Early Detection of Pre-Seizure using Wearables | Projects

 

During his time before arriving at White Box, Senior Data Scientist Mohamed Shakir worked on a number of groundbreaking projects. Integrating technology, data science and machine learning (to name a few), Mohamed produced devices and systems for stakeholders all over the world, including an embedded wearable system that uses Electroencephalography (EEG)(a complex method to monitor and record the electrical activity of the brain) in order to detect the signs of pre-seizure.

What was done?

The Early Detection of Pre-Seizure system

Using the pieces of equipment shown above to monitor the electrical activity of the brain, Mohamed was able to study likely pre-seizure symptoms using ‘control by thought’.

 The first experiment was to make sure he was reading the right signal from the brain. To prove this, he created an algorithm to connect to a wheelchair, controlled by the users thoughts. To connect to the users thoughts, a EPOC EEG headset was used to tap the real-time raw EEG brain wave signals. Then he visualized and processed information to clearly classify (using a Machine learning algorithm) various thought process of a human being in real-time to map it to actions in a wheelchair (front, back, right, left and stop)

After it was successful, another probe was used to tap seizure information and processed to predict seizure of human EEG.

A standard scientific data science research methodology was used to validate results and algorithm to scale it into an embedded wearable device which looked like the devices displayed in the images below.

example of the wearable pre-seizure detection embedded system

A great visual output of the results of the project is displayed in the below figure. These are called 3D Fuzzy Logic Decision Surfaces with (a) showcasing normal brain states, (b) showcasing a pre-seizure state, and (c) showcasing a seizure state.

3.png

What’s the relevance for me?

Control by thought is a complex machine learning process requiring a tonne of analysis and manipulation, coding and machine learning. However, most firms can benefit from less complex adoptions of machine learning to help them better understand their business and its customers, providing more valuable insights and inspiring better solutions. 

Data Science including data classification, data segmentation, and statistical analysis is becoming more and more vital as businesses incorporate Big Data Analytics to produce enhanced, sophisticated solutions for clients and customers. This particular project utilised MatLab numerical computing, PowerBI, and Tableau. These tools enabled the production of an Internet of Things (IoT) device reflected as a wearable system which enabled for advanced data visualisation and further research into a genuine problem experienced by many in society.

Most firms will not require the level of AI and machine learning complexity associated with this particular project. However, the data science processes as well as the incorporation of cloud computing and a customised range of software and tools optimal for the achievement of the goals of the project are all factors that can generate great ROI for any business wishing to unlock the potential of its own data.


If you would like help collecting and organising your data to create insightful and dynamic dashboards and reports, or to learn more about our specialities in AI, machine learning or data science, please get in touch today for more information on ways we can help you.

 

To keep up with all things data and White Box, follow us on our LinkedIn page.