Hardware
Devices
JNEEG
JNEEG Intro

JNEEG

A brain-computer interface with a GPU on board.

JNEEG is an open-source shield for the NVIDIA Jetson Nano. It plugs onto the 40-pin header and measures 8 channels of biosignals (EEG, EMG and ECG) with a Texas Instruments ADS1299. Because the Jetson Nano has a GPU, you can run neural networks on the EEG right on the device: real-time classification, neural decoding, or EEG combined with a camera.

JNEEG shield on a Jetson Nano

⚠️

JNEEG is not a medical device. Run it from a 5 V battery only, never from the mains. Read Safety before the first recording.

At a glance

Channels8 biosignal inputs, plus reference and bias
ConverterTI ADS1299, 24-bit, simultaneous sampling
Sample rate250 SPS
SignalsEEG, EMG, ECG
HostNVIDIA Jetson Nano Developer Kit
Link to the JetsonSPI on the 40-pin header, DRDY on pin 7
ElectrodesDry or wet Ag/AgCl, 2.54 mm header
Power5 V power bank, 3 A, up to 10,000 mAh
SoftwarePiEEG Server (--device jneeg), Python scripts

How it works

 Electrodes 1–8 ─┐
 Reference (ear) ┼─► JNEEG ──SPI──► Jetson Nano ──► PiEEG Server
 Bias (ear) ─────┘   (ADS1299)      (GPU)          ├ dashboard :1617
                                      ▲            ├ WebSocket :1616
           5 V power bank ────────────┘            └ your ML models
  1. Measure: the ADS1299 digitises all 8 channels at the same moment, 250 times per second.
  2. Transfer: the Jetson Nano reads each sample over SPI when the shield signals data ready.
  3. Use: stream it with PiEEG Server, or process it on the GPU with your own models.

JNEEG kit: the shield on the Jetson Nano, ear clips, EEG cap and power bank

JNEEG or PiEEG-8?

PiEEG-8JNEEG
HostRaspberry Pi 4 / 5NVIDIA Jetson Nano
Channels88
ConverterADS1299ADS1299
GPUNoYes, 128 CUDA cores
PiEEG Server--device pieeg8--device jneeg
Best forGeneral BCI, researchReal-time ML on EEG, EEG + computer vision

What you can measure

SignalWhat it showsTypical first test
EEGBrain rhythms: delta, theta, alpha, beta, gammaClose your eyes and watch alpha (8–13 Hz) rise
EMGMuscle activityClench the jaw or a fist
ECGHeart activityElectrodes on the chest or wrists

See a real recording in Signals & Datasets.

Get started

Related

Citation

If you use JNEEG in research, please cite:

I. Rakhmatulin. Device JNEEG to convert Jetson Nano to brain-computer interfaces. Short report. arXiv:2301.11110, 2023. arxiv.org/abs/2301.11110 (opens in a new tab)

Support