This Blogpost extends the short post on BioHakDag #1, with BioHakDag #2 and #3 (Feb. 2019 & 2020) See also https://pong.hku.nl/blog/BioHackDay/ for an impression of earlier BioHakDays.
BioHakDag : “I have some good news and some bad news”.
The 3d BioHakDay took place on 12 February 2020 at HKU IBB. It was great to experience so much interest in the topic of BioSensing amongst participants from Media, Interaction Design, Fine Art, Music & Technology, ECT and some non-HKU external participants Mimeto (MBO).
After a short and basic introduction to senses, sensors, biosensing & biohacking we made an assessment of the hardware present and got started. After the lunch break we had a short inspiration presentation from the people behind https://www.goldenemotionanalytics.com
When we talk about BioSensors one can think of sensors with which the heart rhythm, muscle tension, skin humidity, breathing, movement and brain activity can be measured. We explore how different sensors for biofeedback work and how they can be used to measure the physical change as a result of a variable experience. This type of data enables designing for specific user experiences using sensors for implicit feedback (user centered design). We can thus focus on “the inner workings of the user”: what does the user experience? Think for example of a game that becomes more difficult, more exciting, or easier when the player’s heart rate increases, a meditation application, an installation that you can influence with your brain activity, or an application that gives you feedback based on the muscle tension in your arms.
In addition to measuring muscle tension, a special experience in BioHakDag # 2 was to control muscle tension through Electro Muscle Stimulation (EMS). With an electrical current in the right place, a muscle pulls together from the user. So you can, for example, let someone play the piano ; )
Examples of biosensors can be found on the websites from: Heartlive , MYO , MindWave, Bitalino A number of the sensors are available in the HKU werkplaats ‘uitleen’, but participants were also invited to bring their own sensors for further exploration.
“First the Bad News”
Although some of the hardware is very easy to use (think of a blood pressure meter and the HeartLive with the included Windows software), getting most of the hardware up-and-running (anew, for the first time or ‘again’ after a year) is quite time consuming.
An important find in these BioHackDays is that the time required to find out how to get ‘it’ working, or what seems to be a problem (debugging), etc. was quite a challenge and demotivating.
Sometimes this is depending on the technical background of the person, but sometimes this has also to do with deprecated software, drivers, support and sometimes faulty hardware.
Some of these last finds include the Mindwave (even the new one, Mindwave 2), the difficulty to get Bitalino running, getting the relevant Myo data (via Myo Sender), HeartLives etc.
Some of the HeartLive sensors easily break at the joint of the ear clip, are very ‘defect’ and repaired with rubber bands & hair ties. Besides that, the newer versions of Unity do not support UnityScript anymore. The updated script seems to have some audio filtering, making the raw output of the Heartlive sensor not able to visualise. This is something to further elaborate on in another BioHakDag.)
Another challenge is to get meaningful data out of the BioSensors. Data in many devices is ‘closed’, so that it requires some ‘hacking’ to get (some) of the data for creative purposes. Many devices (such as MindWave, Muse, HeartLive, Myo, etc.) offer their own applications, but do not offer any connectability to other applications.
Bio data can be ‘noisy’, varying to different contexts and users. Some of the hardware has built-in calibrations and some ‘smoothening’, like the Bitalino en Heartlive.
“Then The Good News”
The basic concept of a quantified self, as well as the opportunities that BioSensing offers for user centred (interaction) design are very inspiring. The students that participated in the last couple of BioHakDays were highly motivated to realise some of their work with biosensing. Some interesting end results are discussed in this blogpost and include the work from Jasmina Avellanedas and Guillem Góngora (2019) https://gaudeamus.nl/en/events/senses-working-overtime-2019-day-1/hku-students-programme/. In All the better to see, participants can hear their breathing and heartbeat and influence how sound is experienced in space, as if the environment becomes an extension of the body.
The domains of BioArt, BioSensing and BioHacking are in an enormous lift and the number of (commercial) applications is growing every moment. It was a welcome, inspiring, presentation by Golden Emotion Analytics, a new startup that has the ambition to have some of the ‘quantifiable biodata’ connected to optimizing user experiences involving media and entertainment.
Another interesting find was the commercial applications around “Plant Music” → Plant Sensing, with sound / music as a result.
One of the student groups started working with a concept of plant personalities and some ‘presence sensing’ required to give auditive feedback to a visitor on whether to approach (or not) the plant. This early prototype of ‘plant touching’ was realised with the Bare Conductive Touch Board.
