Inspiration comes from many different places.
For one young innovator it came from a character in a popular animated movie – an automated robot that could monitor and immediately diagnose medical conditions. This curious source of inspiration has now led to an award-winning piece of research involving artificial intelligence, machine learning, and a mobile app that can differentiate between people with and without Parkinson’s.
In today’s post, we will discuss this interesting unpublished research from an inspiring individual, who is trying to help us better diagnose and monitor Parkinson’s.
Have you ever watched the movie ‘Big Hero 6‘?
It is the story of a boy named Hiro who goes on an adventure with a robot called Baymax.
Baymax is a personal healthcare companion that is designed to diagnose and treat medical conditions instantly.
After watching the movie Big Hero 6, Shreya Ramesh became fascinated with the idea of the character Baymax. She began wondering how a machine could be made to be smart enough to analyse the medical conditions, make a diagnosis, and then offer remedies.
So she began reading a great deal about machine learning and artificial intelligence technologies. Then she collected a large data set of information from people with and without Parkinson’s for analysis.
Sounds interesting. Then what did she do?
Next, she designed, developed, and tested a smartphone application (using Python scripts) that could potentially one day help with early diagnosis of Parkinson’s.
And Shreya presented her research at the Intel International Science and Engineering Fair, and she is now seeking to write up and publish her results in a scientific journal.
Wow. That’s really impressive!
Yeah. And she did all of this while still going to all her classes in high school.
Oh, did I forget to mention that she’s just a high school student?
Shreya’s story begins after watching the movie ‘Big Hero 6‘. She decided that she wanted to explore the idea of a robot that could analyse, diagnose and treat medical conditions.
In order to be able to do this, she decided to focus her attention on machine learning.
What is machine learning?
Machine Learning is a branch of computer science that focuses on training machines by feeding them lots of data and asking it to understand the patterns and variations. For example, by feeding the computer different images, it can quickly learn to differentiate between a Chihuahua and a blueberry muffin, or a mop and a Komondor (Hungarian sheepdog):
I’m a Chihuahua & blueberry muffin kind of guy. Source: Microage
Shreya wanted to apply machine learning in the medical field to enable people to get the help they need right away, ideally from the comfort of their own home. She wanted something along the lines of the Big Hero 6 Baymax robot.
As she did more research, Shreya came across Parkinson’s and she noted a lack of tools allowing for the early diagnosis and monitoring of the condition.
And these are REALLY important issues.
As research progresses with regards to disease-halting and neuroprotective therapies (Click here for a SoPD post reviewing this topic), there will be an urgent need for methods that allow for the early detection of Parkinson’s. The earlier we can detect Parkinson’s, the sooner the experimental corrective/preventative strategies can be applied. And by stopping/slowing the condition at its earliest stages, affected individuals will be able to enjoy a better quality of life long term.
So Shreya decided she would apply machine learning to help with the early detection of Parkinson’s.
But what tests did she use for early detection?
Shreya decided upon the Spiral Test and Voice Tests (Telemonitoring) as two tests that could potentially be used by machine learning to determine Parkinson’s.
The spiral test involves individuals drawing an inwardly turning spiral (this test has recently been proposed by Australian researchers to be a very useful determinant of Parkinson’s – Click here to read more about that study). People with Parkinson’s usually draw a smaller spiral than controls (and people with essential tremor – a different kind of Parkinsonism) – as displayed in the image below:
Voice Tests (Telemonitoring) are another method by which machine learning could be used to diagnose and monitor Parkinson’s. Audio files can be easily collected from people with and without Parkinson’s (no needles or biopsies required), and analysis of those file can isolate particular patterns that may be disease – or even stage – specific to Parkinson’s. And it should be noted that Shreya is on to a good idea here as other researchers have also attempted to use voice tests in the past to try and monitor Parkinson’s (Click here for an example of this).
But, hang on a second, if she is in high school, where did Shreya get spiral tests and audio files from?
There are actually many sources of raw data available. And for her analysis, Shreya used the files provided by this study:
Title: Collection and analysis of a Parkinson speech dataset with multiple types of sound recordings.
Authors: Sakar BE, Isenkul ME, Sakar CO, Sertbas A, Gurgen F, Delil S, Apaydin H, Kursun O.
Journal: IEEE J Biomed Health Inform. 2013 Jul;17(4):828-34.
In this study, Turkish researchers collected data from 20 people with Parkinson’s (6 females & 14 males) and 20 healthy individuals (10 females & 10 males). After conducting their own analysis and publishing their results, the researchers made the raw data available for others to analyse (those files are available via the UCI Machine Learning Repository (Click here to find the files).
Shreya was able to use the voice files and Spiral Test images from people with and without Parkinson’s to measure both the non-motor skills and the motor skills, respectively. She then took these different training sets and developed two computer algorithms – one for each type of test. And these algorithms were able to train her computer to differentiate between people with and without Parkinson’s on both tests (with high accuracy).
Now all of this is impressive enough for a high school student, but Shreya took the research a step further: she developed a mobile app for smart phones.
Using her algorithms, she designed a smart phone app that could allow people to do both the spiral test and a voice test and get the results back immediately. And such a app could be used to monitor people over time.
Shreya has written up her research results in the form of a poster:
(Click here to see a high-def. version of the poster)
As I mentioned above, Shreya has been able to present her research at the Intel International Science and Engineering Fair, where 1800 projects are selected out of nearly seven million science projects.
And while this particular post is focused on Shreya’s amazing effort, readers really should have a look at some of the other Parkinson’s-related projects that high school students presented at the Intel International Science and Engineering Fair, such as that of Miss Erin Smith – the founder of FacePrint.
FacePrint is an app which uses facial recognition software and machine learning algorithms to detect early-stage differences in facial muscle movements in Parkinson’s. Ms Smith has now launched a study in partnership with the Michael J. Fox Foundation – Click here and here to read more about this). Or you can watch Ms Smith in action at her TED talk:
But now, back to Shreya.
She has won the Intel Excellence in Computer Science Award in the Georgia State Science Fair for her research, and she is now in the running for the Project Paradigm, a science fair for high school students where all kinds of prizes and scholarships are up for grabs for innovative young scientists.
I would highly recommend SoPD readers to help her out by voting for her project by clicking here to go to her page. The voting button is on the top right hand side of the page.
So what does it all mean?
Every so often I like to write a post about young and inspiring individuals who – under their own initiative – are making a difference in the world of Parkinson’s research. A good example of a previous post along this theme is the one for Jeremiah Pate – a university undergraduate who is running a space technology company at the same time as devising a novel therapy for Parkinson’s and doing his undergrdaduate studies at University – Click here to read that post).
Shreya Ramesh is another individual who fits into this category perfectly.
Not only has she come up with an intriguing concept, but she has acted on the idea and made it a real physical thing.
It’s really inspirational stuff!
And I can’t help but wonder, when the producers of the ‘Big Hero 6‘ movie were developing their project, do you think they ever dreamed that young innovators developing technology for Parkinson’s would be one of the results of their efforts?
EDITOR’S NOTE: The author of this blog does not know Miss Shreya Ramesh and has no association with her. He simply came across her research via her twitter account (@PD_Prediction) and thought it was very interesting. He contacted her regarding the possibility of writing a post for the SoPD website. And his endorsement of her Project Paradigm is based solely on her research and an admiration of her initiative.
The banner for today’s post was sourced from Playvilla