End to End Framework for Neurological Health Monitoring

The prevalence of neurological disorder has drastically increased over the years resulting in burdening the economic growth of developing countries. One main reason can be the lack of facilities and attention given to the health sector in Pakistan. A recent study has shown that on average 13000 peop

Project Title

End to End Framework for Neurological Health Monitoring

Project Area of Specialization

NeuroTech

Project Summary

The prevalence of neurological disorder has drastically increased over the years resulting in burdening the economic growth of developing countries. One main reason can be the lack of facilities and attention given to the health sector in Pakistan. A recent study has shown that on average 13000 people commit suicide in Pakistan and 96% are disease related due to mental disorders. 

it had been established in research that most neurological disorders result in significant change in the patterns of activities performed by individuals in daily life such as eating, drinking, sleeping, indoor and outdoor activities etc. Therefore, it is important to develop an end to end system that distinguishes a pattern in everyday life, with the help of human activity recognition and detects neurological disorders through miniaturized Inertial Measurement Units (IMUs) such as smart watches and mobile phones.

Our project aims to develop such framework to collect data from IMUs and upload it on a remote server, where appropriate data processing is performed in-order to detect abnormalities. We also develop a dedicated mobile application that helps in collaborative communication between the patient, caregiver and doctors so that in case of emergency they can be immediately notified.

The embedded system saves the data in the cloud through wireless network where an appropriate machine learning algorithm is applied to detect neurological disorders. The developed mobile application provides an appropriate interface to the subject and the caregiver for self-analysis and event reporting. We wish to provide scalable, affordable, robust, dependable and fast solution to this problem.

Project Objectives

The main objectives of our project are:

1. Sensor data collection for different Activities of Daily Living (ADL)
2. Application of appropriate signal processing techniques for filtration and initial processing of sensed data
3. Identification of pattern associated with daily living such as walking, running, driving, eating, using appliances and sedentary activities etc.
4. Long term monitoring of abnormalities in ADL as early warning system for various neurological disorders
5. Development of remote database and its linking with mobile application
6. Conducting the survey for software Quality of Experience (QoE) and improving MyNeuroHealth app for wide scale deployment.

Project Implementation Method

We used both pulse sensor and accelerometer for data collection. We integrated the pulse sensor with the help of arduino and sent the sensed data to the application over Bluetooth module HC-05. There are built-in accelerometers in the mobile phone so we have used the one embedded inside the phones inorder to collect data.

We have done both time and frequency domain analysis on the sensed data using Matlab. First we linearized the data in order to refine our results. Whenever the subject wearing the device is involved in an activity, the resulting dataset in time domain analysis is tested for zero crossing.The number of oscillations in the signal is known as number of zero crossings. The person wearing the device and performing certain activity can be predicted by the same pattern of zero crossing. For time division, data is calculated for zero crossing, mean, variance and standard deviation.

images/End to End Framework for Neurological Health Monitoring _1582921436.

After processing the data we were able to classify activities of daily living into 4 categories that are 

1. Stationary Class
2. Light Ambulatory Class 
3. Intense Ambulatory Class 
4. Abnormal Class

These classes further defined wether the subject was sitting ideal or jogging or is in need of emergency help. 

https://www.dropbox.com/sh/q9a6f0ldbr8vr0w/AAAEt1uxBEIkAyd5x4_Tt2sma?dl=0

The project is based upon any subjects’ involvement in different kinds of Activity of Daily Living. These activities are then identified in reference to movement parameters which classify them from one another. Feature identification includes the data collected through the sensors. The project is using the sensor, Accelerometer the movement of the patient, heart rate sensors and temperature sensors to monitor the changes. The data output from these sensors are converted to CSV file format which are accessible by MATLAB. We used accelerometer from the smart phone device to receive the data at any particular time generated in regard to the subjects’ movement

https://www.dropbox.com/sh/pagku4iaw32d0hw/AADwLgA7JpWJXSgSGwKoCwHia?dl=0

We developed a remote database using MySQL and Firewall and integrated it with android application. The database creates individual entries for each subject and records its respective data according to the activities performed by them. 

