BCILattice is a desktop application for brain-computer interface (BCI) research. It provides a visual paradigm designer (NeuralFlow), a machine learning pipeline builder (ML Suite), multi-subject batch training, experiment tracking, and one-click report generation, all running locally on your machine without requiring coding knowledge.

Yes, the Free plan is available at no cost indefinitely. It includes 3 saved pipelines, 500 MB cloud storage, 1 upload per month, and access to all local processing features. Paid plans (Researcher, Lab) add higher limits, model export, XAI, and team workspaces.

No. BCILattice is designed for researchers who prefer visual, point-and-click workflows. You can build full preprocessing + ML pipelines without writing any code. An optional Custom Code block lets experienced users inject Python snippets if needed.

No. BCILattice supports EEG, fNIRS, and EMG data. Preprocessing pipelines adapt per modality: Beer-Lambert law conversion for fNIRS, rectification and envelope extraction for EMG. Future versions will extend support to additional biosignal modalities.

Yes, all processing and training runs entirely on your local machine. Internet access is only required for the initial download, account creation, cloud sync, and accessing community pipelines. The Free plan can be used indefinitely without internet.

BCILattice's NeuralFlow paradigm designer includes blocks for Motor Imagery (left/right hand, feet, tongue, rest), Motor Action (reach/grasp, hand open/close, gait), Cognitive tasks (arithmetic, working memory, mental rotation, reading, spatial navigation), SSVEP, P300, N-back, and more. It supports any event-related paradigm via the Marker block.

BCILattice is developed by BCINexus. The platform was built by and for BCI researchers, with the goal of making state-of-the-art ML tools accessible without requiring software engineering expertise.

Minimum: 8 GB RAM, 4-core x86-64 CPU, 5 GB free storage, Windows 10 / macOS 12 / Ubuntu 20.04. An NVIDIA GPU with CUDA is recommended for faster training but not required. See the System Requirements page for full details.

No. BCILattice bundles Python 3.10 internally along with all required ML libraries. You do not need any existing Python installation.

The standard installer on Windows requires administrator access to install to C:\Program Files\BCILattice\. A user-level installer (no admin rights) is available on the same download page. On macOS and Linux, no root access is required.

This means the local backend (PostgreSQL + FastAPI) didn't start. Most commonly another process is using the same port. Try restarting BCILattice. If the issue persists, check View → System Info for the error log and contact support.

Yes. BCILattice runs natively on Apple Silicon with Metal acceleration for ML training. The macOS installer is a Universal binary, no Rosetta required.

Yes. Your BCINexus account licence covers one active installation. If you need multiple concurrent machines, contact support. Lab and Institution plans include more flexible multi-device policies.

EEG: EDF, BDF, GDF, FIF, VHDR (BrainVision), SET (EEGLAB), CNT, CSV, XLSX. fNIRS: SNIRF, FIF, TXT, CSV. EMG: EDF, BDF, GDF, FIF, TXT, CSV.

Yes. Use Import Folder in Data Manager. BCILattice scans a root directory recursively (up to 5 levels), assigns subject IDs automatically, and creates a single multi-subject session. You can then assign all subjects to a pipeline in Workflow Designer.

All session data and preprocessed files are stored locally in your BCILattice data directory (configurable in Settings). If cloud sync is enabled, session metadata and pipeline configs are backed up to BCINexus cloud, raw recording files are never uploaded unless you explicitly upload them via Data Manager.

Yes. From Data Manager, right-click any session → Export Preprocessed. You can export to EDF, CSV, or FIF. Preprocessed NumPy arrays can also be exported directly for use in external scripts.

No. BCILattice never modifies source files. All preprocessing creates new derived datasets stored separately. Your original recordings are always preserved.

Local imports have no hard limit, they are only constrained by your disk space and RAM. Cloud uploads are limited to 1 GB (Free), 1 GB (Researcher), or 5 GB (Lab) per file.

NeuralFlow is BCILattice's visual paradigm designer. You drag timing blocks onto a timeline canvas to define the sequence of stimuli and events in a BCI experiment. The compiled output is a .nflow file that can be loaded during data acquisition or simulated offline.

