EliOS is a system for understanding, predicting, and collaborating to treat mood disorders.

Through brain scanning, mood tracking, and artificial intelligence, EliOS has the power to change what it means to have a mood disorders.

The system was built for (and named after) our wonderful friend Eli Scott. Eli is a student of Cognitive Science and has Bipolar disorder.

The idea of EliOS was born in a conversation about the role of technology in mental health. We realized that the systems that, at times, make social media like Facebook and Instagram invasive and manipulative could be turned around and used to understand and help people with their mood disorders.

There is also a compelling business case here: None of the mood tracker apps on the market actually helped Eli manage her condition. They didn't visualize the trends in her mood very effectively, let alone make predictions or recommendations. Furthermore, none of the apps had any way to make the process of building and reaching out to a support network any easier.

It also just so happens that the team had an interest in Neurotech, so we brought along a Muse EEG scanner. Why not incorporate brain scanning into the system to further optimize our predictions?


To make sure we met the needs of our stakeholders, we conducted a psychological & neurological literature review and a user consultation meeting. Our overall design goals were as follows:

  1. Make a better data entry system.
  2. Visualization data in insightful, interesting ways.
  3. Make useful predictions.
  4. Enable technology-assisted peer support system.

How It Works

User Experience

We spent a lot of time planning exactly how we would implement the idea because we knew user experience is central to the success of the product.

  • Important: Cross-platform:
  • Highly efficient UI you can tap through with ease
  • Automate as much as possible:
  • Minimalist design with clear, concise visualizations.
  • Make the process of getting help from support network as easy as possible


To accomplish the goal of creating an artificially intelligent system for mood tracking, brain scan analysis, and support system management, we used a lot of tools.

We decided to try Vue for the first time for our multi-platform web-app.

  • Vuex for state management
  • Vue-Charts for data visualization

We used Google's FireBase for our database, authentication, storage, and hosting, and NumPy, Pandas, and Pytorch for our signal analysis and ML.

Plenty of Jupyter Notebooks for testing out ML and signal analysis ideas, too.

We were also inspired to use Standard Library's text API to create the support system.

It can be very difficult to reach out for help when you have a mood disorder, so removing friction is fundamental.

Standard LIbrary's API helped us make it easier to reach out for those who use EliOS, and they made it easier for us to use awesome API's efficiently and effectively.

What We've Learned

Over the course of this project we picked up many new skills. Coming in, we had no experience with the Vue platform on which our application would be based, nor did we have experience interfacing brain scan data with the web at all. This posed many of its own challenging, from structural constraints to changing in the way that we develop.

We also encountered many challenges in the back-end working on the Machine Learning. Starting with the Digital Signal Processing of the EEG, we developed the new skills required to analyze brain wave data. We could then apply these skills to the Google Cloud, posting our machine learning server to be accessible by our application at a moment's notice.

Next Steps

Next steps for our project include: 1) A live data capture for EEG data within the application, 2) Deepening the learning models we use to predict mood swings as we gather more data 3) Integrating common smart technology data to get more info about the people we want to support. 4) Integrate Google Calendar so that the system can react to life events in real-time.

Built With

Share this project: