Mobile Application to Measure and Monitor Parkinson's Symptoms

Summary

Parkinson’s disease, a degenerative neurological disorder, has a main symptom of hand tremors. Currently, there is no standard diagnostic test for Parkinson’s. Doctors must monitor patient's symptoms over time, forcing the need for frequent patient visits for evaluation. To increase efficiency and accuracy of symptom monitoring, I have developed a smartphone application that tests, scores, graphs, stores, and shares tremor reports. The application uses the smartphone’s built-in, tri-axial accelerometer sensor, Google Fusion database, and the phone’s email app. This allows for an efficient quantitative measurement and analysis of Parkinson’s disease. The application successfully measured intensity and frequency of tremors. The recorded graph data clearly shows a significant difference between healthy (static) and abnormal (simulated tremors) conditions. The tremor history is recorded and saved in a database which allows doctors to keep a historical and comparative account of their patients’ condition before, during, and after treatment. This may allow the doctors to vary the patient’s treatment regimen based on the measured results and thus help in the early diagnosis and better prognosis, leading to potentially improved medical outcomes.

Question / Proposal

My goal is to develop a smartphone application to assist in the prognosis of Parkinson’s Disease. The app can help doctors both in the initial diagnosis of a patient’s symptoms and in monitoring a patient's response to the treatment regimen. The app will help quantify the intensity of hand tremors and identify the results will be displayed in numerical and graphical values (the measurements on X, Y, Z axis from smartphone's accelerometer sensor) as normal or abnormal based on the severity of tremors.

The app, Parkinsensor, focuses on assisting with early diagnosis of Parkinson’s Disease. Parkinson’s is a neurological disorder that leads to shaking, stiffness, and problems with balance and coordination. The most common symptom of Parkinson’s is tremors in the hands, arms, legs, jaws, or head. All symptoms of Parkinson’s are degenerative. Currently, there are no medical tests to diagnose Parkinson’s disease. If the disease is suspected, the doctor must monitor the patient’s symptoms over a period of time. Many cases, however, go undetected as people assume their symptoms as part of the aging process. Although there is no medical test to diagnose Parkinson’s Disease, doctors often monitor their patients’ symptoms. They also administer trial doses of Levodopa, a symptom-relieving therapy for Parkinson’s which supports the diagnosis. The main problem, though, is that many people do not report their symptoms. In order for Levodopa to be issued, the doctor must have a way to check their patient’s symptoms. This is where Parkinsensor comes into play.

Research

Parkinson’s Disease is a neurological disorder that leads to shaking, stiffness, and problems with balance and coordination. According to WebMD, the disease is caused by neuron deterioration in the substantia nigra portion of the brain. In a normal condition, these neurons produce dopamine, which allows for “communication between the substantia nigra and another area of the brain called the basal ganglia. This communication coordinates smooth and balanced muscle movement. A lack of dopamine results in abnormal nerve functioning, causing a loss in the ability to control body movements” [1]. The most common of Parkinson’s degenerative symptoms is tremors in the hands, arms, legs, jaws, or head. All symptoms of Parkinson’s are degenerative. Currently, there are no quantitative, efficient and easy to use medical tests to diagnose Parkinson’s disease. If the disease is suspected, the doctor must monitor the patient’s symptoms over a period of time. According to the Parkinson’s disease Foundation, about 10 million people are living with Parkinson’s disease worldwide and 60,000 people are diagnosed each year. The average age of onset for Parkinson’s disease is 60. Many cases, however, go undetected as people assume their symptoms as part of the aging process. Medical professionals also administer trial doses of Levodopa, a symptom-relieving therapy for Parkinson’s which supports the diagnosis. In order for Levodopa to be issued, the doctor must have a way to check their patient’s symptoms. The most important challenge is finding an objective method to analyze disease progression and treatment effectiveness through measuring the frequency and intensity of tremors [2]. To fill this gap, I've developed “Parkinsensor,” a smartphone application that focuses on assisting with early diagnosis of Parkinson’s disease. In today’s age, health applications are gaining more importance especially because smartphones allow for real-time measurement, data storage, and the ability to communicate easily and electronically. Previous research in quantitative assessment of Parkinson’s disease includes the use of laser-based displacement transducers and electromyography (EMG) as well as placing accelerometers on different parts of a patient’s body [3]. There has also been research on a portable Parkinsonian tremor assessment system [4]. Parkinsensor, an application for smartphones, brings many benefits to the medical field. For example, the wide and remote use of smartphones lets patients monitor their tremors anytime and anywhere. It also will allow doctors to efficiently monitor their patients’ symptoms and response to treatment. Furthermore, the U.S. National Institute of Health finds that early diagnosis and treatment of Parkinson’s results in better prognosis and patient quality of life, as well as potential treatment cost savings, leading to an overall improved patient outcome [5].

