What would you do if you could easily connect a Muse to a RaspberryPi?


Hey guys,

My goal here is to generate a little bit of discussion on the topic of hooking a Muse to a Raspberry Pi. I’m currently developing a system application to do so and I would like to know how I can match your needs as much as possible. Here is a summary of what it consists in, let me know what you think and what kind of application you would like to develop with it.

  • Documented and easy-to-use for beginning, intermediate and advanced developers.
  • Available for all RaspPi platform.
  • Available as a friendly-user application for any unix-based system (such as Raspbian) or a high performance open-kernel, based on OpenWRT.
  • Data preprocessing built-in for timeseries, fft, and/or band power extraction.
  • Statistics, linear algebra and machine learning function libraries built-in for application level data analysis.
  • Tutorials provided in the package.
  • GPIO and LCD support through wiringPi.

Myself, I see several advantages to such a system, when compared to laptop or even smart-phone based systems. This system results in a much cheaper, but still stand alone and even wearable, platform. An easier access to GPIOs to interface with any electro-mechanical systems (we speak of an embedded systems application).

So please let me know guys, how interested are you in this kind of systems and what would you be able to achieve, if you had this in your hands.



Portable demos are a big deal for me, and the raspberry pi is an ideal middle ground between a phone and a laptop. I could still run my python based demos, and display them to a screen via HDMI without the hassle of setting up a full laptop!


Yes, indeed. I made some kind of arcade game making use of the Muse (see the transparent box, with LEDs in the picture). It’s still in development, but I’m often asked to make demos.

I bought a power supply for my RPi. When I plug it, the RPi boots in about 20-30 seconds. I have a script that launch the processes automatically. The system then waits to pair with the Muse and I’m ready to go.

In addition to the RPi handling all data processing, it is also taking care of the LEDs control signal and a solenoid which opens the door of the “safe”.


Disclosure - I am working on this project as a co-founder.

The choice of python is an interesting one because of the performance hit, although I see it’s utility. What python extensions would be of interest? Could tinypython (I think it’s called that)?

Secondly, alot of the boot time magic is because our actual image is sub 100-150mb, so usually the peripherals are the last to load. This is for pure boot effiency and we do not currently support a GUI. Now having said that, an ncurses or ASCII output alongside LEDs for example, is sufficiently supported at this moment.

The kernel is very optimized compared to the rasbian one. I spent alot of time benching memory allocators, param tweaks etc… This should also be more attractive for those wanting to use the A+ with less RAM.

As per other, ideas - what about usage as device to enhance presentations? The performer goes into a state to change slides or cause some sort of interactive such as lighting?


Would it be possible to analyse data from the muse on the rasperry pi while wearing it in my pocket. For example, I could monitor my brain activity while doing physical exercise and use the data to obtain results on how my brain reacts to effort?


The Raspberry Pi can be powered in many ways, among them a small and handy battery pack. So yes, it is possible. At this point, a smart phone could do just as well however. Where I think the RPi could bring a plus value is in the fact that if you want to lend your recording device to someone else, rent it at the gym or something like that, you might prefer the system to be so personal as a smart phone, but also cheaper.

With our platform it will be possible to dump the raw or processed data on a USB storage device connected to the RPi for post analysis, but given the easiness with which any hardware can be wired to it, I’m sure many will find clever ways of providing the user with feedback.

I’ve been asked by many people what does a RPi solution brings on top of a smart phone application and I admit this is a very pertinent question, at least from a business perspective. In the latter context, a smart phone app is called an indirect competitor to our solution.

We do not believe that a RPi system will replace the smart phone as the wearable EEG processing hardware, but we believe that it complements it in many situations where an expensive and personal hardware is not well suited. We like to think that it will be pertinent for gaming arcades, arts demonstration and general hacking.


Hey @atom2626

I have built portable, EEG controlled art installations and a wacky garage door opener etc. using the Pi v2 and running NodeJS.

Using PM2, it loads immediately on startup. The only challenge I have had is dealing with BT pairing which isn’t always reliable. The BT pairing is pretty poor on the MacBook Pro, the MS Surface and the Pi so I can only assume it is an issue with the BT radio in the Muse itself.

Running in NodeJS gives us an advantage to dealing with interoperability on the IoT front. I have been able to connect huge art installations over websockets to brain data whilst storing everything to a mongo database that is purged every 3 minutes to provide medium term analysis.

