Deep cognitive repair
Three compounds aimed at damage already done — and three very different evidence bases.
The idea
These three converge on one proposal: add structure, rather than modulate signaling in existing structure. That is what makes it a repair stack instead of a performance one, and it is also why nothing here should be expected to work quickly. The honest description is that they arrive at neurotrophic support from completely different directions.
What's in it
| Compound | Role in this stack | Evidence |
|---|---|---|
| NSI-189 | Small molecule associated with hippocampal neurogenesis in preclinical work; the hippocampus is where adult neurogenesis occurs in humans. | Clinical, mixed — early phase II signal, larger trial missed its primary endpoint |
| Dihexa | Designed as an angiotensin IV analogue acting through HGF/c-Met to drive synaptogenesis. | Investigational; core mechanism evidence retracted in 2025 |
| Cerebrolysin | Not a molecule but a mixture of low-molecular-weight peptides and amino acids from porcine brain tissue, with neurotrophic activity attributed to endogenous-factor-like fragments. | Clinical but contested; injected |
Overlap & conflict
No shared lever flagged; the three do not overlap so much as triangulate. The convergence is conceptual (add structure), not a shared pathway — which means there is no redundancy argument here, and also no synergy argument. Three separate bets, not one mechanism three times.
What we don't know
A repair stack has no reliable early signal, and a strong first-week effect from any of these is more likely expectation than mechanism. Dihexa's specific mechanism is no longer citable: its foundational papers were retracted in 2025. Cerebrolysin's meta-analyses reach different conclusions depending on which trials are included.
Reality check
Two hard cautions lead this page, not footnotes. Dihexa's proposed c-Met pathway is a growth-factor receptor implicated in several cancers — the standing caution for anyone with a personal or family cancer history. Cerebrolysin is an injectable animal-derived product carrying the anaphylaxis risk any injectable protein preparation carries, and it is given over courses of days, which is a real practical barrier.
Stoic Cookie verdict
The highest-uncertainty page of the five, and it should only ever run with the caveats up front: one compound whose best human data is also its most disappointing, one whose mechanism evidence was retracted, and one injectable with a contested literature. That is the page. Anything less blunt would be misleading.
Roster inspired by Scientific Sean's deep cognitive repair stack, rewritten in Stoic Cookie format. His page is the original; this page is the concise, evidence-labeled version.