MobiTest - Parkinson's Monitor

MobiTest is a multifunctional Parkinson's Disease monitor which can track motor functionality using tests and machine learning, and provides a custom keyboard for communicating non-verbally.

Comment

Inspiration

I was inspired by my personal experiences with Parkinson's Disease in my family, so I decided to implement a mobile app that could help in measuring the progression of Parkinson's Disease, and to provide tools that would help those who are suffering from it to communicate with their caretakers. I also took inspiration from this study which introduced potential at-home diagnosis of Parkinson's to build my app.

What it does

MobiTest is a multifunctional Parkinson's Disease monitor which can track motor functionality, and provide a custom keyboard for communication with Parkinson's patients with a speech disability.

This app will use three tests to gauge the severity of Parkinson's Disease in a patient. In the Static Spiral Test, the user must trace a spiral on the screen within 6 seconds, keeping the finger as stable as possible. In the Dynamic Spiral Test, the user must trace another spiral on the screen within 6 seconds; however, the spiral will flash on and off to test the user's short-term memory of the location of the spiral. Lastly, in the Stability Test, the user must press on a specific part of the screen without moving their finger for 12 seconds, all while performing mental arithmetic, which could affect motor functionality.

The user also has access to the charts showing the history of their test scores. The Statistics page stores the results from all of the 3 tests and can display the history and trends of your results over time.

The last feature, also unique to MobiTest, is an Accessible Keyboard. This keyboard drastically improves communication with someone who has speech impediments, a common yet overlooked result of Parkinson's Disease. This keyboard is unique because it refers to a premade dictionary of words with the selection of words narrowed down by a selection of the first letter.

The Suggestions page is very unique as well, because it doesn't just keep a list of the most tapped words and display the top 12, it instead uses a Natural Language Processing model called Naive Bayes. Each word in the dictionary has its own list of most probable words that should follow, which in turn results in those words being displayed upon selection of the first word. This model is updated in real-time as more words are selected from either the Accessible Keyboard or from the Suggestions tab.

How I built it

I built Mobitest using the Flutter framework that uses the Dart language. I chose Flutter because Flutter can be compiled to both Android and iOS. I used several flutter packages for drawing, chart-making, and for using shared preferences to store data locally on the phone.

As a sidenote, the Android APK is too large for me to upload, so I can only link the Github repo.

Accomplishments that I'm proud of

I am proud to be able to implement a fully-fledged application that can actually help those who are suffering from Parkinson's Disease or any other neurodegenerative disorder. I am also proud that I have used the advanced level of statistics that I learned in high school and machine learning that actually improves the app in a significant way.

What's next for MobiTest - Parkinson's Monitor

I am already working with a practicing Neurologist who is helping me make the 3 tests more accurate and useful; for example, the Neurologist recommended the mental arithmetic part for the stability test. I am also working with other doctors to get MobiTest into the hands of patients who are suffering from Parkinson's Disease to get feedback from the end-users who would be benefiting from this app. My goal is to make this app available to hospitals and their patients with neurodegenerative disorders and to conduct research around how to detect Parkinson's Disease in early stages.

Share this project:

Updates