# Felisyn early stop 023: protocol 023 ended NOT INTERPRETABLE at its first instance

The result package of protocol 023, the confirmatory plaids protocol, dated 8 November 2026. The protocol was to run six fresh network instances of the 2 mm cat circuit model and ask whether the plaids pilot's departure from a relay holds on average over them. It ran one. A declared integrity check refused that run's export, the declared early stop ended the queue, and the declared analysis wrote the outcome **NOT INTERPRETABLE**. No confirmatory reading was made.

- **Declared** at 2026-11-07T20:25:30Z, before any 023 run. Spec SHA-256 `1e20a4a90a474d2a2f4348f095133a820120f0a4e85dc1095e6b32ec00cffac3`, published in [journal note 23](https://www.felisyn.com/journal/the-model-moves-to-a-laptop) before any run. The design is published as [protocol-023-public](https://www.felisyn.com/research/protocol-023-public/README.md), with the [spec](https://www.felisyn.com/research/protocol-023-public/frozen/spec.md) (six home folder paths redacted; that README says how to check it against the published hash).
- **Frozen** at 2026-11-08T00:46:40Z by a manifest with SHA-256 `6020fe1b7aaed781f40b90eadd7bb35da4f0de89c7f516ab7b4234a7f8ba1429` (73 files).
- **Run** on the 8 GB laptop that replaced the lost research server: the queue started at 00:47:21 UTC on 8 November, and its first run, s7, ended at 01:53:28 UTC.
- **Stopped** at 01:53:34 UTC by the declared early stop, after the server check refused s7's export. The local check refused it too at 01:55:30 UTC.
- **Analysed once,** on the frozen contract: `results/results.json` was written at 01:55:52 UTC (its file time). Its SHA-256, `7836cdb06942563f02e2bc42e909801cae058c8451fddc48f1d7d057d8879bc0`, was recorded in the project's handoff before any other reading of 023 data.

Simulation only. This is a 2 mm reduction of a published model of cat primary visual cortex, not a biological measurement and not biological validation. It is not a controller test. No token, wallet, exchange, key or real order was involved.

## The declared outcome

The analysis wrote this text (`results/results.json`, field `text`), fixed in the spec before any run:

> Protocol 023 cannot be read as declared: the early stop: the server check refused the export of pynn_seed 7 right after its run, for a fault taken to be shared by every instance, and the local check refused it too (the LGN lead-in identity: X_OFF 1799 of 1800 trains identical). No confirmatory reading is made. The data and the failure are kept; a rerun needs a new declared revision.

Declared next step (field `what_follows`): "A rerun needs a new declared revision."

The spec asks that one limit be printed beside any outcome text of this protocol, because its other outcome texts call the instances ones that "had never been built": the server half of the seed search ran with round 3's script on 3 November 2026 at 00:29:38Z, the final script ran only on the Mac, and the server's files can no longer be read.

## What the protocol asked

In the spec's words: "Over network instances of the same 2 mm reduction of LSV1M that have never been built, does the intact circuit, on average, combine two superimposed drifting gratings differently from a relay, in the direction instance A's pilot found? It is asked for each of the three measures that triggered the pilot's GO (XOS, Dc and ds), with the same definitions and the same margin of 0.10." Six slots (pynn_seed 7 to 12) and two reserves (13 and 14), each with its own declared stimulus order; a one-sided t-test of the six instance estimates per measure, with Holm's correction over the three at a family-wise 0.025. The pilot itself is [plaids-022](https://www.felisyn.com/research/plaids-022/README.md) and [journal note 22](https://www.felisyn.com/journal/two-gratings-at-once).

## What ran

| Item | Value |
| --- | --- |
| Run | `m23-s7-intact-a1`, pynn_seed 7, the first slot, attempt 1 |
| Command | the pilot's intact command with the three 2 mm overrides and `pynn_seed 7`, through the memory wrapper, under a 3.7 GiB memory cap with a 3 GiB disk floor (`receipts/queue/s7-command.txt`) |
| Start and end (UTC) | 00:47:21 to 01:53:28 on 8 November, 3,965 s |
| Exit | 0 |
| Memory | peak footprint 3.186 GiB at 3,814 s; peak resident 2.112 GiB; swap 1,092 MiB at the start, no growth |
| Network | 14,712,684 connections; NEST seed 42, 4 threads |
| Export | `exports/s7.json.gz`, 24,493,818 bytes, SHA-256 `736c2b3ca2daf4e55ab1ed8ca7bbb2582f12f2844879702e718ecee05c8841a2`, the same on the laptop and after the fetch (`receipts/data-sha256.txt`) |
| Run folder files | the same SHA-256 at the start and the end of the run (`receipts/queue/s7-sha256-start.txt`, `s7-sha256-end.txt`) |
| Machine | Apple M2, 8 GB, macOS 26.2 build 25C56, on mains power, no thermal warning (`receipts/queue/launch-environment-20261108T004720Z.txt`, `s7-therm.txt`) |

No other seed ran. Seeds 8 to 14 were never built.