Method / Testing and Redesign

Materials and Methods

To allow for efficient, quantitative diagnosis of Parkinson’s disease, the smartphone application was developed on the Android platform using MIT App Inventor 2 and Java programming language. The application leverages the smartphone’s built-in accelerometer sensor for testing, graphing, and scoring the simulated tremors. This tri-axial accelerometer sensor measures static (gravitational) and dynamic (vibrating) acceleration in the X, Y, and Z axes. The accelerometer sensor’s range of measurement allows it to accurately and precisely detect the oscillatory Parkinsonian tremors. 

 

Application Features  

Parkinsensor is wireframed so that the layout of the screens is easy to navigate and retain information from. Upon initialization of the app, the user is shown a welcome in which they have two options, which are to visit the testing screen or the graphing screen. The screens are explained in greater detail below. All the data is based off of tests and graphs that were performed in a simulated environment. Figure 2 is the flowchart of the application design.

Welcome Screen

When Parkinsensor is initialized, the user is shown a Welcome Screen displaying the logo and an option to select the Testing Screen or the Graph Screen. Figure 3 is a screenshot of the app’s Welcome Screen.

Testing Screen

Parkinsensor uses the phone’s accelerometer sensor to measure tremors of the user’s hand in the X, Y, and Z axis with a unit of m/s2. Through multiple simulated tests, the data can be classified into two ranges:

●       Normal (no tremors): acceleration values in all three axes do not deviate from initial position by more than ±1 m/s2

●       Abnormal (tremors detected): any values outside of the normal range are considered abnormal

For each 3-second interval, data is recorded in a Google Fusion Table. The app uses 10 “trials” to determine a patient’s condition. It calculates a “score” out of 10. For each trial, the score stays constant and does not increase if a tremor is detected whereas a normal condition increases the score by 1. At the end of the 10 trials, if the score is 8/10 or below, the user is notified to seek help. If the score is above 8, the user is shown a “success” message. Figure 4 is a screenshot of the app’s Testing Screen while testing is in progress.

Graph Screen

Parkinsensor graphs the user’s tremors in the X, Y, and Z axes. A straight line (with less fluctuations) represents no or few tremors. Many fluctuations in the graph show more severe symptoms. Figure 5 shows a graph of normal condition tested in simulated environment. Figure 6 shows a graph of abnormal condition tested in simulated environment.

Sharing Data & Graph

On both the Testing and Graphing screens, the user can share test data and graphs through their phone’s email app.  Figure 7 shows screenshots of the results sharing screens.

 

Results

In this project, the data was represented by deviation from the initial (resting) position. Deviation was measured using the smartphone’s built in accelerometer sensor.

 If no tremors are detected, the user has a “normal” condition. If tremors are detected multiple times, the user has an “abnormal” condition. Through research of quantitative analysis of Parkinson’s and the simulated tests, I was able to determine the deviation value of ±1 m/s2. When the “Start Test” button is pressed on the app, the initial position is recorded in the database. Each subsequent trial of the ten total trials is compared to the starting position and if the X, Y, and Z values deviate by more than ±1 m/s2, the data is indicated as abnormal. If the values deviate by less than ±1 m/s2, the user’s condition is deemed “normal.” Having this “range” is vital because it gives patients and doctors a standard against which they can observe the intensity of tremors. Table 1 shows the sampling of data in both normal and abnormal simulated conditions. Figure below shows the plot of normal condition and the plot of abnormal condition.

The shown research data has added a new standard to measuring tremors in order to assess Parkinson’s disease. The use of technology allows doctors to achieve more accurate and numerical results in an efficient manner. Technology can remove subjectivity, which is found in qualitative tests. This will help doctors vary treatment regimens and patients will have a better prognosis of the disease as opposed to having a qualitative test in which doctors ask patients a series of questions and try to observe tremors. The use of similar technology in medical research can help to make advances in the way we diagnose and assess similar diseases.