https://www.dropbox.com/sh/dy02ms92dledcye/AABitqHf0F6fMkL9TWBsaA7ra?dl=0

We will conduct surveys and distribute question ires to gain feedback about the product. This will help us in regularly updating the application with respect to user requirements and make it more user friendly.

https://www.dropbox.com/sh/e1qfm88w38m95t8/AACAJGtpV-1oef5u56sQi96sa?dl=0

We have a Facebook page where you can further view details about MyNeuroHealth. 

https://www.facebook.com/MyNeuroHealth/

Benefits of the Project

Neurological disorders have continued to hinder the daily life of the patient and hence effected his/her overall performance.With the help of our application the user can not only monitor its activities of daily living but will also be able to tell if it deviates form its daily routine, it will be able to detect the neurological disorder the patient might be leading towards. The results of using this application can be fruitful because beforehand you will be informed of the type of disorder you may encounter and help monitor healthy day to day routine. 

Neuropsychiatric disorders affect the functionality and wellbeing of individuals and pose an economic burden for society. These disorders are not only costly with respect to treatment and other services related to disorders, including special accommodation, social services and informal care, but even more so when considering the indirect costs, such as lost income and productivity owing to not finishing school or reaching only low grades, inability to train for a job, lower academic achievement, absence from work or early retirement. Improving functionality of patients suffering from neuropsychiatric disorders will lead to a reduction of the costs of these disorders for society and governments in the range of about $90 billion in Pakistan.

By introducing this idea in the market can be favorable because now majority of the people own smart phones so our application is easily accessible. Being a free and a user friendly application it can be easily used by a majority of people.

Long term advantages of using this application will help improve the performance of an individual as well as the society as they can themselves see and eliminate those poor habits from their life and hence lead a more healthy and prosperous life.

Technical Details of Final Deliverable

Hardware requirements for this project consists of an accelerometer sensor that is mostly built in installed in smart phones devices these days. The sensor is at first calibrated and tested with different activities. The output of the sensor is then used as data input to the mobile application. Mobile GPS sensor defines the position of the subject and helps the caregivers locate the subject in need. Some of the other sensors such as embedded heart rate sensors would be quite useful to achieve accurate results in terms of activity recognition.

Another hardware requirement is the pulse sensor, Bluetooth module (hc-05) and Arduino UNO. We assembled this hardware in order to integrate another sensor with our application so that we can accurately identify irregularities in activities of daily living. Heart Rate is a sensitive parameter that is bound to change while performing different activities of everyday life. 

An android application is designed to monitor health changes which are then linked with raw-data from accelerometer. Users have to register to the mobile app which helps to store these readings in the cloud based database. Each user has a unique id to differentiate and keep the records intact. The sensor takes readings for every 30 seconds and deduces the type of activity the person was involved in. In case of identification of abnormal activity the emergency call is made to selected emergency numbers and the caregivers are notified via SMS. The subject is prompted if there was incorrect abnormal behaviour detection which can prevent giving trouble to the concerned people. In cases other than abnormal activity, the person is not notified but the response is uploaded on the database. Similarly the record is updated for each session and the user can monitor these activities through application as well.

We are using MySQL in our application after buying the web hosting package containing domain name with high bandwidth facility. The code for the database is written in a server side programming language PHP.It is then tested using an API testing tool Postman, to verify the connection establishment between PHP script and server. There are number of ways to set up connection between tested API and the java class such as GET and POST method. Upon connection build up the user is registered and logged in to their respective account. The table structure for a “users_table” contains the primary user information such as ID, Name, Email and password. Similarly, these fields are given the data types and their desired lengths for each record. Subject’s ID is used as the primary key and incremented automatically upon entering a new record.

Final Deliverable of the Project

HW/SW integrated system

Type of Industry

Medical , Health

Technologies

NeuroTech, Wearables and Implantables

Sustainable Development Goals

Good Health and Well-Being for People

Required Resources

If you need this project, please contact me on contact@adikhanofficial.com