Plans for our proposed neural learning network
Our project began with the simple goal to create a tool that could assist students in one of their most pertinent issues- cultivating mental health. As students, we must constantly contend with high-intensity environments in which stress is continually a factor. With the understanding that stress is a key component of mental health, we decided that a quantitative view of an individual's specific stress level throughout the day could create a positive difference in one's stress management by allowing for greater awareness and encouraging de-stressing activities, ultimately resulting in lower stress levels and improved mental health.
What it does
Our system employs a Muse Headband (EEG) to measure brain activity as prompted from the GUI of a custom-designed computer application. Using an algorithm, the program interprets the brain activity and using research data determines from the wave intensities the level of stress an individual is experiencing at the current moment. This information is then displayed to the user through plain language as well as expressing ones specific quantity of stress graphically.
How we built it
Starting with only a Muse Headband, we developed a program that took the EEG input of the machine acquired over a test scan of 4 seconds (resulting in over 50,000 data points created!) using the headband's API to read brain activity data. Using research data we discovered on correlations between brain activity and stress, our program (built in Java) extrapolates this data into a quantified stress level. We then created a GUI in Java to allow the user to interact with and see the output of the program. Using a button interface for simplicity, we designed our program to take a brain scan of the user when prompted, turning the data into measured output which is then reported to the user in the form of language as well as graphically for ease of understanding.
Challenges we ran into
Throughout the course of the project, we were confronted with a wide range of problems to which we had to adapt our product. Originally designed to measure emotion, we concentrated the focus of our program to measuring stress after discovering the lack of available data on emotion versus brain wave activity hence preventing us from utilizing a neural network we had designed to determine specific emotions. We were also faced with the task of adapting to the Muse API which was both unfamiliar as well as often confusing, adding extra time to the project. Finally, our team had not had significant experience in creating a GUI until now, resulting in a steep learning curve as we learned the necessary skills in a limited amount of time.
Accomplishments that we're proud of
This project required us to adapt to new types of programming, and we are proud to have rapidly acquired proficiency in utilizing the Muse SDK as well as in creating an intuitive user GUI. Although we were not able to implement our neural network under the current research limitations, we were nonetheless proud of our initiative.
What we learned
We learned hard work and cooperation are the most important skills.
Look out for us in the headlines!