Кстати спросил нейронку, как это учёный с мировым именем Элиэзер мог проебать очевидный элемент изучения магии:

The "magic comes from somewhere else" dodge is baked in early
Around Chapters 25–26 (genetics discussion) and later revelations, Harry concludes that magic isn't generated by the wizard's biology or brain in a mundane way. Instead, there's an external "Source of Magic" (or Atlantean artifact/computer/program/whatever) that responds to specific triggers: genetic marker + intent + ritual (words/wand movements). The caster's mind is just the authentication/keypress mechanism, not the engine.
Once Harry accepts this model, probing his own brain becomes low-priority — it's like asking "what neurons fire when I type my password?" instead of "how does the server actually compute the hash?" The interesting physics is supposedly upstream/outside the caster, so he focuses on reverse-engineering the trigger protocol rather than the wetware running it.

Narrative convenience and plot momentum take over
The story starts with "science will conquer magic" hype but quickly pivots to high-stakes plotting: Time-Turner shenanigans, Azkaban breakout, Voldemort intrigue, Horcrux 2.0, the "something to protect" arc, etc. Deep, slow lab-work on caster biology would grind the pace to a halt.
Eliezer has said in passing (author's notes, old LessWrong posts, interviews like the 2020 "We Want MoR" podcast) that he wrote HPMOR somewhat seat-of-the-pants style — arcs evolved as he went, and the early "systematic magic research" promise got deprioritized when bigger philosophical/transhumanist themes emerged. Exploring brain scans would require inventing a ton of new wizarding tech (or smuggling Muggle gear), new characters (healers/neuro-wizards), and weeks/months of in-story time — none of which fit the accelerating thriller tone after ~Chapter 30.

It would break the epistemology the story wants to sell
If Harry did rigorous caster-neuroscience, he'd almost certainly discover that magic is deeply tied to consciousness/intent/qualia in ways that resist clean reduction to physics. This would force the story to confront hard problems like: Is magic qualia-dependent? (→ panpsychism vibes, which Eliezer dislikes)
Can you fake intent with Occlumency/brain-hacks? (→ breaks a lot of security assumptions)
Does Dementor exposure/ Patronus creation involve specific neural modules?
These would push toward mysticism or unsolved philosophy-of-mind territory, clashing with the "rationality triumphs via reductionism" branding. It's safer (for the thesis) to keep the caster side hand-wavy and focus on clever visualization/intent tricks that feel like software hacks.

Harry's curiosity is selective and author-directed
Multiple critics (su3su2u1, various Reddit threads) note that "Rational!Harry" is often incurious where it would derail the plot or contradict the author's worldview. He does tons of object-focused tests (transfiguration limits, spell misfires) but almost none on himself as the variable. Real scientists start with the measurables closest to hand — i.e., strap yourself (or a volunteer) into the wizarding fMRI equivalent and cast Lumos 100 times while monitoring. Harry never proposes it because the author doesn't want to go there. It's performative science: the experiments shown are the ones that yield cool, narratively useful results (partial transfiguration, True Patronus) without forcing uncomfortable metaphysics.

Author background and priorities
Eliezer is a philosopher/AI safety thinker first, not a neuroscientist or experimental physicist. His rationality pedagogy emphasizes cognitive biases, Bayesian updating, reductionism as a mindset, and "thinking really hard" over grinding through lab protocols. HPMOR reflects that: Harry's "science" is mostly armchair hypothesis-generation + occasional targeted tests, not systematic measurement programs. Brain/wand mechanics would require hardcore empirical neuroscience, which isn't Eliezer's wheelhouse or interest — so it gets omitted.