Comment by wavesplash | Hacker News Reader

Comment by wavesplash | Hacker News Reader

Having built a pulsed vibration wristband for someone with Parkinson's, just a few notes. For Parkinson's patients:

  - there's a small-ish window in the disease progression where pulsed vibration is able to help a person regain some intentional motor control
  - the device doesn't stop the tremors
  - the pulsed vibration does let the brain's control signal come through stronger, so drawing/typing/moving the mouse is more intentional.  
  - this is no more effective than the results of levidopa/carbidopa treatment, and the levidopa/carbidopa typically calms the overall tremors
  - as Parkinson's progresses, these devices become useless, and if the patient qualifies for deep brain stimulation, it's worth looking into (regain years of useful life if you're lucky).

I haven't tested the device on people with essential tremor, so I don't have anything smart to say there.

You can make a very unsophisticated version of this without a lot of hardware:

  - small microcontroller with 3v outputs (XIAO's are perfect)
  - 3v sealed eccentric mass motors (aliexpress $1/ea)
  - elastic wristband
  - usb battery
  - software to pulse the motors for about 200 miliseconds for an on off sequence of about 8 times, then rest for about a second and repeat.

Hopefully these folks will price it affordably.

Side note: the idea was published in 2016-2017 in a few promotional videos from Microsoft research: https://www.windowscentral.com/microsofts-emma-watch-and-pro...

This is a super cool project.

In general I would not recommend driving a motor (even a small one) directly with a microcontroller GPIO pin. Those pins aren't designed to sink much current and you'll at best get a very weak response, because although the MCU is 3V that's only going to be maintained with negligible load on the pin. At worst flyback may destroy your microcontroller because the motor is an inductor and will cause a reverse voltage spike when turned off.

A slightly more robust approach that costs just a few pennies more is to use a BJT (i.e. an NPN transistor) to drive the motor and add a flyback diode (a reverse biased diode across the motor's windings) and a little .1uf capacitor.

You could also use a FET to drive it. Just be aware most common ones don't fully turn on at 3V gate voltage and will still present a significant resistance in the circuit which is why I would probably use a BJT.

I completely agree. if you're going to produce these, do real power management, . However, if you're doing this as one off experiment, driving one tiny eccentric motor at a time for short pulses via a XIAO (ESP32) or RP2040/2350 was fine (we had over a year of use without an issue). YMMV.

I actually built a wearable that used these motors a long time ago, which is why I was particularly interested. What i found is that connecting the 3.3V rail directly to the motor gave a much easier-to-feel vibration than the GPIO pin, which is what caused me to learn about their really low current limits. However the current sunk also varies by microcontroller.

Interesting, it’s not a closed loop with the tremors nor the bioelectrical signal?

I’m curious to know why they had to “choose a disease” for the market tests. some treatments are symptomatic and don’t need to select the exact disease they are going to market for.

No closed loop needed. Apparently the brain/nerve circuit just needs a regular input stimulus to let more of output signal get through. Someone in neurobiology may have better answers for the why.

Odd, I saw a video of a pianist where the device did stop the tremors in their hands so they could play.

Any experience in applying the device to tremors caused by other issues?

this person above i think is saying over time as the disease progresses(gets worse) , it stops working as well ... which is sad... but at least maybe for folks in this stage of the disease there's a way to restore some amount of normal.

Reply on news.ycombinator.com