The reference layer: everything the tracks draw on. Consult it, rather than working through it.
This site has two layers, and confusing them is the fastest way to get lost in it.
Modes cut across all three — you can walk a track alone, in a hosted session, or (eventually) inside a research programme. The mode axis is independent of the topic axis: any track can be worked in any available mode.
So: what you are learning is the track, how you are using it is the mode, and what it is built out of is this page.
Long-form reference documents at instructor depth: neuroanatomy, ultrastructure, imaging, infrastructure, cell types, connectomics, proofreading and case studies. Each entry runs a full argument with worked examples, misconceptions, and real references, rather than summarising one.
Reach for it when: You need the depth behind a claim a unit made in one sentence.
Every term carries a definition, typical values with units, why the term matters, and what it is commonly confused with. The confusion field is the useful one: most vocabulary errors in this field are near-miss substitutions, not blanks.
Reach for it when: A word stops you. Look it up immediately rather than reading past it.
The literature organised by topic, each entry stating what the paper established, what it did not establish, and the specific thing worth arguing with. Discussion prompts are written for a group but work alone.
Reach for it when: You want to judge a claim, or you are looking for the paper behind a method.
The lookup layer: spatial and temporal scales side by side, canonical dataset dimensions, the acronym set, and the numbers that recur across the curriculum, with units attached and sources named.
Reach for it when: You need a number, a scale comparison, or to decode an acronym mid-paper.
The operating knowledge research training assumes you already have: technical practice, reading and judging, lab norms, career mechanics, meta-learning, belonging, and handling conflict. Written down because it is otherwise transmitted only by proximity to someone who already knows, which distributes it unequally by background.
Reach for it when: You are unsure what is expected of you, and asking would cost something.
What public volumes exist, what each can and cannot answer, how to get access, and the acquisition-to-interpretation workflow they sit inside. Includes the version-pinning practice that every downstream analysis depends on.
Reach for it when: You are choosing data, citing data, or trying to open some.
The failure mode this layer exists to prevent is reading a reference document front to back because it appeared in a list. Here is what each surface is actually for.
| You are… | Go to | Not to |
|---|---|---|
| Stopped by a word | Dictionary | A unit. A large share of apparent difficulty in this field is vocabulary, and it is fixable in about a week. |
| Trying to do a specific technical thing | Content library | The module that mentions it. The library entry is the depth; the module is the session. |
| Checking a number, a scale, or an acronym | Atlas and reference | Searching the units. The atlas exists so you do not have to. |
| Assessing whether a claim in a paper holds | Journal club | The abstract. The entries state what each paper actually established and what it did not. |
| Unsure what is expected of you, socially or professionally | Hidden curriculum | Guessing, or asking someone in a way that costs you. |
| Choosing or citing data | Datasets | A paper’s methods section alone. You need the version, not just the citation. |
Two things on this site look like core material and are not.
Modules and technical units are path content. They have an order, they build on each other, and each ends in an artifact. They belong to tracks. Where a module and a unit cover the same ground, the module is the session and the unit is the depth behind it.
Session kits, slide decks and worksheets are delivery material. They are generated from the module pages, and they exist for whoever is running the session rather than for whoever is learning. If you are studying alone, you can ignore them entirely. See teaching.
Some material is deliberately off the main paths: coherent, worth doing, and not a prerequisite for anything. The largest is proofreading — around 17,000 words of reference material plus a unit, five worked scenarios, and a portfolio artifact, treated as an allocation problem under a fixed budget rather than as a stage in a sequence.
Correcting reconstruction errors as an allocation problem under a fixed budget. Six stages over the reference material, ending in a portfolio artifact a lab can read and disagree with.
What the category is, why material sits outside the tracks, and what else is named but not yet built.