Composition 2 by Piet Mondrian, 1922
Anima joined the InnovAID hackathon, an international event with teams from the Netherlands, Portugal, and Ukraine. Their goal? To create AI-based solutions for various healthcare challenges. Anima’s innovative work was right at the center of this event. Two teams who worked on challenges related to Anima’s field won the top two spots and were highly praised by the judges. Other teams worked on different challenges, like AI tools for heart disease, women’s health, and emotional disorders.
The teams from the Netherlands and Ukraine used Anima’s anonymized data to build models that can spot depression. They focused only on how eyes move and reached an 80% success rate in distinguishing between minimal and severe depressive symptoms. All this was accomplished in under two days, with time being a significant bottleneck.
The Dutch team’s model focused on how fast the eyes move, how long they look at images, how often they switch between images, and the overall distance the eyes travel across the screen. For detecting depression, they used a complex algorithm that blends various data analysis methods. The Ukrainian team had a different approach. They showed their data using something called spectrograms and also used deep learning models. This made their way of finding patterns even more trustworthy.
Spectrogram graph for each level of depression severity
The Anima team didn’t give their solutions to the participants, but they let them use real user data for practice. This was a big help in tackling real-world problems. After both teams won, they shared what they made with us, giving us new insights into detecting depression.
In the end, these teams created algorithms that can use eye movement to detect depression with great accuracy. This achievement is significant because it highlights the importance of mental health and shows how advanced AI models can help in diagnosing mental health conditions more accurately. This aligns with Anima’s goal of not just identifying but also predicting mental health issues, which can lead to timely and life-saving treatments and improve the quality of mental healthcare overall.
References
