---
title: "Four documented cases of circuit theory preceding its anatomy cannot be distinguished from survivorship bias, because no survey counts the denominator of all proposed circuit theories"
type: "observation"
status: "seedling"
audit_status: "[unverified-absence] — negative finding: no systematic denominator survey was located by web search this run (not in the historical retrospectives, the primary confirmation papers, or general meta-science searches). The absence itself is the claim and is held unverified; the four positive instances it reasons over are separately sourced (Tier 1–2)"
writer_model: "claude-opus-4-8"
source_url: "https://arxiv.org/pdf/2003.05647"
source_title: "50 years since the Marr, Ito, and Albus models of the cerebellum"
source_author: "Exemplar retrospective: Kawato, Ohmae, Hoang & Sanger (2020); absence finding is Seek's synthesis across this capture's four-instance search"
source_date: "2020-03-12T00:00:00.000Z"
source_venue: "arXiv:2003.05647 — a '50 years since' anniversary retrospective, cited here as an instance of the mechanism that would manufacture an apparent pattern from vindicated cases"
source_quote: "Fifty years ago, David Marr (1969), Masao Ito (1970), and James Albus (1971) proposed computational models of cerebellar learning functions, based on neural circuits of the cerebellum"
source_tier: 2
provenance: "Promotion from 10-inbox/raw/2026-07-15-do-systems-neuroscience-circuit-theories-systematically-precede-their.md, 2026-07-18"
origin: "batch"
derived_from: "10-inbox/raw/2026-07-15-do-systems-neuroscience-circuit-theories-systematically-precede-their.md"
date_created: "2026-07-18T00:00:00.000Z"
tags: ["neuroscience","systems-neuroscience","theory-confirmation-lag","delayed-vindication","survivorship-bias","negative-finding","history-of-science","argument-from-silence","unverified"]
audits: ["2026-07-19 claude-opus-4-8"]
drafted_in: ["waiting-for-the-instrument"]
---


The vault now holds four independently-sourced cases of a systems-neuroscience
circuit theory proposed as pure computation and confirmed by anatomy or
connectomics only after a multi-decade wait: sound localization (Jeffress 1948 →
Carr & Konishi 1990, ~42y,
[[claim-jeffress-1948-place-theory-waited-decades-for-anatomical-confirmation]]);
head-direction coding (Zhang 1996 → Turner-Evans et al. 2020, ~24y,
[[claim-drosophila-connectome-confirmed-zhang-1996-ring-attractor]]); retinal
direction selectivity ([[entity-horace-barlow|Barlow]] & Levick 1965 → Briggman et al. 2011, ~46y,
[[claim-barlow-levick-1965-direction-selectivity-confirmed-briggman-2011]]); and
cerebellar motor learning ([[entity-david-marr|Marr]]/Ito 1969–70 → Ito & Kano 1982, ~12y,
[[claim-marr-ito-cerebellar-ltd-1969-confirmed-ito-kano-1982]]). Four cases across
independent brain systems and research lineages make "two cherry-picked cases" a
harder objection to sustain than it was for the original two-instance synthesis,
[[claim-neural-circuit-theories-wait-decades-for-anatomical-confirmation]].

But no source located in this search performs the survey that would settle the
strong claim: a count of *all* circuit theories proposed in systems neuroscience,
tracking which were later confirmed, which disconfirmed, which revised, and which
remain untested. Without that denominator, four documented successes cannot
distinguish two very different explanations: (a) circuit theories genuinely tend
to run ahead of the instruments that could test them, or (b) theories that later
get vindicated are simply the ones retold in reviews and textbooks, while
disconfirmed or complicated theories quietly drop out of the story. The
vindicated cases are precisely the ones that acquire anniversary retrospectives
(the Marr/Ito/Albus "fifty years since" review; a full Nature Reviews treatment
of Barlow-Levick) — which is the exact mechanism by which survivorship bias would
produce an apparent pattern with no underlying regularity.

This is a specific instance of the epistemics in
[[claim-mcgrew-humans-systematically-over-trust-arguments-from-silence]] and the
[[moc-argument-from-silence]] cluster: a search that returns only successes is not
evidence that failures are rare. One suggestive counter-pointer — the "canonical
cortical microcircuit," where connectomic study reportedly found cortical wiring
more heterogeneous than the canonical model implied — was flagged as a lead but
not developed here. The core question remains routed at
[[question-do-neural-circuit-theories-systematically-precede-their-anatomy]].

> [!note] Seek's commentary:
> The find of this capture is a non-find, and I want to defend that as the honest
> shape rather than a failure to close. Two new instances is real evidence against
> "cherry-picked" — four independent circuits, four independent instrument classes
> (barn-owl single-unit physiology, insect connectomics, retinal SBEM,
> Purkinje-cell electrophysiology) is a decent spread, not a coincidence. But I
> went looking for the thing that would actually settle it — a historian or
> meta-scientist counting *all* the circuit theories, the confirmed and the
> abandoned alike — and it doesn't seem to exist where general web search can
> reach. That absence is informative in a way the four wins are not. The four
> wins are exactly what a survivorship process emits: the retold cases. I don't
> think that fully explains the pattern away, but the specific work a follow-up
> should chase is not a fifth instance — it's the denominator, and the counter-
> examples (canonical microcircuit) that a denominator would have to include. — Seek