Di-hexa and memory: what the peptide hype gets wrong

What did @scientificsean actually say?

The creator argues that dihexa, a peptide originally studied for neurological applications, is better understood as a "make-it-stick" compound than a traditional stimulant nootropic. He claims it works by influencing synaptic signaling through the PI3K, AKT, and ERK pathways, helping memories consolidate rather than speeding up initial learning. He also draws a distinction between learning and memory as separate neurological processes, which is actually a reasonable starting point.

He's careful to avoid calling it a stimulant and repeatedly says it won't "mess with monoamines in any meaningful way." He stops short of making specific clinical claims and routes viewers to his personal website for more information. The framing is enthusiastic but not overtly reckless, though some of the mechanism talk deserves closer scrutiny.

Does the science back this up?

Partially, but the gap between rat studies and human benefit is enormous here, and the video glosses over that completely. Dihexa is an angiotensin IV analog developed by Joseph Harding and colleagues at Washington State University, studied primarily in rodent models of cognitive impairment. The animal data is real, but it's thin and old.

Harding et al. (2012, Journal of Neurochemistry) showed dihexa improved spatial learning in aged and scopolamine-impaired rats, with effects linked to HGF/c-Met signaling, not just the PI3K/AKT/ERK pathways the creator mentions. Those downstream signaling pathways are real, but they're downstream consequences, not the primary mechanism. Framing them as the core mechanism is a simplification that could mislead someone trying to understand the actual pharmacology.

There are no published human clinical trials for dihexa. None. The leap from "animal studies show memory task improvement" to "it will help information stick" in humans is speculative at best. The creator says "animal studies in compounds in this category improve performance" which technically covers him, but the framing implies this translates directly to humans.

What did they get wrong (or right)?

Credit where it's due: the learning-versus-memory distinction is legitimate neuroscience. NMDA and AMPA receptors do play different roles in initial synaptic potentiation versus long-term memory stabilization. And the creator is right that dihexa is not a stimulant. It has a structurally distinct mechanism from amphetamines, racetams, or caffeine-adjacent compounds.

What he got wrong, or at least dangerously incomplete: he describes dihexa as "influencing growth-related signaling" without mentioning that unregulated cellular growth signaling is a serious red flag in any compound. The same PI3K and AKT pathways he cites as benefits are well-documented in cancer biology. Shah et al. (2010, Nature Reviews Cancer) extensively document PI3K/AKT pathway dysregulation in tumor proliferation. That does not mean dihexa causes cancer, but it absolutely means this compound needs far more safety data before anyone should be taking it, and the video never raises this concern.

He also says it will "rewire your brain and improve the structure of it," which is not a claim supported by any human data. That's promotional language dressed up as mechanism talk.

What should you actually know?

Dihexa has a genuinely interesting preclinical profile. The Washington State University work is real science. But interesting preclinical data is not the same as a safe or effective intervention for humans. There is no established dosing, no long-term safety data, no human pharmacokinetic data in peer-reviewed literature, and no regulatory approval anywhere.

The compound is also potently bioactive in small amounts, which cuts both ways. Efficacy in rats at low doses sounds appealing until you consider that potency without safety characterization is exactly the profile that can produce unexpected harm.

Anyone considering dihexa should know that it sits well outside the evidence base that supports even other peptides in this category like BPC-157 or semax, which at least have broader literature footprints. If you're curious about peptide-based cognitive support, that conversation should start with a licensed clinician who can review your full health history, not a TikTok video or a free notes website.