Instructor Notes — Chapter 23

Nootropic and Neuroprotective Peptides — Semax, Selank, Cerebrolysin, and the Cognitive Claims Part 4 · intermediate · prerequisites: Chapters 5, 6, 22

What this chapter is actually teaching

Do not teach this as a compounds chapter. The compounds are the occasion.

The subject is how to evaluate evidence you cannot fully assess — and the chapter is built so that the two obvious moves are both wrong. Dismissing an unfamiliar literature is parochial and is a rule-4 violation (never downgrade with distaste). Accepting it because it exists abandons the standard applied everywhere else in the book. Students will want one of those two exits. The entire pedagogical task is to keep them in the uncomfortable middle long enough to discover that the middle has a procedure, not just a feeling.

The secondary subject — §23.7 and §23.8 — is the most transferable material in Part IV, and many students will rate it the most useful content in the book. Budget time accordingly.

Time budget (three 75-minute sessions, or one long seminar)

If you have only one session, teach §23.3 and §23.7 and assign the rest. Those two sections carry the chapter.

The three places students reliably go wrong

1. They resolve the ambiguity. Students find "⚠️, access-limited" unbearable and convert it, usually within a week, into either "so it probably doesn't work" or "so there's probably something there." Watch for this in written work. The tell is that the type label disappears while the symbol survives.

Intervention: make them write the two triggers (what clears the access flag / what changes the rating) every single time. A rating with two live falsifying conditions is much harder to quietly collapse into a prior.

2. They mistake skepticism for rigor. A confident negative verdict on a literature the student has not read is the same error as a confident positive one, and it usually goes ungraded because it sounds careful. It is not.

Intervention: the mechanical test from §23.3 — can you name the study-quality problem, or are you gesturing at a country? Apply it out loud, in class, to a student's own answer, with their permission. It lands far harder than saying it abstractly.

3. They agree about practice effects and then are impressed by a before/after result anyway. This is nearly universal and it is why case study 23.2 exists.

Intervention: before assigning the case study, put the clinic's study on a slide and take a vote: is this evidence the product works? Record the vote. Work the six-term decomposition. Vote again. The gap between the two votes is the lesson, and it is more durable than any explanation you can give.

Handling the chapter's political temperature

Some cohorts will bring strong priors — in either direction — about research from Russia, and a few students will try to make the session about geopolitics. The chapter is written to make that conversation unnecessary, and you should hold that line firmly and without drama.

Useful moves:

• Redirect to the mechanism of the claim. "What specific study-quality question are you asking?" If there isn't one, there's nothing to discuss yet.
• Deploy the symmetry immediately. The oxytocin literature (Ch 21) is English-language, well-funded, indexed, and had a decade of confident findings that did not replicate. Publication bias, industry funding effects, and selective outcome reporting are documented features of the most accessible literatures in the world. This is not diplomacy; it is accurate, and it defuses the framing faster than a rule announcement.
• Name the constraint explicitly if needed. "The argument in this chapter is about evidentiary standards and accessibility. It is not about anyone's competence, and the chapter does not make that argument because that argument is not supported."
• Do not let students name individuals. The chapter deliberately names no Russian investigator or institution. Keep it structural.

If a student argues that the chapter is too generous to the Russian literature, the honest answer is: the chapter makes no claim about that literature's quality at all, which is the point, and the student should say what specific finding would settle it. If a student argues the chapter is too stingy, the answer is the same sentence.

Two demonstrations worth doing live

Demonstration A — the search that changes the sentence (10 minutes). Project ICTRP or ClinicalTrials.gov. Search one of the chapter's compounds. Then search a second database. Have students write, before and after, the sentence they would use to describe the evidence. The sentence almost always changes from a claim about the world to a claim about a search, which is exactly the §23.3 lesson arriving on its own.

Demonstration B — the multiplicity arithmetic (5 minutes). Put eleven columns on the board. Ask how many you would expect to "move" by chance under conventional thresholds. Then write the two sentences side by side: "significant improvements in verbal fluency and processing speed" versus "the pre-specified primary endpoint was delayed recall; here is what happened to it." Ask which study you would rather read. Nobody chooses the first one once it is on the board, and yet the first one is what most of them have been reading for years.

Assessment guidance

• Grade on reasoning shape, not conclusion. See _scratch/answers/ch23.md. A student who concludes ⚠️ and one who concludes ❌ can both be right; what separates them is whether they attached the rating to a claim, separated assessed from unassessable, named a specific problem, and stated a falsifying condition.
• The high-value exercises are D, O, P, R, T, X, Y, AG, AJ, and AL. If you assign a subset, assign those.
• Item O is a personal-honesty exercise. Grade on the specificity of the replacement question, not on the confession. Never read one aloud.
• Penalize invented numbers. In this chapter specifically, a student who produces a percentage they cannot source has committed the error the chapter is about. Reward explicitly labeled illustrative reasoning ("if the practice effect is typical, I'd expect…") and mark unlabeled fabrication down hard, with the reason stated.