This exploratory study looked at how well different methods could guess the mood and excitement level of a conversation. The research involved 51 pairs of people. Scientists measured heart rate, skin temperature, and breathing while also recording audio and video. They compared these physical signs against human ratings and the people's own written descriptions of how they felt.
The results showed that self-report questionnaires were the most accurate method, reaching 79.5% accuracy. Using human reviewers to watch the video and listen to the audio achieved 69.6% accuracy. When relying on the physical body signals like heart rate and skin temperature alone, the accuracy was 51.5%.
The study is considered exploratory because the team tested many different computer algorithms. Because the research is early and exploratory, these results should not be used to make medical decisions or change how people communicate. More research is needed to confirm these findings before they can be applied in real-world settings.