Parkinson's disease diagnosis system

Hello, there. I plan to build up a system containing multiple IMUs and an Arduino board to diagnose PD via detecting its motor symptoms. However, I am just a beginner (I used Arduino Uno only when I was in Y1), so I completely have no idea about which Arduino and sensors should I use... anyone can help?

@hl-zhang00, your topic has been moved to a more suitable location on the forum as this is not related to covid.

First step is to determine what properties you are planning on measuring to detect PD. That will lead to what sensors may work which will lead to which processor you need.

Pls define: system containing multiple IMUs.

This is a noble and worthwhile project…
I have a couple of family members with PD.

It’s a pretty steep starting point for a beginner.
My suggestion would be to either find an experienced buddy, or put it in the open-source domain to attract some stronger developers to work with you.

There’s a lot more to the final result than just reading the IMU’s, but the goal could well be worth the effort.

Good luck.

So, something like a wrist-attached device to record the small tremors for later analysis?

It would be an ambitious project even for an experienced programmer. I recommend that you do a lot of breadboard testing. Buy an Arduino, the Uno is OK for testing. Buy an IMU and start experimenting.

1 Like

Depending on which particular symptom or symptoms of Parkinson's you intend to measure will effect which IMUs and how many you will need.

I have a working prototype with similar goals for working with Essential Tremors. I started with an Uno but quickly exhausted it RAM capacity with one IMU and a fast data logger. Now using a Mega but intend to switch most likely to a Teensy 4.1 I would advice you to consider separating the analysis s/w from the data collecting s/w.

Some other problems you will need to solve:

  1. How to power your project. Tethered or free movement
  2. How to position your sensors so your measurements are repeatable with some fidelity
  3. Adjunct to 2. Quick and easy to put on and take off
  4. Adjunct to 2. The sensors rig must not interfere with natural movements
1 Like

If you want to accurately measure movement in the human arm/hand excluding fingers and thumb, your need three tri-axial IMUs as there are 3 Degrees of Freedom (Dof) of the upper arm, 2 Dof of the forearm, and 2 Dof of the hand. Adding in the forefinger and thumb, adds 2 more IMUs.

With Parkinson's besides tremors and movement of the arm, you need to look a gait. That adds 3 more IMUs for the 7 Dof of the leg/foot. Are you going to measure both left and right? You now need 16 IMUs and you should probably add 1 more for head movement. For my own peace of mind, I would add one more on either the saddle of the hips or my preference is to add one on the beast bone and measure everything relative to it.

Even though the measurement and recording of data at 100hz is a challenge, analysis of the data may be a order of magnitude greater problem IMOSHO

1 Like

Thanks, what I was really looking for was; " inertial measurement units (IMU )"

Are these sensors actually acceleration sensors?

Haven't come across that one before! :rofl:



what is an imu sensor

Tom... :smiley: :+1: :coffee: :australia:

There are accelerometers and then there are accelerometers. If all your are interested in measuring is Parkinson's 'resting tremors' you might be able to get by with a $5 3-axis accelerometer. On the other hand, if you are going to 'deep dive' into diagnosis of Parkinson's and for example test "tandem gait', which should include a measurement of posture, your going to need some a bit more sophisticated than a single accelerometer.

Now we will return to our normal programming..."Decide on what aspects of Parkinson's you are going to measure and make your design decisions based on that".

Here is a helpful reference: Home - PMC - NCBI

Caveat emptor! You can be buried in information from this source. A search for 'parkinson's diagnosis' yields 100686 articles. Far too many for one person to digest. I suggest you time limit your searches to the last 10 years. You will find others that have use 'modern technology' to diagnose PD.

Good luck


The first time I googled IMU I got a ton of hit for Don Imus :slight_smile:

Hi, I was planning to build up a device to collect data only, any more advice?

I am just a Y3 student, so it's gonna be still great if I can design a device positioned on wrist (like apple watch) to collect data from PD patients for later analysis.
Any help?

To only measure tremors in the hand/wrist, I used a
Adafruit 4485 Adafruit LSM6DS33 + LIS3MDL - 9 DoF IMU with Accel/Gyro/Mag - Stemma QT Qwiic
mounted on a Velcro strip.
I recorded the data using a ELEGOO MEGA R3 Board ATmega 2560 and a Seeed SD card shield.
The software I used was exFat SD data logger and the Adafruit library for the sensor.

This will not run on an Uno as there is not sufficient ram.

Thanks for telling me that, I plan to use an Arduino board that can use Bluetooth to transfer data