It’s especially nice when a concept is realised in the limited time of one ‘BioHakDag’. The realisation of ‘breathing the life into a digital Pinguin’ with an E-health sensor and PinGui is especially noteworthy: The documentation of E-health sensor and PinGui can be found on: http://laczkojuli.net/biohackdag-hku-ibb-feb-2020/
What we explored / worked on / with in BioHakdag #2 and #3:
As a potential ‘presence detector’, ‘pose detection’, but also heartrates the Kinect is a promising sensor. It’s good to mention that Kinect won’t work on OS X Mojave, because openNI is not supported. For pc use NiMate to ‘translate’ skeleton data to OSC. “Beware: use Microsoft drivers, not the one NiMate wants to install” Check for more information : https://trello.com/c/s0eVb1zk
Tobii Eye tracker
(no updates, only works on Windows), see https://github.com/hku-ect/BioData-Interfaces One group of students realised an auditive memory game with visual control using the Tobii:
Bose AR ready headphones : Frames
A new development we see for BioSensing is the simple implementation of gyroscope data in Bluetooth devices, such as headphones (Bose AR ready headphones). A special mention and worth exploring is the Bose Frames (sunglasses with built in miniature speakers and ‘AR-ready’). To start experimenting with these frames : https://developer.bose.com (or ask author Than van Nispen for the Unity demos) Step 1: Firmware updates (required) Bose Update application (works with Google Chrome) Bose Frames werkt via USB Audio Support ook in Unity op Mac. Echter… is een USB kabel benodigd. Builden naar een mobile device kan voor beide apparaten (en dan dus wel handsfree). USB Audio Support (BOSE NC 700 HP (doesn’t support USB Provider) If your Bose AR device supports USB Audio, the device should enumerate as an audio output device on your operating system shortly after connecting to the device at runtime. Currently, the following Bose AR devices support USB Audio: • Bose Frames To hear audio through your supported device: Connect to your device via the USB Provider Go to your system audio settings and select your Bose AR device as the default audio output. Next level: (including Bose Frames for Max/MSP 8 ea) chrome://flags/#enable-experimental-web-platform-features Chrome for Desktop [PREFERRED]: enable Web Bluetooth by going to chrome://flags/#enable-experimental-web-platform-features and check “Experimental Web Platform features” http://localhost:8888/bose-ar-explorations/
Heartlive & Unity
The HeartLive is a simple way to sense ones bloodflow via photophletysmography, using an infrared earclip. The clips can be bought for < € 10,-, but the HeartLive costs around € 100,-. The HeartLive offers callibrating hardware, which presents as a soundcard and a windows application. As mentioned some of the HeartLive sensors easily break at the joint of the earclip, are very ‘defect’ and repaired with rubber bands & hair ties. Besides that, the newer versions of Unity do not support UnityScript anymore. The updated script seems to have some audio filtering, making the raw output of the Heartlive sensor not able to visualise. This is something to further elaborate on in another BioHakDag.)
Microphone input does work Heartlive does not give any form of response (but the data can be seen in Audition). Sander: “I think some filtering or buffer size cancels the signal. To be continued…
More on HeartLive on the ECT-github repo
Although debatable whether this device is a ‘biosensor’ the MakeyMakey is a very simple usb interface that literally “Connects the world to your computer” (-as is the company’s slogan), by offering 6 inputs to connect ‘conducting things’ to, like bananas, tin foil, humans, or anything else in the world. . . Makey Makey work just like a USB keyboard or mouse, sending keyboard (WASD, spacebar, etc) and mouse (left click, right click, etc) signals to your computer.
Makey Makey Classic works through opening and closing circuits, just like any other button. Instead of the circuit being closed underneath your keyboard, the circuit is closed through the conductive objects you connect with alligator clips like your hand or your lunch or some tinfoil. When the circuit is closed, the Makey Makey sends a command to your computer, just like a button pressed on a keyboard.
(Relatively new is the “Makey Makey GO” which is much smaller and works through capacitance).
For rapid prototyping and for instance to recognize a person ‘touching’ an object, the MakeyMakey is considered as a relevant device in BioSensing.
There are many variations based on the MakeyMakey principle, but a special mention goes to the Touch Me ‘sensor’
One group used the BareConductive for a ‘Plant communication’-project and modified the Bare Conductive, so they could use the output as midi. You can either solder, or use electronic paint on the midi-connection. See : https://www.bareconductive.com/make/on-board-midi-mode/ for more information. The group was able to use two different plants as capacitive sensors, however the values were quite noisy. The detection of plant touched / not touched was very clear, but the amount of touch, or place of touching not accurate enough.
http://diagnostics.myo.com ECT previously developed the MYO OSC sender https://github.com/hku-ect/Myo to make use of the MYO data possible, via sending it over the network as OSC. The application works fine in Mojave. A new feature request for this MYO Sender was to have myo/emg data in OSC (as can be seen in https://github.com/Sindel/myo-osc ). The 8 EMG-data pods are relevant for using the muscle tension and is an important addition to the gyroscope data and pre calibrated gestures. Another question -and feature request- is whether the updated application could send ‘pings’ to the Myo (can you have the MYO trill on command?) These features will be implemented later this year.
MindWave 1 & 2 (new)
Niet aan de praat gekregen. Documentatie doet vermoeden dat ‘het’ moet werken, maar geen installaties mogelijk onder 10.14.6 . Advies contact opnemen met bedrijf voor support.
