EMOTION RECOGNITION FROM ELECTROENCEPHALOGRAPHY USING DEEP LEARNING: A SINGLE‑SUBJECT STUDY

This paper investigates the use of artificial intelligence for emotion recognition from EEG signals collected with a low‑cost headset in ecologically valid conditions. We adopt a design with substantial variability of EEG data, which is a large, multisubject dataset. By concentrating on one individual, we aim to build a highly accurate, personalized AI model that predicts that person’s emotional states. Overall, this paper demonstrates that precise, reliable, and personalized emotion recognition is achievable even with a lightweight headset and real‑world protocols. This focused approach provides a practical springboard for scaling to larger populations and integrating emotion detection into diverse human-computer interaction (HCI) modules.