## The checks

The blind checker reads only the 30 % single gratings, the LGN lead-in and the export's own records, and computes no plaid quantity. It ran twice on the same export with the same frozen code: the server check, right after the run (since the move to the laptop, "server" means the machine that ran the model), and the local check, after the fetch. Their check values are identical.

| Check | Role | s7 |
| --- | --- | --- |
| The export is the declared run (run name, arguments, resolved seeds and weights, kernel, steps) | refusal | passes |
| LGN band: X_ON and X_OFF counts at 30 %, within 0.90 to 1.10 of the pilot's | refusal | passes: X_ON 7.959 (0.987 of 8.06), X_OFF 5.772 (1.007 of 5.73) |
| **LGN lead-in identity: every LGN spike train in the 504 ms lead-in equals the pilot's reference** | **refusal** | **refused: X_ON 1,800 of 1,800, X_OFF 1,799 of 1,800 (cell 2738 differs)** |
| Layer 4 band: 30 % own singles of the 45° and 135° groups, within 0.80 to 1.25 of the pilot's | reported only, never a refusal | 45°: 1.085; 135°: 1.282, outside the band |
| Layer 4 group sizes (45°, 135°, 90°) | reported only | 921, 1,117, 774 cells |

The layer 4 values describe a run whose LGN input was faulty in one cell, and every cortical train of s7 differs from a repeat of its inputs over the sequence, the first at 75.2 ms (`diagnosis/`, exploratory). The spec gives the layer 4 band no decision role, and it has none here.

## Why the early stop fired

The spec: "Every refusal kind above is one that a fault shared by every instance would produce (the queue's command, the engine's records, the input stage, the engine itself), so the first refusal of the server check is taken as systematic: the protocol is NOT INTERPRETABLE with that refusal as its condition, once the local check of the same export, after the fetch, has refused it too."

So the queue stopped at once with exit 5 (`receipts/queue/early-stop.txt`, `queue.log`, `queue.out`), no further attempt or reserve ran, the local check refused the same cell (`checks/check-s7.json`), the summary recorded the stop (`checks/instances.json`), and the analysis wrote NOT INTERPRETABLE. The spec named the cost of this rule in advance, as its risk 14: "A fault of that one instance (an export damaged as it was written, say) would then end the protocol where the replacement rule could have replaced it; the rule accepts that cost for the server time it saves when the fault is shared."

## Exploratory diagnosis, after the outcome

`diagnosis/diagnosis-public.md` summarises an exploratory diagnosis made after `results.json` was written. It changes nothing in the outcome, and every claim in it is labelled measured, inferred or untested. In short:
- **measured:** two exact repeats of s7's inputs, d1 (5 steps) and d5 (full length), passed the lead-in check, and a third, rate-r1, equals d5 in all 18,600 spike trains. Against d5, s7 differs in one LGN cell only, 2738, which fired 47 extra spikes until its train became identical again at 14,344.6 ms;
- **inferred:** s7's difference was not a fixed result of its declared inputs, and the lead-in check caught a real execution fault on the laptop's run path. The early-stop rule then ended the protocol as declared, a cost the design accepted in its risk 14 (above). How often such a fault occurs on this run path is close to unknown (below);
- **untested:** the cause. The best fit is that cell 2738's luminance input read as zero for its first 2,048 input values, the size of one 16 KiB memory page, and three places where that could happen fit equally well;
- **measured:** full-length runs on the laptop checked spike by spike: 1 fault in 4 (G4, s7, d5, rate-r1). **inferred:** with four runs the rate is close to unknown, a 95 % interval (Clopper-Pearson) of 0.6 % to 81 %;
- **a procedural gap:** the spec asks that any exploratory reading of 023 data have its own manifest, frozen before it reads any 023 export. The diagnosis had none. Its records say it computed no plaid quantity.

