MNE Scan Hardware Support Matrix
This page enumerates every plugin shipped under src/applications/mne_scan/plugins/, the device or data source it interfaces with, and the build-time / runtime constraints that determine whether a given plugin is usable on your system.
Plugins fall into three roles:
- Sensor — acquires data from a piece of hardware or a network stream.
- Algorithm — processes data inside the running pipeline (no hardware contact).
- IO — writes results to disk or downstream consumers.
Only the Sensor plugins are tied to physical hardware; the others are pure data-flow stages and run wherever MNE-CPP itself builds.
Sensor Plugins
| Plugin | Device / Source | Default Sampling Rate | Channel Counts | OS Support | Vendor SDK / Driver | Build Flag |
|---|---|---|---|---|---|---|
babymeg | BabyMEG (Boston Children's Hospital pediatric MEG) | 1000 Hz (configured by acquisition host) | Full BabyMEG channel set delivered by the BabyMEG server (magnetometers + references + trigger) | Linux, macOS, Windows | None for the plugin itself — relies on the BabyMEG TCP server (babymegclient) running on the acquisition host | Built by default |
brainamp | Brain Products BrainAmp EEG amplifier | 1000 Hz (driver runs internally at 5000 Hz and downsamples) | Up to the amplifier's configured montage (16/32/64/128 channels) | Windows only | BrainAmp SDK header BrainAmpIoCtl.h; USB driver installed on host | -DWITH_BRAINAMP=ON |
eegosports | ANT Neuro eego sports EEG amplifier | 512 Hz (also 500 / 1000 / 2000 Hz) | Up to amplifier model (8 / 32 / 64 / 88 channels) | Windows only | eemagine SDK (eego-SDK.dll, eego-SDK.lib); USB driver; per-device license files for older SDKs | -DWITH_EEGOSPORTS=ON |
fiffsimulator | Pre-recorded FIFF file streamed through mne_rt_server | Inherited from the source FIFF (typically 600–5000 Hz) | Inherited from the source FIFF | Linux, macOS, Windows | None — uses the in-tree mne_rt_server | Built by default |
ftbuffer | FieldTrip Buffer (TCP, default port 1972) | Inherited from the upstream producer (e.g. Neuromag / MEGIN via neuromag2ft) | Inherited from the FieldTrip header | Linux, macOS, Windows | None for the plugin; the producer side (e.g. neuromag2ft) must be installed on the acquisition computer | Built by default |
gusbamp | g.tec g.USBamp EEG / biosignal amplifier (up to 4 chained units) | 128 Hz (also 256 / 512 / 1200 / 2400 / 4800 / 9600 / 19200 / 38400 Hz) | Up to 16 channels per unit × up to 4 units (master + 3 slaves) | Windows only | gtec_gUSBamp.h and gUSBamp_x64.lib / gUSBamp_x86.lib from the g.tec SDK; USB driver | -DWITH_GUSBAMP=ON |
lsladapter | Lab Streaming Layer (LSL) outlets discovered on the local network | Inherited from the LSL stream | Inherited from the LSL stream | Linux, macOS, Windows | None — uses the in-tree mne_lsl library | Built by default |
natus | Natus Quantum amplifier (clinical EEG, phase-2 epilepsy monitoring) | 2048 Hz | Up to the amplifier configuration (typical 32–128 EEG + auxiliary) | Windows only (requires Natus SDK v8.4 host service SelectStreamingMachine) | Natus SDK v8.4; startDataTransmit.bat running on the streaming machine | -DWITH_NATUS=ON |
tmsi | TMSi Refa / Porti EEG amplifier | 1024 Hz | Up to amplifier model | Windows only | TMSiSDK.dll installed at C:\Windows\System32\TMSiSDK.dll; vendor USB driver | -DWITH_TMSI=ON |
Live-Status Checklist (per Sensor Plugin)
Use this checklist whenever you bring up a new acquisition setup. The plugin will not start streaming unless every item is green.
- Vendor driver installed and the device enumerates in the OS device manager (
lsusb/ Device Manager). - Vendor SDK library file is present in the search path (
*.dllnext tomne_scan.exe, or*.so/*.dylibunder the install prefix). - License / activation file present (EEGoSports only, for SDK versions older than the bundled one).
- Amplifier powered on before MNE Scan is launched.
- Sampling frequency and block size configured in the plugin's Preferences tab and match the amplifier's allowed values.
- FIFF info reports the expected channel count after
Init(visible in the Channel Selection widget). - No
[ERROR]lines tagged with the plugin name in the MNE Scan log dock when you press the green Run button.
Algorithm Plugins
| Plugin | Role | Notes |
|---|---|---|
averaging | Real-time evoked-response averaging with baseline correction and artifact thresholds | Triggered by stimulus channels in the input FIFF info |
covariance | Online noise-covariance estimation | Output is consumable by rtcmne |
hpi | Continuous HPI coil fitting and head-position monitoring | Requires a MEG input with active HPI coil channels; coil-frequency presets are read from hpi.json |
neuronalconnectivity | Real-time functional connectivity estimation | Input: source-localised data from rtcmne |
noisereduction | SSP projection, gradient compensation, and real-time filtering | Replaces several offline pre-processing steps |
rtcmne | Real-time minimum-norm source estimation | Requires forward solution (rtfwd or pre-computed) + noise covariance |
rtfwd | Real-time forward-solution recomputation, driven by HPI updates | Pairs with hpi for moving-subject MEG |
IO / Utility Plugins
| Plugin | Role | Notes |
|---|---|---|
dummytoolbox | Reference / template plugin for new plugin authors | Not intended for production use |
writetofile | Streams the pipeline output to a FIFF file on disk | One file per Run; filename pattern configurable in the plugin's GUI |
Plugin Count
18 plugins in total: 9 sensor, 7 algorithm, 2 IO/utility.