Quick & EZ operation for connecting your Muse to an OSC Output

Connecting a Muse device to your computer is a mess, with LSL libraries being out of date, impractical and unstable. This repo aims to simplify that with a streamlined python script. This allows you to stream raw or processed EEG data to an OSC port of your choice.

Simply run :

python __muse_monitor.py__ --EZ --osc

This operation will automatically do a few things in sequence.

1. Find the USB address of the BLED112 device
2. Find the Muse (must be in connection mode) with the BLED device
3. Start streaming to the default OSC port for TouchDesigner or other uses

data flow

send data from an input -> transform -> output

from eegstreamer.inputs import RandomEEGInputStream
from eegstreamer.transforms import ConstantFactorEEGTransformer
from eegstreamer.outputs import JSONFileOutputStreamer

# create the output stream
output_stream = JSONFileOutputStreamer('eeg_output.json')

# create the stream to transform the data
transform_stream = ConstantFactorEEGTransformer(factor=0.2)
# tell the transform stream where to send its data
transform_stream.connect(output_stream)

# create the input stream
input_stream = RandomEEGInputStream(sample_rate=10, sample_duration=10)
# tell the input stream where to send its data
input_stream.connect(transform_stream)

# start the input stream
input_stream.start()

input can be streamed to multiple endpoints

graph TD;
    Input-->Transform;
    Input-->HttpEndpointOne
    Input-->ScreenOutput
    Transform-->FileOutput
    Transform-->HttpEndpointTwo
    Transform-->ScreenOuptput

Loading

from eegstreamer.outputs import HttpPostEEGOutputStreamer, JSONFileOutputStreamer
from eegstreamer.transforms import ConstantFactorEEGTransformer
from eegstreamer.inputs import RandomEEGInputStream

endpoint_one = HttpPostEEGOutputStreamer("https://sample.com/rawd")
endpoint_two = HttpPostEEGOutputStreamer("https://sample.com/processed")
file_output = JSONFileOutputStreamer("eeg_output.json")

transformer = ConstantFactorEEGTransformer(factor=10)
transformer.connect(file_output)
transformer.connect(endpoint_two)

input_stream = RandomEEGInputStream(sample_rate=10, sample_duration=10)
input_stream.connect(transformer)
input_stream.connect(file_output)

input_stream.start()

Development

Due to new requirements for packaging python modules, the following command needs to be run prior to doing development:

pip install -e .

This creates a symbolic link so that python knows that the eegstreamer package can be found underneath the src directory.

Also, requirements are now part of the pyproject.toml file, so this will also install the latest dependencies.

If you also want to use the behave testing framework for development, this will install the extra dependencies needed:

pip install -e .'[dev]'

Muse device

brew install labstreaminglayer/tap/lsl

OR

conda install -c conda-forge labstreaminglayer/tap/lsl

THEN

pip install -e .'[muse]'

Known issues

1. If two applications try to connect to the same muse device. The thread running inside bgapi throws an exception with no way of catching it.
2. The system will run usually for about 20-30 min. If headset disconnects mid stream (?), just re-run the script
3. We have seen an error where everything will be physically connected but the Muse will not connect. In those instances, disconnect and re-seat all physical connections (i.e. USB)