- Home
- Skills
- APIs & Backend
- bci system architect
Works with the AI Agents you already use
bci system architect
Design and optimize end-to-end Brain-Computer Interface (BCI) architectures with a focus on real-time neural decoding.
$200
/moSubscription · cancel anytime
Secure checkout via Stripe
- Always the latest version
- Delivered live to your agent via MCP
- Cancel anytime, access ends at period end
bci system architect
Example session with this skill installed
Architect a real-time motor imagery BCI system using non-invasive EEG for controlling a motorized wheelchair. Focus on latency and artifact rejection.
- Read your context and instructions
- Compiled the bci system architect
I have designed a real-time BCI architecture for motorized wheelchair control using a non-invasive EEG paradigm. The system prioritizes low-latency motor imagery decoding and robust artifact rejection to ensure user safety.
BCI System Architecture: Motor Imagery Wheelchair Control
| Layer | Specification |
|---|---|
| 1. Signal Source | Non-invasive EEG (focus on C3, Cz, C4 motor cortex areas) |
| 2. Acquisition | 16-channel dry/semi-gel electrodes, 500Hz sampling rate, 24-bit ADC |
| 3. Preprocessing | Bandpass (8-30 Hz), Notch (50/60 Hz), CAR re-referencing, ASR for motion artifacts |
| 4. Feature Extraction | Common Spatial Patterns (CSP), Log-variance of band power |
| 5. Decoding | Linear Discriminant Analysis (LDA) with adaptive bias |
| 6. Actuation | ROS2-based wheelchair controller via ZeroMQ; Haptic feedback for state |
Latency Budget
- Acquisition Buffer: 64 ms
- Preprocessing (ASR + Filtering): 20 ms
- Feature Extraction: 10 ms
- Inference (LDA): 5 ms
- Actuation Command: 40 ms
- Total Closed-Loop Latency: ~139 ms (Well within the 250ms target)
Recommended Tech Stack
- Streaming: LSL (Lab Streaming Layer) for synchronized data ingest.
- Processing: Python (MNE-Python for offline, Scikit-learn for online decoding).
- Communication: ZeroMQ for low-latency messaging between the decoder and the wheelchair's ROS2 node.
Fail-safe: Heartbeat monitor; if signal confidence drops below 60% for 3 consecutive epochs, the wheelchair initiates an emergency stop.
Next steps
- Perform a spatial filter calibration session to compute user-specific CSP weights.
- Validate the ASR thresholds against typical wheelchair vibration noise profiles.
- Conduct a pilot test in a controlled environment to tune the LDA classification bias.
Subscription · always the latest version
- Secure checkout via Stripe
- Cancel anytime
- Delivered live via MCP
About this skill
Designing a Brain-Computer Interface (BCI) requires balancing biological signal constraints, hardware sampling, and real-time latency. This skill provides a structured framework for architects to design, review, and optimize neural interface systems ranging from non-invasive EEG wearables to high-bandwidth intracortical implants. It translates neuroscience requirements into deployable engineering specifications.
What it does
- Architectural mapping provides a six-layer system design covering signal source, acquisition, preprocessing, feature extraction, decoding, and actuation.
- Modality selection evaluates tradeoffs between EEG, ECoG, fNIRS, and intracortical signals based on invasiveness and resolution.
- Pipeline optimization defines specific cleaning recipes for EEG artifacts and intracortical spike sorting.
- Latency budgeting calculates end-to-end closed-loop timing to ensure motor control stays under the critical 250ms threshold.
- Algorithm selection recommends decoders like CSP, LDA, or Transformers based on paradigms like P300, SSVEP, or motor imagery.
How it works
- Define the paradigm by specifying the intended use case such as neuroprosthetics or silent speech.
- Layer analysis proceeds through the canonical six-layer pipeline to identify bottlenecks.
- Budgeting and stack selection determines the hardware sampling rates and software frameworks like LSL or ONNX.
- Safety and compliance review applies regulatory frameworks (FDA/MDR) and fail-safe protocols to the design.
Frameworks & tools
The skill supports Python-based processing with MNE and SciPy, low-latency execution in Rust or C++, and real-time data streaming via LSL (Lab Streaming Layer). Inference is optimized for ONNX Runtime, TensorRT, and CoreML.
Why this beats prompting it yourself
Generic models often fail to account for the specific temporal resolution requirements of neural signals or the safety density limits of cortical stimulation. This skill enforces a rigorous six-layer architecture that prevents common pitfalls like ignoring non-stationarity or hidden buffer latency.
Use cases
- Design a real-time EEG motor imagery pipeline for a rehabilitation exoskeleton.
- Review the signal-to-noise ratio and artifact rejection strategy for a new BCI wearable.
- Develop a latency budget for a high-performance intracortical cursor control system.
- Map out the regulatory and safety requirements for a clinical-grade neural implant.
Known limitations
Does not provide raw medical advice or surgical procedures. Architecture designs require validation against specific hardware data sheets and clinical safety standards.
Find more skills like this on the Claude Skills Marketplace, the Codex Skills Marketplace and the Gemini Skills Marketplace, or browse Claude Skills for APIs & Backend.
Does not provide medical advice. Requires hardware-specific data sheets for exact buffer tuning. Architecture designs must be validated by clinical safety professionals.
How to use
- 1
Connect your agent to the Agensi MCP
Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup
- 2
Sign in to Agensi
Your agent shows a sign-in link. Sign in and click Allow.
- 3
Ask for what you need
Your agent loads this skill and uses it on your task. It always asks before buying a skill.
Reviews
No reviews yet
Be one of the first to try it. Every listed skill goes through the trust checks below.
Security scanned
Checked by our 8-point scan before listing
Fresh listing
Recently published to Agensi
Free to use
No account required to browse
Trust & safety
Security scan
Security scan passed 22 days ago
Permissions
No special permissions declared or detected