The options for a new declared revision are listed there without a recommendation. The owner decides.

## What went wrong on the way

- **The first freeze stopped at the disk floor,** on 7 November at 21:15 UTC, before any run. Our own mistake: the workspace's full verification script ran beside the freeze, and one of its tests copies a 294 MB folder 18 times. Free disk fell to 0.18 GiB. No manifest was written. The second freeze passed at 00:46 UTC on 8 November.
- **The first tier 2 call was refused** by its disk guard at 20:25:42 UTC on 7 November (2.98 GiB free) and made nothing. The second passed at 20:37:37 UTC.
- **The local check was first written under a name the summary does not read** (`s7-check.json`) and then renamed `check-s7.json`. The summary made in between is kept as `checks/instances.pending-wrong-name-20261108.json`; it records the instance as pending. The summary made after the rename is `checks/instances.json`.
- **A diagnosis run, rate-r2, stopped at the disk floor,** because another program on the laptop filled the disk. It is not a model result.

## Limits

- **One instance ran.** Nothing here measures the plaids effect in a fresh instance. The pilot's GO is unchanged and still rests on one network instance.
- **The stated limits of the laptop run,** which the spec asks every result of this protocol to repeat:
  1. **A new platform.** The pilot ran on x86-64 Linux; this run ran on an arm64 Mac with Apple clang, libc++, macOS's maths library and Python 3.9.6, on one process.
  2. **Two engine patches,** the only changes to pinned sources: the LGN noise pair order and a NEST block size ([laptop-move](https://www.felisyn.com/research/laptop-move/README.md)).
  3. **A memory wrapper** around the unchanged runner, shown byte-identical on a short run.
  4. **The second-machine check is lost.** The server check and the local check both ran on the laptop, under two Pythons, so a fault of this machine would show in both.
  5. **The server's half of the seed search** (the limit quoted above).
  6. **What the laptop's identity with the pilot rests on:** the tiny regression and G4, one instance on one sequence.
  7. **One shared laptop,** on a nearly full disk.
  8. **What G4 shows:** one instance on one sequence reproduced exactly, not that a fresh seed behaves.
- **A spec sentence found wrong for this laptop, by the exploratory diagnosis after the outcome.** The spec says of heat and other work that "the simulation is deterministic, so they change only the time". One full-length run in four differed from an exact repeat of itself (measured); with four runs, a 95 % interval for the rate runs from 0.6 % to 81 % (inferred).

## Redactions

Every file here is a byte-for-byte copy of its source in the project's private workspace except those listed below. They differ only by exact replacements, each token absent from the original, so putting the original strings back gives the original files; the builder checked that for every one. `manifest.json` gives each original's SHA-256 and the lines.

