The real cost in a lab workflow is not any single analysis. It is the handover — a student finishes, and what remains is a folder of scripts with names like final_v3_actually_final.py, a spreadsheet whose column order nobody remembers, and a figure that cannot be regenerated.
Structuring the work as a study rather than a folder is what fixes that. Recordings, every processing stage with its parameters, the trained models, the results and the report live in one project, so what a colleague inherits is the work rather than the archaeology.
The study as the unit of work
A study holds the imported recordings, the named stages each step produced, the ML pipelines that trained against them, the results, and the report. Because stages are named and non-destructive, opening someone else's study shows you what was done in order, not a final state you have to reverse-engineer.
Data Manager handles the import side: single file or whole-folder multi-subject import, modality selection, label column confirmation, and quality warnings surfaced before anything is added to the study.
Working as a lab, not as individuals
Team studies
Shared studies sync per section, so two people working on different parts of the same study are not overwriting each other.
Co-authors
Invited at publication and required to accept, so authorship is recorded rather than assumed.
Roles
Plan-level roles separate who can run, who can publish and who can administer, which matters once a study leaves one person's laptop.
Review
Community review is a structured process with a claimed decision, accumulated comments, a recorded outcome and a quorum rule.
Reports and the results notebook
Reports come in two forms: a structured Fields editor that builds the report from the study's own results, and a document editor for prose that surrounds them. Numbers in a report come from the run rather than being typed in, so a report cannot quietly disagree with the study it describes.
The results notebook runs Python in-process against the finished study for anything the report does not cover — a bespoke figure, a statistical test the results page does not implement — without exporting to a separate environment first.
Handover, forking and reuse
A published study can be forked as the starting point for new work, which is the mechanism that turns a graduating student's project into the next student's baseline instead of into a folder nobody opens.
Datasets and models are separate published objects with their own pages, so a recording set can be reused by a study that shares none of its analysis, and a trained model can be evaluated against data it never saw.
Where the data actually lives
Recordings and intermediate stages stay in the project folder on the machine that ran them. Nothing is uploaded as a condition of using the software, and publication is an explicit action with a choice about what goes.
For labs whose ethics approval forbids sharing recordings, that is the difference between a usable workflow and a non-starter.
Questions
Can two people work on the same study?
Yes. Team studies sync per section, so simultaneous work on different parts of a study does not overwrite. Team features sit on the paid plans.
What happens when a student leaves?
The study remains a complete, openable record: recordings, named stages with their parameters, models, results and report. It can also be forked as the basis for the next project rather than starting over.
Does the whole lab need paid accounts?
No. The free plan covers two studies with the full analysis suite and 118 of the 119 ML blocks, which is enough for individual work. Team, clinical and enterprise features are what the paid plans add.
Can I keep everything private?
Yes. Publishing is a separate, deliberate action, and a study can be run start to finish without ever leaving your machine.
Read the reference
Related
Reproducible BCI Research
A methods section is a summary written from memory weeks after the run. A study that carries its own stages, parameters and fold scores is the run itself.
BCI Research Platform
BCILattice runs on your machine and covers import through model training. BCINexus is where the finished study is published, reviewed, cited and reused.
BCI Analysis Software
Eleven steps sit between a conditioned recording and a labelled design matrix. In BCILattice each one shows you its result before you commit it, and records what it ran with afterwards.
EEG Analysis Tools Compared
The honest answer to "which EEG tool should I use" is that they solve different problems. Here is what each is strongest at, and where BCILattice fits.