BCILattice Docs

Getting Started Guide

Install -> import data -> first experiment in 30 min

FAQ

Focused answers for app, cloud, data, teams, privacy, and troubleshooting

Feature Status

What's production-ready versus experimental, in the app's own words

Lab Plan & Team Sharing

Private teams, dataset manifests, roles, and activity records

Public Study Publishing

Publish to BCINexus Studies: version control, visibility

Study Sharing & Imports

Share local studies, cloud packages, private links, or team workspaces

Analysis Suite Reference

All 11 analysis steps: ICA, epoching, time-frequency, selection, features

Feature Extraction

71 feature families across 5 domains, and how the design matrix is built

Stages & Dependencies

What each step writes, and which input each step accepts

Technical Overview

Architecture, modules, data flow, and network topology

Technology Stack

PySide6 desktop UI, local FastAPI service, PyTorch, MNE-Python, and the engineering decisions behind them

System Requirements

OS, hardware, GPU, storage, and software dependencies

Security & Compliance

Implemented controls, compliance readiness, encryption, and audit records

Driver API (partners)

Hardware driver integration for device manufacturers

Legal & Terms

Terms of Service, Privacy Policy, Cookie Policy, IP

Security & Compliance

Security controls and compliance-readiness guidance

Enterprise Guide

On-premise, air-gap, legal review, dedicated support, how to buy

Lab Plan Guide

Team sharing, shared data, collaboration controls

Using the ML Suite

MLFlow, Workflow and Training: how the three steps combine

Smart Training

One checkbox that configures, optimises and diagnoses a run — and records every decision

Multimodal & Media

Text, image, audio and video bound to the recording clock: sets, the four binding routes, six blocks, the paradigm runner

Pipeline Recipes

Channel selection, PCA, cross-validation, LOSO, domain adaptation

Block Index

Every block we make, in one line each, with a link to its full description

Custom Blocks

Inputs, outputs, graphs and the everyday BCILattice palette

Advanced Blocks

Control flow, domain adaptation, continual learning, training

Extensions

24 curated packages, or any spec, turned into sandboxed palette blocks

Validation & Reporting

Hold-out, cross-validation or Evaluation Protocol: one design per run, what each scenario reports, never mixed

Results Validation

ANOVA, t-tests and post-hoc families across sessions, pipelines and experiments

Results Export & Notebook

Every result table as CSV or Excel, and a notebook to write the analysis we did not build

Doc Editor

Write the whole report in the study: templates, real pages, pipeline snapshots, linked results

Report Integrity

Intervals, chance level and per-class results on every number; a check before every export; marking, releases and sign-off

Model Catalog

The bundled BCI model classes and their constructor parameters

BCINexus Edge

Signed model files on Raspberry Pi and Jetson: rate, source, your code, device checks

Using a model in your code

model_loader, loading a model file from any path

Limitations

Hardware, models, timing, connectivity and security, stated plainly

AI Assistant Guide

Ask about your experiments and published BCINexus work, with citations

Build with AI

Generate a starting graph for either canvas from a plain-language description

NeuralFlow Paradigm Builder

Visual drag-and-drop BCI paradigm design without code