| File | Lines | What was replaced | SHA-256 of the original | SHA-256 here |
| --- | --- | --- | --- | --- |
| `receipts/queue/launch-environment-20261108T004720Z.txt` | 4; 15; 39, 51, 66 | the laptop's host name, replaced by `redacted-host`; the laptop's battery identifier, replaced by `id=redacted`; 3 copies of the home folder's path, replaced by `<home>` | `9c6fcb039c2ee579be2e9200234fee3989dcad4dba8df1d2b2aa34d2c8ffd361` | `b9310b7c03e69e5f0858c2ad107a037c43b5ea3758b50acf8f420fcb2dc7ea86` |
| `receipts/queue/queue.log` | 1 | the home folder's path, replaced by `<home>` | `b96ffad59c8bab6309ecdb73eede190ad8a832b13f20615bfd99c95f052a8c9b` | `08e14897a88bd31fac31f2b959fdd60b16b7db7cbdc2b3f6bec39312b691b36b` |
| `receipts/queue/s7-env.txt` | 1, 4, 11 | 4 copies of the home folder's path, replaced by `<home>` | `0fcf58bd6604b7c869817e76dcfa1c5aa7ab5c4be22f3cc88180c6e39ea8df5d` | `947df5be96bef763408f57533603b4b8b0dbc117ca46826ca322444adf109251` |
| `receipts/queue/s7-memory-end.txt` | 17 | the home folder's path, replaced by `<home>` | `5d2dba0ff658d345130770c6a85d2c2c4aeef45a92f99a3ab692f7889e23d238` | `532f71af4410db01d3decdabceeef87872f17e7a47e6e4f49d47a9ba23f86486` |
| `receipts/queue/s7-server-check.json` | 184 | the home folder's path, replaced by `<home>` | `27fc061f78e7dfd00f083c9c21a5adc5a373837979d53b38fdce74026b02a250` | `7aaf8050e9f516708286af7e242f6ca4ccbb37244f18663c22e6eec0fc0cf7be` |
| `receipts/queue/s7-server-check.mem.json` | 7, 9, 11, 12 | 4 copies of the home folder's path, replaced by `<home>` | `3115a2237900c0a638a3b54ebea4b94fa5a77041719afe91b507309dbcd50801` | `16d18e02ce54034e8b3f54cf11783e9b9a81c95ea1857281cd317fb17482345c` |
| `receipts/queue/s7-sha256-end.txt` | 553 lines, 1 to 591 | 553 copies of the home folder's path, replaced by `<home>` | `f1f6bbbac3be4249fc5cd59f5e33b8250c43e0d31ab4dbe9da54ba0de790be4b` | `f731f47914fbe524a516f11181afe875ca3bd20195c73d648711c3e7960c5c86` |
| `receipts/queue/s7-sha256-start.txt` | 553 lines, 1 to 591 | 553 copies of the home folder's path, replaced by `<home>` | `f1f6bbbac3be4249fc5cd59f5e33b8250c43e0d31ab4dbe9da54ba0de790be4b` | `f731f47914fbe524a516f11181afe875ca3bd20195c73d648711c3e7960c5c86` |
| `receipts/queue/s7.log.gz` | 30, 7640, 7666 | 3 copies of the home folder's path, replaced by `<home>` | `b24323742b5ac97646f28d40ed859ecc1334f7255dd99849dc01471e4a5671ff` | `0bf4642d76521cd7239c532c0c7ed12285fe12ab5cf5e3a4c56091453c272639` (gzipped) |
| `receipts/queue/s7.mem.json` | 10 | the home folder's path, replaced by `<home>` | `5b4f42e593c2caa19b17501371eb2dd25b240ba7f64105c738d5615dbe2f93ab` | `3857838d1222322533a85e20e23e7cd2d4c35d0dd6443bf8c5e45113773911c4` |
| `checks/check-s7.json` | 195 | the home folder's path, replaced by `<home>` | `8676b17b7b437d34a0d8c7f724f0d301a93a9b240d46d2bdae44f03f10b0fadb` | `b5a8065ec9f3e1df7922fc942febc59d402e540b81bb8c7996d4fd8e83def2c5` |
| `checks/instances.json` | 139 | the home folder's path, replaced by `<home>` | `340bee05f2e4263b74c0dd6bf4967c52a8cd68e8953c2f8f801994e583c0e841` | `602d51120981acea942f7b6f2a22f7782245be91f7db9971842b7f86a8c3b143` |
| `checks/instances.pending-wrong-name-20261108.json` | 120 | the home folder's path, replaced by `<home>` | `a46a6eebfb05ac940d1d009a6238fc34271e4c8b15a4c80561ad839a5defe975` | `ae51dae84af90f0057375f086bc4565d18d89dbccd810de55a085227ce4cb5c1` |
| `results/results.json` | 360 | the home folder's path, replaced by `<home>` | `7836cdb06942563f02e2bc42e909801cae058c8451fddc48f1d7d057d8879bc0` | `ebf180e850f19fe517f53ebe9e56283b6de9ab1f6cca5fb6e3cac66b8b60e2b9` |
| `diagnosis/runs/d1-s7first5/d1-s7first5.mem.json` | 10 | the home folder's path, replaced by `<home>` | `8c1ede8e98ddce4093df244fc92e13ec7537d59d6a82680e34e9ad44cff39f23` | `633a34e3a3b3e0bc9d74ee724be5f56082308e047fd8f62a4438b84ada9a4c5c` |
| `diagnosis/runs/d1-s7first5/input-sha256.txt` | 1, 2, 3, 4, 5, 6 | 6 copies of the home folder's path, replaced by `<home>` | `c959daffadebddc34399f11969c33c65f222843f4bd297fde91d665f9991674e` | `6a6458a19f6a028f220086c8845816fe34206563cdac2f764916788eb5447213` |
