Since the COVID-19 pandemic, college students have experienced notable changes in academic performance, social connection, and personal well-being closely tied to stress. We set out to develop a more accurate method for measuring student stress levels by creating a system that collected brain signals using wearable sensors and utilized artificial intelligence to analyze these signals and classify stress levels and causes in users.
We used a headset to read electrical brain activity and blood oxygen signals from test subjects. These signals, electroencephalogram (EEG) and functional near-infrared spectroscopy (fNIRS), provided valuable insight into subjects’ mental states. We cleaned the raw signals to remove excess noise and errors, then analyzed the resulting signals and used them as inputs for our machine learning model, which classified users as either stressed or relaxed.
Our system correctly identified stress in most test cases, allowing for quick assessment of someone’s current mental state. This project demonstrates how signal processing and machine learning could be combined to create an effective classification system. Our final design allows for future expansion to support real-time usage and wearable health applications, and the project could serve as a test bed for further research in student stress.