We've gotten a flood of offers for help since this story took off. Brian is now in touch with a programmer and together they're capturing the information he needs.
If you still have an appetite to help Brian and Melissa, this Facebook page will tell you how.
Before the tumor things were going well for Brian Lande. He had a PhD in Sociology from UC Berkeley and was working as a research faculty member at UC Santa Cruz. He had been married seven months to his wife Melissa, who was pregnant. Life was good.
When Melissa was 26 weeks along, the unthinkable happened; complaints about migraines revealed a brain tumor, leading to a seizure that left her in a coma. Three months later she was rubbing her belly, still comatose, and safely delivered a healthy boy without ever waking up. It’s a year later now and she is very slowly emerging from her coma, just minimally conscious today.
A researcher himself, Brian was used to approaching problems creatively, from first principles. He started looking into how doctors communicate with their comatose patients, wanting to do everything he could to help his wife communicate with their newborn child.
What he discovered was surprising. Physicians use a qualitative scale for classifying the movements of coma patients: they observe how they move, apply certain criteria to determine if the movement was intentional or not, and disregard motions that don’t register as deliberate on the scale.
Let’s say her thumb just moved: was it an involuntary twitch or a partial thumbs up? The best guess prevails, and it changes from physician to physician, therapist to therapist.
Dissatisfied with this low-data approach, Brian came up with a plan. He spent years as a program manager at DARPA, where he learned about the power of wearable sensors by using them to model social interactions. He realized that a wearable accelerometer would help him model Melissa’s movements over time, and see if she had some kind of routine or activity schedule. He thought maybe a Fitbit would do the trick, but he couldn’t get access to the raw data and the device wasn’t sensitive enough to be useful. He needed a different kind of wearable sensor.
Enter the Myo armband. Myo’s nine-axis IMU was great for tracking Melissa’s daily movements, but the data from Myo’s EMG sensors is what Brian finds truly exciting. EMG sensors directly sense muscle activity; even better, they can read the intensity of that muscle activity in a very refined way.
Brian’s hypothesis is that this information is the key to determining whether a hand motion is a non-purposeful reflex or a deliberate hand gesture. He even published a paper with Liz Torres (Rutgers University) on the topic before he discovered Myo. Now that he’s got one, he needs a technical mind to help unlock its potential.
What Brian needs now isn’t complicated: a small script or program that will store the raw data from Myo as an HDF5 file. That’s it. If he can make that happen, he can do a powerful time-series analysis on the data that will give him deep information about Melissa’s movement and muscle activation.
What he doesn’t have right now -- with his work, his 11-month-old, and a wife requiring constant care -- is time.
So we’re reaching out to you, our awesome dev community -- people who have shown us that they want to improve the world with technology. We hope there’s someone out there who can help.
If you’re able to produce a script or program like this, that will capture the information Brian needs in the format he needs it, we’d be deeply grateful. Just write to firstname.lastname@example.org and we’ll put you in touch.