I think maybe PACO would allow that. I’m reading the manual now 
My mom has a tremor that varies. When it’s pronounced, she can’t snap her fingers as quickly as when the tremor is dormant. Maybe you could do a quick self-assessment to see if you experience the same thing? Just start a 10 second timer and see how many snaps you can do when the tremor is present vs dormant? If there’s no difference, perhaps just experiment with different more complex motions, like playing “jacks” (bounce a ball, pick up a small object, catch the ball) or threading/re-threading a needle or whatever? All the best, Dean
1 Like
These are very good ideas. I’ll try a few different things and report back.
A quick search dug up this open source app. Don’t know if it still works.
Otherwise, there are quite a few “seismometer” apps that will tell you how bad your tremor is on the Richter scale…
Hi Gary,
I played around with tracking hand movement by attaching a phone holster to a glove. There are STEM apps like Physics Toolbox Sensor Suite that will capture phone movement data that you could look at to see if there is useful information coming off it. You could probably just rest the phone on top of the back of your hand.
I don’t know if they could compile it for the apple watch.
Good luck.
1 Like
I’m just back from vacation and catching up on the forum - @OP_Engr thank you for pointing me to the Physics Toolbox Sensor Suite. I think this is just what I needed. I downloaded it and tested it just now. The G-Force meter gives a very clear picture of the tremor.
My method of holding the phone is pictured below. My goal is to perform an action that is:
- Likely to induce the tremor. In this case, deliberately attempting to hold my left thumb steady always induces the tremor.
- Is easy to do the same way all the time. As this picture shows, this way of holding the phone is pretty easy and repeatable.
It took a bit of practice to figure out how to hold it, but I think I’ve got it figured out so that it will comparable from measurement to measurement.
Here is a screen capture that shows my tremor:
I just had a friend try it - she doesn’t have a tremor. She held the phone in the same way and this is the picture. Clearly different.
The app also provides a simple csv export - hooray! What I’d like to do now is convert this measurement into a score I can use for my left and right hand. I’m thinking that this score will be my dependent variable. I’m especially interested in how the tremor varies over the course of a day. I have a sense that it varies a lot, but I’m really not certain about the pattern or cause.
1 Like
I’m thinking about how to construct my “tremor score.” Ideas appreciated if anybody can help. You can see the total g-Force bouncing up and down, and there ought to be a simple way to tell the height of each bounce, and then to take the average height. Since the tremor is pretty steady over short periods, the total measurement time shouldn’t matter very much, which is good for keeping my self-test protocol simple. (“Do it for a while until the tremor is visible on the graph.”) But I don’t know what formula to use, or if I have to take the sampling rate into account. I’m over my head. Advice?
Here is an example of the data that comes from the sensor:
Here is the csv file: tremor data.csv (6.7 KB)
An idea that didn’t work:
My concern is solely with the strength of the tremor, so how about if I just measured the amount of g-force exerted in a set amount of time while holding my phone in the measurement position (as still as possible)? A worse tremor should exert more force. Yes, but although I can hold the phone still enough that the tremor pattern is easily visible in the graph, there are still some jumps from small movements of my wrist or arm, and these jumps are too big not to mess up a measurement based on total g-force alone. I’m going to need a measurement of the rhythm.
to measure “rhythm” you might consider the fast fourier transform (FFT), which can convert your data from the time domain to the frequency domain, and give you the intensity of frequencies in the data (so if your tremor is at a frequency of 3hz, you could see the intensity of that frequency in the output, vs. the intensity of other frequencies in the data)
it would also potentially allow you to measure changes in the frequency of the tremor (but only larger changes, I think, because the FFT works by putting frequencies into bins of a certain size)
1 Like
Thanks @Beau_Gunderson - I’m learning as a go, here and that link is useful. I’ve been thinking that it’s just the amplitude that’s important (even though I said “rhythm” in my past post). It’s possible that the frequency changes too; I don’t think so but then again I’m not sure I would notice if the change is small.
an FFT would also give you intensity of a given frequency, and if the frequency doesn’t change then you’d know which frequency to look at to measure the intensity
@Agaricus do you want to share a CSV export? it should be very easy to see whether an FFT approach will yield a useful result
Thanks Beau for thinking about this. Here are a couple recording made with my left hand. I changed my measurement method to produce a bigger wiggle in the second recording. (More of the phone’s weight on my thumb.)
Tremor-aug28-11:58-am.csv (127.2 KB)
Tremor-aug28-11:56-am.csv (170.5 KB)
In the second file there is a pronounced peak around ~5hz, here I used the gFTotal column which is a sum of g-force over all the axes:
If I instead use just the Z axis you see this, a dominant peak around ~6.5hz in the first file:
I think using the gFTotal column is probably the best way to measure as the hand likely moves in multiple axes while holding the phone no matter the orientation and no matter how carefully held.
Do you have additional recordings @Agaricus? The longer the better!
2 Likes
Hi Agaricus, Is the tremor in a specific axis that could be associated with a specific muscle group in your arm? I am wondering if you could record the muscle signals and get useful data from that.
This is super useful. Thank you. I will record some longer files in the next few days. Not sure about the muscle groups but I will try to notice.
Left hand, 8:15 am.
sensor.csv (574.9 KB)
Right hand, 8:20 am
sensor-1.csv (387.6 KB)
(Right hand usually better, subjectively.)
These look similar in terms of the frequency and differ primarily in terms of magnitude:
I wonder if binning the data would make the magnitudes closer, though, since the right hand data seems to have a wider range of frequencies centered around 6.1hz and the left has a more defined peak at 6.5hz. We’re also approaching the limits of my knowledge about FFTs, and so I can’t definitively tell you whether the magnitudes I’ve calculated are comparable between the two files… It does seem like this approach allows for detecting both the frequency and the magnitude of the tremor, though, and that the frequency detected is within the range @ejain pointed out.
re: recording muscle signals, something like this looks promising: https://www.sparkfun.com/products/13723
and I actually have one of these somewhere, if the drivers still work I’d be happy for @Agaricus to have it! https://www.extremetech.com/extreme/149335-the-myo-gesture-control-armband-senses-your-muscles-movements