Which brainwaves are each of the 4 channels associated with?


So it says that: [SIZE=16px]
Muse detects a full range of brainwave activity. Brainwaves are typically broken up into five bands and Muse detects all five bands. The five bands are:[/SIZE]

  • Delta waves which are most present during sleep.
  • Theta waves which are associated with sleep, very deep relaxation, and visualization.
  • Alpha waves which occur when relaxed and calm.
  • Beta waves which occur when, for example, actively thinking or problem-solving.
  • Gamma waves which occur when involved in higher mental activity and consolidation of information.
but which waves are read by which sensors?


Hi to_the_sun,

I am just a beginer, and started a research in this matter just a few months ago, when I became interested in Muse. This forum is just begining and the Muse team takes to long to answer questions (we can understand, since they must be extremely busy since the launching four months ago). So, if I can be of any help I would sugest you to do a google search - you will find thousands of useful information (the MIT Open Courseware is a good source for neuroscience and also programming).

Just to start I could say that:

  • all four sensors (channels) detects only raw data (that is, all frequencies mixed together, read from each sensor position over the brain skin), in voltages up to less than 2000 microvolts in time.
  • the separation into the various know frequencies is made by software (muse-io with the --dsp parameter) using high math - Fast Fourier Ttransformations (FFT) that broke those reandings into the desired frequencies.
    Take a look at the Muse Developer Site https://sites.google.com/a/interaxon.ca/muse-developer-site/museio/osc-paths/3-4-0 and you will see where all those data are saved into your computer (the OSC paths). Then you can use another application (like muse-lab) to see those waves in separated graphs.
    [B]/muse/dsp/elements/low_freqs ffff[/B]
    1-8Hz, log band power (dB)
    [B]/muse/dsp/elements/delta ffff[/B]
    1-4Hz, log band power (dB)
    [B]/muse/dsp/elements/theta ffff[/B]
    5-8Hz, log band power (dB)
    [B]/muse/dsp/elements/alpha ffff[/B]
    9-13Hz, log band power (dB)
    [B]/muse/dsp/elements/beta ffff[/B]
    13-30Hz, log band power (dB)
    [B]/muse/dsp/elements/gamma ffff[/B]
    30-50Hz, log band power (dB)

HTH, Eduardo.


Ah! I spent all day yesterday looking at that page! How did I not see the most important parameters at the bottom? Thanks for the response, however, my headband does not seem to be transmitting these readings. I’ve been reading all the others fine (I’m using Max/MSP), but the /muse/dsp/ messages don’t come through. Is there something different about these ones? Something you have to do to start seeing them? You mention the --dsp parameter in association with muse-io… To get muse-io running i’ve been typing

muse-io --preset 14 --device <DEVICE_NAME> --osc osc.udp://localhost:5000

into a command prompt; does a --dsp need to be added here or something? In the tutorial it goes onto to talk about an
oscdump 5000

command, but I’ve never done that part. Could that be the problem? After typing the first command into the prompt and getting Muse to connect, there’s no more prompt to type into and I was already getting messages, so I hadn’t worried about it.
Thanks in advance for any info!


Unfortunatelly I did not receive my Muse yet, so I cannot test it, but based on what I have read: Yes, I believe that you need to add the --dsp parameter to the command line to get those algorithms pre-processed data, like bellow:

muse-io --device <DEVICE_NAME> --osc osc.udp://localhost:5000 --preset 14 --dsp (note: --device <DEVICE_NAME> is not needed if you have not changed the bluetooth name of your Muse.

Once you get muse-io connected to your Muse, it will start broadcasting data to your own computer (localhost = IP on port 5000, according to what is defined in preset 14.

Then you can use any of the tools already released to see all the data it’s sending, either as pure text screen dump, or in a graph (using muse-lab).

oscdump is one of these tools (screen dump) - try oppening another command prompt window and type:

oscdump 5000

to see all data, or if you want to filter just some data (like alpha or beta …) try:

oscdump 5000 | grep /muse/dsp/elements/alpha
(or /muse/dsp/bandpowers/alpha case OSC Paths versrion 3.4.0 was not yet released)

muse-player is another command line tool that accepts different tyes of INPUTS and OUTPUTS (I believe that if you type only muse-player it will list all parameters accepted and their syntax).

but, give also a try to muse-lab and you shoud see graph plots of the data you select (muse-lab is a java app, just double-clik it), check its tutorial at:
https://sites.google.com/a/interaxon.ca/muse-developer-site/muselab/tutorial (it may be a bit outdated, but the basic functionallity should be the same)

Let us know what you get, ok ?


Thanks guys this is interesting information. I want to observe if certain things change brain wave frequencies. Or how long after a session i remain calm.


Thanks! I’ve got all the dsp info streaming now. Haven’t tried out Muse Player or Lab yet and I honestly probably won’t, as Max is my language of choice and i’ve got graphs and everything i need there.