| `diagnosis/runs/d1-s7first5/memory-end.txt` | 14 | the home folder's path, replaced by `<home>` | `d9f628dc3c35c012e97cc62d67245a26b79d63f992865f5c99d402b59941b219` | `ba6a6ad0fb01ff63e718f4d4aaa240b9014efd20468989337b7800eda1ba1f71` |
| `diagnosis/runs/d1-s7first5/leadin-check.json` | 2 | the home folder's path, replaced by `<home>` | `677b43b34ce43921bb9e8f24709ee7e5d1fa42d4ef1c1604d9579c30567e0892` | `70039c9c7de31cb939935de0214eba955dc84a59548934861b51ee5c994a2b5f` |
| `diagnosis/runs/d1-s7first5/vs-s7.json` | 3, 5 | 2 copies of the home folder's path, replaced by `<home>` | `cc537f83f185ad70470148dd9a24a1a801afa54e376e6f3db5873a0ef97cc0b6` | `e4a46c37a1df049368519897c906782bf7e25e6895919e68bd270da3bfabcbfc` |
| `diagnosis/runs/d1-s7first5/vs-d5.json` | 3, 5 | 2 copies of the home folder's path, replaced by `<home>` | `7aabb9f6831074bc8c7ed194073cf6206d3f596b14045adc9d176389566624f0` | `9722cabef9d123d276279780b01a34a9f060fe6914574c76195305f6d6327c88` |
| `diagnosis/runs/d5-s7repeat/d5-s7repeat.mem.json` | 10 | the home folder's path, replaced by `<home>` | `478f756b772dba8e23d3e78d69f72509abae83cb4b320c6f84f5620862971ca1` | `15d86dcd9f8d03c823c1c2cec39179e3acdd7b402b07493125ae21a6a91d9cb6` |
| `diagnosis/runs/d5-s7repeat/input-sha256.txt` | 1, 2, 3, 4, 5, 6 | 6 copies of the home folder's path, replaced by `<home>` | `5fd73b82368258b3875c1ba42746f378d9a1328fe863f89fd8d4314f54122e14` | `b7c998950c5a27e8075ebc474c4e0c51d0099cbb4832249d609af6cca872f7a3` |
| `diagnosis/runs/d5-s7repeat/memory-end.txt` | 14 | the home folder's path, replaced by `<home>` | `e654bba9e62916272cbc70cbfecb271e293802ecae8f37fcbfb65a1415404903` | `de33ae3b3d3b3901f14b700d67179201a692c23100c1904e01986db03925aef8` |
| `diagnosis/runs/d5-s7repeat/leadin-check.json` | 2 | the home folder's path, replaced by `<home>` | `d61053653c6882f13840318e8f3447cb3abfadfc1dd4856b580f7c81b2337bfc` | `9ff96159aee4fd207d38557c634badd7c2efd236936ecd36439d3959274fdcec` |
| `diagnosis/runs/d5-s7repeat/vs-s7.json` | 3, 5 | 2 copies of the home folder's path, replaced by `<home>` | `ed70befdf570681ecef2f18ad84f794c9fdb955c32337204c5a0d4a42828ce51` | `f77118458bc0db22d60a40533b2127152f56b2db923ced5d6ff5962884b11ae2` |
| `diagnosis/runs/rate-r1/rate-r1.mem.json` | 10 | the home folder's path, replaced by `<home>` | `55ee71d4db562362577d24914b3593dff2e37a3a2e83b680bd1956545f1ebee8` | `7d808b445fe9ef8d774804b8663058d159a64b03a1d082ae117dca397a25f715` |
| `diagnosis/runs/rate-r1/input-sha256.txt` | 1, 2, 3, 4, 5, 6 | 6 copies of the home folder's path, replaced by `<home>` | `90a8e6f912abccec6dbdfd80769a9ec197ee93e744ce5e1616c8a1adbc6bb75f` | `4b761c11d52b470f10746924d7ab325fc24594eaa8fe1c32b29c53f541bd36f3` |
| `diagnosis/runs/rate-r1/memory-end.txt` | 14 | the home folder's path, replaced by `<home>` | `1a34ba4b3a8d3fa15fc0393a30c308f90f8281cc32233537e2033595c5ff4480` | `30fc449aa68af10b66189e9b8a7bbf134d3808d354c036d6f404da580fcb6aae` |
| `diagnosis/runs/rate-r1/vs-d5.json` | 3, 5 | 2 copies of the home folder's path, replaced by `<home>` | `c464bae601aac2fc6fa8a0258230121f64b2719a0c49c127ec6dc20c1051dbf3` | `fc7d2eb78bcaec392ae7da7ae2be325bf501cfddc3cdb22478706675f242a0c5` |
| `diagnosis/runs/rate-r1/vs-d5.txt` | 10 | 2 copies of the home folder's path, replaced by `<home>` | `5cb5f945a78654dd8383cc31de6b3a269a5065bf5247189aafaff1ad597dc131` | `78b514dbac4c2f73474daba38eab4345fac70f48787dea27a53f4efd9f5cdb14` |
| `diagnosis/runs/rate-r2/rate-r2.mem.json` | 10 | the home folder's path, replaced by `<home>` | `baf5aea37d8b70cb0f4343574dd1f6707a04df9c4ace0d8edff99d53a9f9078a` | `0b94374133d4073605edc0135263c4845b7e41e96d9d6a197d6dbb9d8c3a7cc0` |
| `diagnosis/runs/rate-r2/input-sha256.txt` | 1, 2, 3, 4, 5, 6 | 6 copies of the home folder's path, replaced by `<home>` | `216dab5d36e37cac5448526e384f52383e4aad9acb3af30a4935e62b62c319f3` | `5bbf8e8bc51e40813e4a1dbe9c856755bbbd8606b7dbb38f157bbc18146488d1` |
| `diagnosis/runs/rate-r2/memory-end.txt` | 14 | the home folder's path, replaced by `<home>` | `7ba4bfbacb95701fde096c352d19bd35f06277741fb3e16afd2da393dd514aef` | `4da7a9d6c51eb15c595f11858710feeea3370f598023350a495b6333aa3637b0` |

