Feature Status
What's validated and production-ready in BCILattice, what's still experimental, and exactly what that means before you rely on it.
Why This Page Exists
Every module in BCILattice is at a different point of maturity. Rather than presenting the whole app as equally finished, this page lists each module's actual status, in the same language the desktop app itself uses at the point of use, so what you read here matches what you see when you open it. If a feature is still experimental, we say so here and in the app, not just in one place.
How We Label Maturity
| Label | What it means |
|---|---|
| Production | Used across the standard research workflow, stable, no outstanding hardware-validation gap. |
| Experimental | Ships and works, but has not yet been validated against physical acquisition or deployment hardware. Flagged in-app with an "EXPERIMENTAL" tag and a dismissible warning banner. Off by default in the sidebar can be hidden entirely via Settings. |
This is a maturity label, not a regulatory one. See Research & Clinical Scope below, nothing on this page, Production or Experimental, is a claim of medical-device clearance.
Production Modules
The standard research pipeline, stable and in everyday use:
| Module | What it does |
|---|---|
| Data Manager | Import, scan, and organise EEG / fNIRS / EMG / ECG / EOG files from disk |
| Preprocessing | Signal filtering, cleaning, and artifact removal |
| Neural Flow | Visual BCI paradigm design |
| Analysis | Feature extraction and physiological signal analysis |
| ML Suite | MLFlow pipeline design, Workflow Designer, and local model training |
| Experiment Hub | Run history: compare, inspect, export, and restore experiment state |
| Model Analysis | Explainability: SHAP attributions, weights, activation maps |
| Reports | Statistical summaries and exportable PDF reports |
| Multimodal & Media | Text, image, audio and video stimulus sets bound to the recording clock by four routes, six Media blocks, the paradigm runner with a measured onset band, BIDS and PsychoPy export, and consent-gated response capture, see Multimodal & Media. Audio and video generation is not shipped: it is gated on a measurement that has not been run |
| Report Integrity | Confidence intervals, chance level and per-class results on every reported number; a pre-export check for unresolved values, uncited references, uncaptioned figures and empty sections; confidentiality marking, released versions and sign-off, see Report Integrity |
| Results | Study-wide results across experiments, best metrics, comparison, and statistical validation (Results Validation) |
| Results Export & Notebook | Every result table as CSV / Excel plus a starter .ipynb (free on every plan), and an in-app Python notebook over those tables for custom charts and custom validation (running it is Researcher and up), see Results Export & Notebook |
| AI Chat | Chat with your own connected provider using session context. Outputs are not validated for regulatory, clinical, or safety-critical use, see Legal |
| Participants & Clinical Mode | Local-first participant/consent registry and subject-linking. Informed-consent and subject-tracking bookkeeping for IRB-approved research, not a clinical decision system, see Research & Clinical Scope |
| Community | Browse and import community pipelines |
Experimental Features
These two ship today and work, but the underlying hardware path has not yet been validated against physical devices. This is the exact wording the desktop app itself shows in its sidebar and on the feature page:
| Feature | In-app disclaimer |
|---|---|
| Record Session (live LSL streaming acquisition) | "EXPERIMENTAL (unvalidated on acquisition hardware)" — stream discovery, channel labels, and sample rates are read from whatever the LSL outlet reports; verify them against your amplifier before relying on a recording. |
| Edge Devices | "EXPERIMENTAL (unvalidated on physical hardware)" — deployment to Raspberry Pi, Jetson, Arduino, and ESP32 has not yet been validated on physical hardware; treat results as unverified until confirmed on your own device. |
Research & Clinical Scope
BCILattice is research software. It is not a medical device, diagnostic product, treatment system, or a substitute for professional clinical judgment, Production or Experimental, nothing in this app carries a medical-device clearance. The Participant registry and Clinical Mode are informed-consent and subject-tracking tools for IRB-approved research; enabling Clinical Mode requires a Participant to have a consent record on file before a recording session starts, it does not verify that consent was properly obtained, and the "Doctor" account designation is an identity label only, it grants no clinical authority within the app.
Full terms: Legal & Terms, Section 9 (Research and Clinical Disclaimer).
Compliance Readiness
Security controls, encryption, and HIPAA / GDPR / SOC 2 readiness are covered in full on the dedicated Security & Compliance page, including what's currently certified versus what requires your organisation's own review (SOC 2: not currently certified; HIPAA: supportable, depends on your environment and a signed BAA).
Report an Issue
Found a feature behaving differently than this page describes, in either direction? Tell us, this page is only useful if it stays accurate.
Email [email protected] with the module name, your app version, and what you observed.