Must admit i was a little disappointed to see muse-io is just command line. Hopefully they’ll come up with a UI front end for it soon. I really don’t look forward to having to explain to my customers what to type (or even where to find the command prompt) and everything, just to get things connected… But if it works i’m well satisfied for now.


Could it be that version 3.4.0 is no longer the newest for osc paths? Thought i’d note that these paths i’m getting


vary from what’s stated here https://sites.google.com/a/interaxon…sc-paths/3-4-0

Accessing DSP OSC channels

Question: how are the numbers for each of the brainwaves supposed to be interpreted? It’s stated that the numbers coming in for each brainwave are the “log band power (dB)” for each specific range of frequencies. So the higher the numbers coming in for alpha for example, the calmer you are, simple as that? Would it be appropriate/ideal to just average the readings coming in from the 4 sensors?


that is a good question, would like to know that too!


to the sun you can use muse lab to visualize it, connect with OSC to 5000 click open port. then choose visualizer in the drop down… But i when i select the alpha beta gamma delta theta i doesnt show anything, whereas if i just visualize the eeg i can what it is reading. And I can influence it with blinking my eyes. But that is about it. Could you tell me how you read that?

And i think yes, the higher the activity in alpha and lower in beta would suggest that you are more calm than working on a math problem i think. I feel when you calibrate muse in the app it asks you think up stuff, by doing so you increase you beta waves therefore it can get a more accurate read. But this is just what i think.


Like Eduardo said up there, try adding the


parameter when connecting the muse to get the actual brainwave readings.


It is true that activity in different frequency bands has associated with certain cognitive processes; just try searching on http://scholar.google.com to get an idea. Importantly, there is no one-to-one mapping between bands and what you are doing or thinking, and it is a complicated issue to make these associations consumable by the general public.

You can say that more alpha power means that you are calm, but you are losing so many details that may be important in determining what alpha is actually associated with. For example, increased alpha power is associated with inhibiting brain regions that aren’t involved in the current task, a decrease in alpha [and beta] occurs when you are paying attention to something or encoding a memory, theta power generally correlates with memory retrieval and encoding, etc., etc. ProTip: If you want to know about alpha specifically, look up the researcher Wolfgang Klimesch.

I actually don’t find the list included in the first post to be all that accurate compared to what I know from memory research (I’m a cognitive neuroscience PhD student), but I am sure this is an artifact of InteraXon and friends needing to simplify the technology and methods for the general public.


Thanks for the insights mollison. I’ll definitely be looking more into the natures of these waves. When i asked my question above i didn’t mean to oversimplify the matter, i was just wondering about the technicality of the numbers themselves. I suppose t can be assumed pretty easily that a higher number means more of that wave (and not, for some reason, the other way around or anything) and i’ll just keep averaging them for now (unless taking the highest reading at any given moment or some other method, might be better?). I’ve been operating under the assumption that the range of these numbers is 0 - 1. Confirmation?

In any case I’ll have to be doing a lot of investigation to determine if there are any reproducible patterns generated during certain activities. I’m interested identifying when certain brain states are reached while playing an instrument.


@to_the_sun and eduardo thanks that does work if i use the OSC dump, but in muse-lab it only shows flatline.


@dammas, hhmmm sorry i can’t help you; i only tried Muse Lab a couple of times and i couldn’t get it to show anything at all. Maybe post about it in the Tech Support forum?


@to the sun. Ok thanks I will try that. @mollison thanks for the insights, could you recommend some articles or a book perhaps which cover the basics on brain waves?


hey guys. I was trying to do oscdump 5000 where I got these errors. /Applications/Muse/oscdump: line 146: /opt/local/bin/gsed: No such file or directory
/Applications/Muse/oscdump: line 150: /opt/local/bin/gsed: No such file or directory
/Applications/Muse/oscdump: error: `/Applications/Muse/.libs/oscdump’ does not exist
This script is just a wrapper for oscdump.

Anyone have a solution?


If oscdump works by itself but not when it is called from a script it probably has to do something with access rights.


Thanks for your reply. I was able to figure it out, probably was a problem around the installation of liblo.
I’m now able to get those data with oscdump 5000, but I can’t figure out how you save those data.
I was trying to do something like
oscdump 5000 | grep /muse/dsp/elements/beta | >> beta_data.txt
but wasn’t able to save anything.
Anything I’m doing wrong?


Hey Ken,

That’s almost correct. If you want to dump to a file you’ll want to try something like:

oscdump 5000 | grep /muse/dsp/elements >> elements.txt

Note >> will append to the end of an existing file, > will overwrite the file entirely.

Also note, if plotting any of these scores they range from 0 to 1, therefore to see them in Muse-Lab you will need to zoom your plots accordingly.

Additionally note that all scores are relative to the current user. Your alpha scores will reflect an increase or decrease in your personalized alpha waves. As such, this score is personalized for each user. We take a snap shot of your frequency patterns and inform you as you reach your own peaks and valleys of delpha, theta, alpha, beta and gamma.