`receipts/queue/s7.log` is published gzipped as `s7.log.gz`, after its replacements; decompressed, it reverses to the original. The run's environment variables in `s7-env.txt` are published with paths redacted; the diagnosis runs' environment files, which record the laptop's full search path, are not published.

## What is here

| Path | What it is |
| --- | --- |
| `results/results.json` | The declared analysis's output: outcome, fixed text, conditions and the frozen check (redacted: one path). Original SHA-256 `7836cdb06942563f02e2bc42e909801cae058c8451fddc48f1d7d057d8879bc0`. |
| `checks/check-s7.json` | The local check record of s7 (refused). |
| `checks/instances.json` | The summary of the instances, from the queue's receipts and both checks: stopped by the early check. |
| `checks/instances.pending-wrong-name-20261108.json` | The first summary, made while the local check had the wrong name (kept). |
| `receipts/queue/` | The queue's 27 receipts as fetched: the launch record and environment, the queue log and exit, s7's command, environment, times, exit, memory, heat, run-folder hashes and log, the server check record (`s7-server-check.json`, SHA-256 `27fc061f…`), its log and memory, and the early-stop record. |
| `receipts/data-sha256.txt` | The fetch's record of the export's size and SHA-256 on the laptop. |
| `exports/s7.json.gz` | s7's export, byte for byte: spike trains of all 18,600 neurons, neuron tables and run records. |
| `diagnosis/diagnosis-public.md` | The public summary of the exploratory diagnosis. |
| `diagnosis/runs/` | Receipts, lead-in checks and spike train comparisons of the diagnosis runs d1, d5, rate-r1 and rate-r2. |
| `diagnosis/recount/`, `diagnosis/tools/` | The recount of cell 2738's trains with its script, and the comparison script. |
| `figures/` | Figure 1 of journal note 24, its data, its two scripts and its caption. |

The diagnosis runs' exports (about 24 MB each) and the rest of the diagnosis stay in the project's private workspace.

## Credits

- **Model.** Antolík J, Cagnol R, Rózsa T, Monier C, Frégnac Y, Davison AP (2024), *A comprehensive data-driven model of cat primary visual cortex*, [PLOS Computational Biology 20(8): e1012342](https://doi.org/10.1371/journal.pcbi.1012342), [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/). Its code, LSV1M v1.0 (MIT), ran with the 2 mm reduction's overrides; none of its files is copied here.
- **Software.** Mozaik (CeCILL), PyNN 0.11.0 (CeCILL), NEST 3.4 (GPL) with the NEST step current module (GPL) and two patches offered under the terms of the files they change, GSL 2.7.1 (GPL).
- **This work.** The protocol, the runs, the checks, the diagnosis, the figure and its scripts are Felisyn's own.
