Requirements

Install

1. Open bcinexus.xyz/download.
2. Choose the installer or package available for your operating system.
3. Install BCILattice using your normal institutional or personal software policy.
4. Launch the app and wait for the local FastAPI service to become ready.

If you run the app from source, follow the repository environment and dependency setup rather than installer instructions.

Account

BCILattice requires a BCINexus account to operate. After sign-in, analysis and model training run on your machine; cloud sync, community imports, publishing, private share links, dashboard access, and team collaboration are optional connected features.

1. Create an account at bcinexus.xyz/register or from the desktop sign-in flow.
2. Verify your email if prompted.
3. Sign in from Settings, Account, or from the module that requests authentication.

Create a Study

1. Open Studies from the desktop sidebar.
2. Create a new local study with name, modality, task type, and description.
3. Save local work as a .blts study.
4. Use the study workspace to move through Data Manager, Preprocessing, Analysis, Neural Flow, ML Suite, Experiment Hub, Results, Model Analysis, and Reports.

Import Data

Use Data Manager as the source for files, sessions, subjects, modalities, and labels.

1. Open Data Manager in the active study.
2. Choose a single file import or folder import for multi-subject data.
3. Select the modality and confirm label columns or label references.
4. Review previews and quality warnings before adding the data.

Preprocess

Open Preprocessing after data import. Start conservatively and document every choice.

Design a Paradigm

Use Neural Flow to define cues, task blocks, rest periods, marker codes, timing, and loops.

1. Open Neural Flow from the study workspace.
2. Add task, cue, feedback, timing, marker, or loop blocks.
3. Set duration, label, marker code, and repeat settings.
4. Compile the graph and resolve validation messages.
5. Save or export the paradigm for use with the study.

Build & Train

Use ML Suite for MLFlow graph design, Workflow assignment, and training.

1. Start with a simple baseline graph such as channel selection, scaler, LDA/SVM, and metric blocks.
2. Use Workflow to bind Data Manager files, subjects, and sessions to the pipeline.
3. Compile before training and fix any graph errors.
4. Train locally on CPU or a compatible GPU.
5. Move to deep models after the baseline and data splits are sensible.

Review Results

1. Open Experiment Hub after training completes.
2. Inspect run records, metrics, and configuration differences.
3. Use Results for metric summaries and cross-experiment comparison.
4. Use Model Analysis for SHAP, weights, activation maps, or other diagnostics where supported.
5. Export run CSVs when you need external statistical analysis.

Generate a Report

1. Open Reports from the study workspace.
2. Generate the report from the current study state.
3. Review the report editor before export.
4. Export PDF or HTML for people, JSON for study documents, or CSV for metric tables.

Use AI Chat

AI Chat connects the desktop to your selected provider or local Ollama endpoint. It is useful for interpretation, troubleshooting, report drafting, and research questions, but it should not replace methodological review.

1. Open AI Chat and connect a supported provider or local Ollama.
2. Ask focused questions such as "Why might subject 3 accuracy be low?" or "What preprocessing checks should I run next?"
3. Review the answer, then make changes deliberately in the relevant module.

Community

Use BCINexus community and study pages when you want to reuse published artifacts or publish selected study outputs.

1. Browse BCINexus Studies or use the desktop community import entry points.
2. Import only artifacts that match your modality, task, data shape, and validation assumptions.
3. Publish selected artifacts through the desktop Publish flow when your study is ready to share.

Troubleshooting