Our brain regulates our emotions and how we feel, and those feelings can be detected by looking into brain activity. One way to do this is through electroencephalography (EEG), a technique that records brain signals from the scalp. These signals can help develop AI models to process EEG signals to recognize how someone is feeling.  

This capability has applications such as diagnosing mental health issues, evaluating student engagement and seeing how people respond to advertisements. However, EEG signals vary greatly between individuals, making it challenging to design models that work well for everyone. Until now, researchers have mainly addressed this issue through AI systems alone, while neuroscience techniques have been ignored.

A technique that can help in this regard is the inter-subject functional correlation (iSFC), which measures how similarly different people's brains respond to the same experience. By identifying these common brain activity patterns, iSFC can help train AI systems to recognize emotions in a more consistent and reliable way, even across people with very different brain activity. As a result, emotion recognition systems can become more accurate and widely applicable.  

These systems can help manage conditions like anxiety and mood disorders, which affect around 23 per cent of Manitoba's population. As a result, in today's AI-driven society, our work aims to build practical tools that make emotional support more accessible for everyone.