Like most things though, it’s great when it’s working. Whenever I do a major event though, I bring the MS Surface Pro 3 and the MacBook Pro. The Pi is just responsible for running visualisations of the live brain readings on the television nearby.

I’d be keen to hear if you’ve managed to get more consistent pairing with many Muse devices (we run up to 4 at the same time).



Hey Mic,

You are bringing an interesting perspective. We are looking into the development of communal art display.

So far we haven’t had any bluetooth connectivity issues, but we only paired up to two devices at the same time. Keep in mind that bluetooth connectivity is limited to a few meters in the best conditions.

I did had trouble pairing with one muse device in particular, but I still haven’t figured out why.

Which library do you use for your data processing? It is all based on the Muse SDK?

Did you had any trouble pairing 4 muses on the same bluetooth dongle? We were wondering when we should start worrying about losing packets.

Just for the heck, here’s a link to our last week-end project:

You have any link to your work?


I just want to add a technical note about Bluetooth and Bluetooth Low-Energy (BLE) from what we have observed from our experiences in spaces “crowded” with wireless devices - BT and these related technologies are using the 2.4x Ghz range which is extremely overpopulated.

If you are WiFI (and everyone else is too), chances are you are on the 2.4Ghz band (you will be lucky if there is 5Ghz or that you connected to an enabled AP), and if you are surrounded by a number of other BT devices - you will have alot of difficulties. There are a number of other factors involved such as specific lighting systems, wireless microphones, and interference such as microwaves/electric motors. I’d recommend checking your issue with a spectrum analyser from time to time - it may be that your environment is just too busy for reliability sake. (or bring a portable Faraday cage ;))

Edit: Thinking about BT on USB, BT uses roundrobin approaches to multiplex multiple connections. This means that on a good day with a very “quick” reciever, you can expect it to switch somewhere between 1.5-2ms and establish a connection within 3.5-4.5ms (this goes for Cypress and another vendor). Secondly, from the USB aspect - you may end up getting too many hardware interrupts which could make your performance suffer… this is something we all will need to pay attention to with Linux on low-end systems (bench bench and bench the schedulers for your application).


Wow @atom2626 such a similar concept to what I did with the tug-of-war, cool stuff! I’m in the middle of getting a website up but you can see my tug-of-war game here: https://www.facebook.com/thewondergamesau/timeline/story?ut=43&wstart=0&wend=1443682799&hash=-2540569583148658833&pagefilter=3


Hey Mic, I’m a bit busy today, but just to let you know that I took a quick look at your setup and yes, we’re on a similar path :wink: nice work! I just notice you are on the other side of the world. Out of curiosity, are you ever in Canada, North America?


Rather than hijack this thread for future visitors, I’ll PM you @atom2626 about my travel plans


I’m psychologist in Canada, Quebec. I wish MUSE could be connected to Robotics in order to offer multiple sensorial feedbacks to users. Like visual, touching, audio, motion, etc. Researches in psychology field demonstrate that; the more you use different sensorial feedbacks pathways, the more your brain reacts and change.

Raspberry Pi 2 platform is small, powerfull and portable, so could be used in colleges, schools, homes, offices, outdoor, sports, on-the-go, etc… and as I go to visit clients needing help about anxiety, stress, insomnia, PTSD, etc… I would use this setting for sure. So, some small robotics platform controlled by Raspberry Pi 2 and monitored by a user’s MUSE brandwaves would be a much better efficiency solution than smartphone feedback alone. Motivation would be 10 times more powerfull.

Even bigger project I’m dreaming of would be that clients would use MUSE home, connect to a special website and control in real time a specific small robot in a real room, in which room there will be many other robots controlled remotly like that by other users. All these people would be involved in a collaborative usefull-humanistic task using ‘Calmness’ as a cue signal to achieve the goal (ex: robots would create a few kind words matching their shape and colors all together, the process would be followed by a Webcam and registered sick children in hospitals around the world could follow the event through Internet). 100% more chances that every user will be highly motivated and will fully participate everyday with pleasure. It would create a 'World web connected group of people in same mood and positively-calmly energizing each other.

Great projects might come out of Raspberry Pi - MUSE pairing together.


Thank you for that input. We will be supporting the Raspberry Pi 2 shortly. I’ll PM you, I have the feeling like we have met before.