): highly variable. The modal positive-responder report is a slow, cumulative mood-bright effect over 2-6 weeks, with gradual improvements in verbal memory access, subjective clarity, and reduced social anxiety. Not acute — there is no "feel it on day 1" experience like modafinil or bromantane. Biohackers who report success usually say they noticed improvement around week 2-4 of consistent dosing.

The non-responder + adverse-responder population reports headache (most common), nausea, fatigue, and sometimes a paradoxical brain-fog + low-mood effect — the opposite of the intended outcome. A nontrivial fraction of users discontinue within 1-2 weeks for tolerability reasons.

Honest framing: Even at the dose that worked best in the Phase 1B trial (40 mg/day), the magnitude of subjective effect in healthy biohackers is generally described as "subtle" — not in the modafinil class, not in the racetam class. People who keep taking it usually justify it on theoretical neurogenic grounds (hippocampus expansion long-term) more than on what they feel day-to-day.

• Tolerance buildup: Unknown / minimal in clinical data. Trials ran 6-8 weeks with steady efficacy.
• Recommended cycle (biohacker convention): 4-12 weeks on, 2-4 weeks off. No strong scientific basis for any specific cycle, but the convention is conservative given unknown long-term effects.
• Reset protocol: Not formally established. If discontinuing for any side effect, no taper is needed (long half-life will self-taper over ~5 days).

Synergistic with (theoretical)

• cerebrolysin: Both target the neurotrophic / neurogenic axis. Cerebrolysin delivers BDNF/NGF/GDNF mimetic activity at TrkA/TrkB receptors; NSI-189 indirectly upregulates BDNF expression. Mechanistically additive on paper. No clinical co-administration data. If Dylan ever ran NSI-189, it would most logically sit during a Cerebrolysin off-cycle, not stacked simultaneously, to keep the variables separable.
• semax / n-acetyl-semax-amidate: Both upregulate BDNF; NASA's longer-acting BDNF curve plus NSI-189's neurogenic action are mechanistically complementary. Again, no human evidence.
• bpc-157: Broad neuroprotective + neurogenic signal in animal data; theoretical multiplier on whatever NSI-189 is doing at the niche level.
• citicoline / alpha-GPC: Cholinergic substrate for new neurons forming new synapses. V4 stack already has citicoline.
• omega-3 DHA: Membrane substrate for new neurons. V4 stack already has 2 g DHA — no need to add.

Avoid stacking with

• dihexa: Both promote brain plasticity via different mechanisms (NSI-189 → progenitor proliferation; dihexa → HGF/c-Met synaptogenesis). Stacking compounds two unproven pro-proliferative compounds with no human safety data — bad risk profile. Dihexa's theoretical cancer concern is the bigger one but NSI-189 adds noise to any signal you'd want to track.
• High-dose racetams + cholinergics during NSI-189 onset — confounds whether benefit/AE is from NSI-189 or from the stack.

Neutral / safe co-administration

• Standard V4 stack (DHA, magnesium, citicoline, NAC, PS, curcumin, rhodiola, theanine, D3+K2, Mg-glycinate, beta-alanine, vitamin C). No known interactions.
• Modafinil — different mechanism, no documented interaction.
• Caffeine — fine, no interaction.

• CYP profile: Not well-characterized. Limited oxidative + glucuronide metabolism in unpublished in-vitro data. No clean public CYP induction/inhibition data. Take this as a warning, not an all-clear — if you're on any CYP-sensitive medication (warfarin, hormonal contraceptives, antiretrovirals, antiepileptics), assume potential interaction and don't combine without monitoring.
• Monoamine drugs: No direct receptor/transporter binding, so theoretically minimal interaction with SSRIs/SNRIs/MAOIs. However, in the moderate-MDD subgroup signal from Phase 2 (2020), the benefit appeared adjunctive to standard antidepressants — not replacing them.
• GABAergics: Theoretical concern given the GABA-A negative allosteric modulator signal — could theoretically reduce efficacy of benzodiazepines or sleep aids. No clinical confirmation.

• Minimal data. No published PGx work on NSI-189 specifically.
• CYP genotype: Without clean CYP characterization, no specific PM/UM guidance possible.
• BDNF Val66Met: Theoretical relevance — if NSI-189 works partly through BDNF signaling, Met/Met carriers (~25-30% of European ancestry, reduced activity-dependent BDNF release) might respond differently. Pure speculation; no data.
• 5-HTTLPR: Probably irrelevant given the non-monoaminergic mechanism.
• For Dylan: 23andMe results land June 2026; if he ever pursued NSI-189, BDNF Val66Met genotype would be the one variant worth checking, though the inference would still be weak.

COA verification is critical. The community split between "amazing, life-changing" and "terrible, made me sick" reports correlates suspiciously with vendor identity in some Longecity threads. Without identity + purity confirmation, you don't know what you're taking.

Identity confirmation if paranoid: Mass spec is the gold standard. For practical biohackers, sticking to vendors with HPLC ≥98% + UV trace + a published method is the realistic ceiling.

• Baseline (before starting):
  • Verbal memory (Rey AVLT or list-recall web tool — Alto used the latter as their biomarker)
  • Working memory (digit span, N-back)
  • Subjective mood (PHQ-9, MADRS if depressive)
  • Sleep architecture (Oura/Whoop trend)
  • Optional: hippocampal MRI volume (impractical for most)
• During use (every 2-4 weeks):
  • Repeat verbal memory test
  • PHQ-9
  • AE log: headache, nausea, sleep, anxiety, dreams
• Post-cycle:
  • Same battery, compare to baseline
  • If verbal memory improved, that's the closest thing to a real biomarker signal. If it didn't, the cycle did nothing measurable.