ML Suite (also called MLFlow internally) is a node-graph pipeline builder with 5,400+ processing and model blocks spanning signal processing, feature extraction, classification, regression, cross-validation, and evaluation. Pipelines are compiled and trained locally.

Workflow Designer extends ML Suite with multi-subject batch capabilities. It adds file selection, graph endpoints, and output blocks for running a single pipeline across many subjects automatically.

Experiment Hub is BCILattice's experiment tracking system. Every training run is saved as a versioned snapshot (v1, v2, v3…) including the pipeline configuration, hyperparameters, and results. You can compare, restore, and export any version.

The AI Research Assistant (Researcher plan+) is a context-aware chat interface that understands your current session, pipeline, and results. It uses your own API key (BYOK, Bring Your Own Key) so your data never leaves your machine. It can suggest preprocessing settings, explain model decisions, and help interpret results.

BCILattice includes SHAP (SHapley Additive exPlanations) for feature importance, activation maps for deep learning models, and topographic EEG scalp maps for spatial visualisation. XAI is available on Researcher and Lab plans.

Yes. Researcher and Lab plan users can export trained models to ONNX and PyTorch formats for deployment in external applications or embedded devices. Available from Experiment Hub → Export Model.

Yes. Go to Reports → Generate Report (or press Ctrl+R / Cmd+R). BCILattice generates a standalone HTML or PDF report including session metadata, preprocessing config, paradigm summary, and per-subject results, ready to share or submit as a deliverable.

BCILattice's ML Suite includes EEGNet, DeepConvNet, ShallowConvNet, EEGTransformer, LDA, SVM, Random Forest, XGBoost, and 5,000+ additional blocks from the community pipeline catalog. You can also use Custom Code blocks to inject scikit-learn compatible estimators.

Yes, when an NVIDIA GPU with CUDA is detected. PyTorch deep learning models (EEGNet, DeepConvNet, etc.) run on the GPU automatically. CPU-only training is always available as a fallback. Apple Metal acceleration is used on Apple Silicon Macs.

BCILattice supports k-fold, stratified k-fold, leave-one-subject-out (LOSO), and leave-one-trial-out cross-validation, configurable via the Cross Validation blocks in ML Suite. Results are aggregated per subject and across subjects.

Yes. Connect a GridSearch or RandomSearch block to your pipeline. BCILattice will train the model across the parameter grid and report the best configuration in Experiment Hub.

Check View → System Info. If GPU is shown as "Not detected", install CUDA matching your driver version and restart BCILattice. Reducing batch size, using a lighter model (LDA instead of EEGNet for a quick baseline), or reducing the number of CV folds will also significantly speed up training.

Yes. Browse pipelines at bcinexus.io/pipelines, click Import into BCILattice, and the pipeline opens automatically. Community pipelines are available to users with a free BCINexus account.

No. Cloud sync is entirely optional. BCILattice works fully offline. Connecting a BCINexus account enables cloud backup of session metadata and pipeline configs, plus community pipeline access.

By default: pipeline configs, session metadata (subject count, recording params, result summaries), and NeuralFlow paradigm files. Raw recording files and preprocessed data are not synced unless you explicitly upload them via Data Manager. You control what goes to the cloud.

Yes. All data in transit uses TLS 1.3. Data at rest is encrypted using AES-256. See the Security & Compliance page for full details.

Primary servers are hosted in the EU (AWS eu-west-1 / Ireland) with replication to US-East (AWS us-east-1). Enterprise customers can request exclusive EU or US data residency. Institution/Enterprise customers can opt for on-premise deployment with no cloud dependency.

Yes, on Institution and Enterprise plans with a signed Business Associate Agreement (BAA). All compute runs locally, patient data never leaves your network unless you explicitly enable cloud sync. See Security & Compliance for full HIPAA guidance.

Free (always free), Researcher ($15/month), Lab ($99/month for up to 5 team members), Institution (custom quote), and Enterprise (custom quote with on-premise option). Academic users get 20% off Researcher and Lab plans. See bcinexus.io/pricing for current pricing.