The Muse headband was under reparation during BioHakDag #3
Bitalino presents itself as a “personal biomedical research system” and “an all-in-one hardware design, with all the blocks pre-connected between them and ready-to-use, making it perfect for biosignal exploration and lab activities. The kit includes all the basic accessories needed to get started, namely the hardware modules, battery, cables and electrodes. Along with our cross-platform OpenSignals software, it enables instant biosignal data visualization and recording out-of-the-box.”
Some of the sensors included in the Bitalino kit we used are Electromyography (EMG), Electrocardiography (ECG), Electrodermal Activity (EDA), Electroencephalography (EEG), Accelerometer (ACC) From the BioHakDag #1 : A Python script to get in the Bitalino data, and send it over OSC, tested on Mac OS Sierra, and with Python 3.7 and libraries mentioned here: https://github.com/BITalinoWorld/revolution-python-api There is also a Bitalino Max/MSP object available
As many students seem to have difficulty in getting the Bitalino up-and-running, one of the R&D-colleagues of ECT experimented with the Bitalino. It is important to follow the online instructions, step-by-step. Also check if the sensor is connected to the correct input (beware : the information of the inputs is on the underside of the board, while it seems more logically to only look on the topside of the board). As Job said : “We zijn geneigd om vanaf boven te kijken en simpelweg te tellen wb poorten” Furthermore : “Data in tekstbestanden verdient geen schoonheidsprijs.” : The data is primarily updated in a text file. This is not very practical for realtime interactive applications. Bitalino offers an Arduino SDK, which should make more interactivity possible (without having to work with updated text-files). It’s also good to mention that the Bitalino hardware gives good and coherent results → apparently there is some Auto-calibration in the hardware. Is you connect the sensors directly to an Arduino this results in very noisy data, practically unusable, so use the Bitalino hardware with the sensors! Furthermore : ignore the urge to break the board in separate components (it really looks very inviting to do so), because this will require to make all the connections anew….
Kodea KD-202F Blood Pressure monitor
Dit is een Blood pressure Monitor waarover weinig te vinden is online. Echter na wat snuffelen blijkt dat de monitor hoort bij de e-health sensor platform van cooking hack. Main tutorial page of the e-health shield Error in e-health arduino library
Het lijkt alsof je met de bloodpressure sensor alleen de gegevens in het geheugen kan uitlezen. Dus je kan alleen meetingen die al gedaan zijn uitlezen. Je verbindt hiervoor de hartslagmeter met de e-health shield die je op Arduino zet. Alleen als je nu in de Seriele monitor kijkt zou je de gegevens moeten kunnen binnen halen, maar dat werkt helaas niet…
Electro Muscle Stimulation (EMS)
student Filipe: “voordeel van EMS vs. trill motors is dat het een zeer specifieke sensatie is” Hans: “unpleasant feeling, too intrusive” Tested on arm, thumb. Experimented with size and form: too small form feels more ‘painful’, original (square) size seems best. More literature : Pedro Lopes EMS
Evaluation by Hans Leeuw: “Electro Muscle Stimulation (EMS) is an interesting way of implementing a haptic experience. The method is much more intrusive though than haptic feedback through vibration motors or vibrating piezo elements.” An email was send to Pedro Lopes by Hans, but there was no real answer to the issue of intrusiveness. The questions to Pedro remain open for now and is interesting to share:
“Finally had some time to look into muscle stimulation today. I did some experimenting together with students who also liked it. I have a number of questions though. I hope I do not ask questions that are answered in a technical article that I am not aware of…: We tried EMS on the thumb and the wrist. The electrodes that come with the TENS/EMS seem a bit large for the thumb. We used some scissors to cut them smaller but then did not really like the feeling. Are there special electrodes for fingers/thumb? Did you do any experimenting with fingers/thumb? I saw some gloves online for EMS. Is that something you have experience with? Overall I (and other people trying it) did not really like the feeling. Is that something you get used to over time? it was already better using short bursts on lower amplitude (I do not yet have an openEMSstim but we faked it by switching the battery of the TENS/EMS on and off). Despite the slight unpleasant feeling I can really see the benefits of EMS for my system and like to continue exploring. I found that I have a slightly different TENS/EMS version then you use in the video. Mine only goes to 100 Hz. I guess that should suffice? Controlling the EMS frequency over time is not something worth while to address I guess? I found this schematic for building the openEMSstim: https://hackaday.io/project/106571-neurocuddl/log/151521-electrical-stimulation-based-haptics Most components are SMD. I guess there are PCB prints somewhere? If not, did somebody already look into DIP alternatives? (Especially considering the Galvanic Isolation) Also considering the stim, if I look at the schematic, am I right that this is only for one channel? What I do not really understand is why there is both on/off (relais) and a Mosfet driver with mosfets. Can the on/off signal not simply be driving the mosfets high and low? Or is that too slow? And related to that, has there been experimentation with the LFO envelope over the EMS signal and what that does to response and possibly ‘pleasantness’ of the feeling?”