 

Conclusion

The developed smartphone application successfully recorded, scored, and graphed the simulated hand tremors to determine Parkinson’s disease. The recorded graph clearly shows the difference between static and simulated tremor conditions and can thereby suggest a need for further medical intervention. In addition, the graph shows the extent or severity of hand tremors. The tremor history is recorded in a database which allows doctors to keep a historical account of their patients’ condition before, during, and after treatment. This may help doctors vary treatment regimen based on the measured results for early diagnosis and better prognosis for improved medical outcomes. Since the developed smartphone application uses phone’s in-built accelerometer sensor, it provides an inexpensive, efficient, and objective evaluation of Parkinson’s disease.

The current app design needs to undergo actual patient/clinical trials as per guidelines for medical apps so that doctors and patients can officially embrace the proposed device and technology and promote its use in the patient community. In some instances today, patients have to undergo multiple physical tests before the exact cause for the disorder can be determined. Various tremor patterns resulting from multiple causes can be incorporated in Parkinsensor so that medical professionals can pinpoint the precise cause of the disease. The developed application can be ported on smart wearable devices equipped with a similar accelerometer sensor. It is likely that the Parkinson’s patients may require constant monitoring of symptoms and the proposed application is designed to handle such conditions.

About me

I am interested in the applications of technology to make an impact in the community. I always viewed computing technology as a complex and intimidating field. At the end of my 6th grade, my participation in an Android development summer camp changed my perspective. Not only I was thrilled to learn about mobile technology, but I also explored some of its applications in solving real-world problems. One of my career goals is to merge two of my interests, computing and social activism. I hope to leverage information and communication technologies to empower civil society and promote social justice. I plan to explore how I can use technology to tackle issues at both community and global levels to advocate for fairness and inculcate empathy in our society. Wanting to help social justice campaigners use computing technologies more effectively, I would like to participate in projects focused on increasing the ability of advocates and activists to use digital technologies in many marginalised communities. The ubiquity of mobile phones means that many citizen have the power, if not always the means, to send or receive information directly via text message. This capability can help transform a global citizen to “witness” or even an informant – through crowdsourcing or crowdmapping. Creative use of technology can allow us to tackle the major challenges of our age: access to water, information, education, healthcare, and many others. Winning the Google Science Fair would be a major stepping stone to help me achieve my goals.

Health & Safety

I tested and modified the application by simulating tremors. I was able to adjust the intensity of the shaking to represent different severities of Parkinson’s Disease and made necessary modifications to my application to improve accuracy and precision. This enabled me to collect, graph, and analyze data for normal and abnormal conditions. The smartphone application was developed on the Android platform using MIT App Inventor 2 and Java programming language. The application leverages the smartphone’s built-in accelerometer sensor for testing, graphing, and scoring the simulated tremors.

Bibliography, references, and acknowledgements

References

1.      "Parkinson's Disease Foundation (PDF)." Diagnosis | Parkinson's Disease Foundation (PDF). Parkinson's Disease Foundation, n.d. Web. 05 July 2017.

2.      WebMD. “Frequently Asked Questions About Parkinson's Disease.” WebMD, WebMD, http://www.webmd.com/parkinsons-disease/guide/parkinsons-faq#1-5.

3.      Rigas, George, et al. “Assessment of Tremor Activity in the Parkinson’s Disease Using a Set of Wearable Sensors.” IEEE Xplore Digital Library, IEEE, 2012, http://ieeexplore.ieee.org/document/6121951/.

4.      Dai, Houde, et al. “Quantitative Assessment of Parkinsonian Tremor Based on an Inertial Measurement Unit.” Sensors (Basel, Switzerland), MDPI, Oct. 2015, https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4634500/.

5.      Pagan, F. L. "Improving Outcomes through Early Diagnosis of Parkinson's Disease." The American Journal of Managed Care. U.S. National Library of Medicine, Sept. 2012. Web. 05 July 2017, www.ncbi.nlm.nih.gov/pubmed/23039866.

6.      National Institutes of Health. "Parkinson's Disease." NIH Senior Health. National Institutes of Health, n.d. Web. 7 Apr. 2017.

7.      NINDS. "Parkinson's Disease Information Page." National Institutes of Health. U.S. Department of Health and Human Services, n.d. Web. 19 Apr. 2017.