Yes. The Researcher plan includes a 7-day free trial, no credit card required. Lab plan trials are available on request via support@bcinexus.io.

Yes. Cancel anytime from your account settings at bcinexus.io/account. Your plan remains active until the end of the current billing period. Data is preserved for 90 days after cancellation.

Your data is always preserved for 90 days after a downgrade. If you exceed the lower plan's storage limit, you'll be prompted to export or delete data before the grace period ends. Pipelines and sessions are never deleted without your action.

Yes, 15% off Researcher and Lab plans. Sign up or log in with your institutional or educational email, fill in your details (institution, department, role), and submit, no photo ID or document uploads required. If your email domain is recognised the discount is applied instantly. See the Academic Program guide.

Yes. Institution and Enterprise plans support purchase order, wire transfer, and institutional billing systems. Contact enterprise@bcinexus.io to arrange procurement.

The Lab plan adds a shared workspace for up to 5 team members. All members share a 50 GB cloud storage pool, a shared dataset library, shared projects, and a real-time activity feed. SHA-256 checksum verification ensures every team member runs experiments on identical data.

Yes. Any BCINexus user can be invited to a team workspace regardless of which institution they belong to. Cross-institution collaboration is fully supported.

Owner (1 per workspace), Admin, Member, and Viewer. Owners and Admins can manage members and delete data. Members can upload and delete their own uploads. Viewers have read-only access. Roles are set in Teams → Workspace Settings → Members.

The Lab plan covers one workspace with up to 5 members, the workspace owner pays. Each member must have a free BCINexus account to connect. For more than 5 members, the Institution plan covers a larger team with a seat count defined per contract.

The Institution plan supports larger teams with a seat count agreed at contract time. Email sales@bcinexus.io with your organisation name and expected headcount for a custom quote.

No. All signal processing and ML training runs locally on your machine. Raw data files are never automatically uploaded. Only session metadata and pipeline configs are synced if you enable cloud backup, and only if you choose to. You have complete control.

The AI Research Assistant uses your own API key (BYOK). Context sent to the AI includes only your current pipeline configuration, summary statistics, and result metrics, not raw recordings. You can inspect exactly what is sent in Settings → AI Assistant → Privacy Log.

Yes. BCINexus is registered as a data processor under GDPR. We provide a Data Processing Agreement (DPA) for Institution and Enterprise customers. EU data residency is available. See Security & Compliance for details.

Yes, on Institution/Enterprise plans with a signed HIPAA Business Associate Agreement. Since processing runs locally, patient data never leaves your secure environment (assuming cloud sync is disabled or restricted). Contact legal@bcinexus.io to arrange a BAA.

Go to Account Settings → Data → Delete All My Data at bcinexus.io/account. This permanently deletes all cloud-stored data associated with your account. For GDPR right-to-erasure requests, email privacy@bcinexus.io.

Verify the file is not corrupted by opening it in another application (e.g., EDFbrowser). Check that the file extension matches the actual format (e.g., a .txt renamed to .edf will fail). If unsure, convert to EDF using EDFbrowser or MNE-Python, then re-import.

The most common cause is a port type mismatch, a data port (blue) connected to a timing port (orange). Check the colour-coded port indicators on each block. Hovering over a port shows its expected type. Also verify every block has all required inputs connected.

Open View → System Info. The CUDA version detected is shown there. Install CUDA matching your NVIDIA driver version from developer.nvidia.com/cuda-downloads. Restart BCILattice after installing. AMD GPUs are not currently supported.

Yes. Use File → Export Session Archive (.bcis). This creates a portable file with your session, pipeline, and results. Open it on the other machine with File → Import Session Archive. Cloud sync (Researcher+) makes this even easier, sign in on the new machine and your sessions appear automatically.

Email support@bcinexus.io. Community forum: community.bcinexus.io. Lab plan users have priority 48-hour response. Institution/Enterprise users have a dedicated engineering contact with a 4-hour SLA.

Yes. Go to Help → Release Notes in BCILattice, or visit bcinexus.io/changelog for a full version history. BCILattice updates automatically in the background when connected to the internet.