# MIRA export — what the strict file cannot carry

**Paper:** `rozak-2026-neurovascular-dl` · **Claims registered:** 2026-03-30 – 2026-09-10

24 claims, 129 typed relations between them.

## Claims by role

- **empirical** — 8
- **control** — 4
- **scope** — 4
- **hypothesis** — 3
- **methodological** — 2
- **prediction** — 2
- **synthesis** — 1

## Relations dropped

**73 of 129 relations (57%) have no MIRA predicate and are absent from the strict export.**

| Relation | Dropped | What is lost |
|---|---:|---|
| `scopes` | 42 | a scope constraint governs another claim's validity |
| `requires` | 10 | a claim depends on another holding |
| `enables-method` | 9 | a result makes a downstream method possible |
| `derived-from` | 4 | a prediction derived from its hypothesis (inverse of entails) |
| `dissociates-with` | 4 | the source and target jointly establish a dissociation — two claims whose difference across a condition, region, population or measure is itself the finding, neither bearing on the other's truth (symmetric) |
| `entails` | 4 | a hypothesis entails its prediction — the deductive step |

The `entails` / `derived-from` pair is the most consequential: together they are the paper's deductive spine. Without them a reader cannot tell which prediction belongs to which hypothesis.

## Relations flattened

`tests`, `confirms`, `validates`, `extends` and `replicates` all become `mira:supports`; `contradicts`, `rules-out` and `in-tension-with` all become `mira:opposes`. Both directions of collapse lose real distinctions — most sharply, evidence *designed* to test a prediction becomes indistinguishable from evidence that merely agrees with it after the fact. `in-tension-with` under `mira:opposes` says more than the relation means: a tension holds between two claims the paper *asserts*, both of which stand, so `mira:opposes` overstates it as one claim standing against the other. `dissociates-with` is no longer flattened here — it is a neutral contrast with no MIRA predicate and is dropped instead (see above).

## Verification records dropped

**19 verification records across 19 claims are absent from the strict export.** MIRA has no node type for the fact that a claim was independently checked, by what code, against what data, with what result. They are carried in the extended file as `cg:VerificationRecord`.

## Questions synthesized

MIRA requires every Claim to address a `mira:Question`; a claim tree has no such node. **3 questions were derived mechanically from hypothesis text and need human review.**

- Is it the case that vessel caliber measured at a single point along a capillary is an adequate estimate of microvessel volume change, so point measurements suffice for quantifying neurovascular coupling without volumetric segmentation?
- Is it the case that a deep-learning segmentation pipeline (UNet/UNETR ensemble) coupled with cross-time-point registration, vertex-wise radius estimation, and graph-theoretic network analysis can produce reliable, automated, network-level measurements of neurovascular coupling in volumetric two-photon fluorescence microscopy of cerebral microcirculation — measurements that prior point-measurement and manual-segmentation approaches could not yield at the scale of hundreds of interconnected vessels per FOV?
- Is it the case that optogenetic activation of cortical pyramidal neurons elicits coordinated, network-level vascular responses that cannot be predicted from individual-vessel measurements alone: dilations and constrictions are spatially organised relative to active neurons, capillary responses correlate with their network neighbours' responses, and overall capillary network efficiency is modulated by stimulation?

Override any of these by adding a `question:` field to the hypothesis's frontmatter and re-running the export.

## Stance: claims this paper does not assert

**1 of this paper's claims are not asserted by it.** They are alternative explanations it entertains, rejects, or attributes to others. MIRA has no vocabulary for that distinction: it types a node as `Claim` and says nothing about who stands behind it.

The strict export still carries the `rules-out` edge that eliminated each one, declared under `mira:opposes`, so a MIRA-only reader can see the direction of the argument. What that reader cannot see is that the paper **denies** these propositions — so it will over-read them as assertions. The stance travels in the extended file as `cg:stance`.

- `alt-point-measurement-estimates-vessel-volume` — rejects

## Alternatives materialised as claims

None: every relation in this paper points at a claim the tree asserts.

## Paper-level scopes with no MIRA target

None.