Enhanced Forehead EEG Using Peripheral Signals For Emotion Recognition In Wearable Devices
Daohong Wei, Dongyi Chen, Zhiqi Huang · 2024
Emotion recognition via EEG signals has garnered considerable attention with the rise of wearable computing technologies. However, EEG data's high dimensionality and lengthy computation times present challenges for real-time use. Moreover, the limited number of channels available on wearable devices complicates the task of achieving accurate emotion detection. To overcome these obstacles, this study introduces a novel approach that incorporates peripheral physiological signals to compensate for EEG in the forehead hairless region. Experimental results demonstrate that a 2-channel EEG setup from forehead positions get accuracy and F1 scores comparable to those from an 8-channel system covering multiple brain areas. The inclusion of anterior signals significantly enhances emotion recognition performance. This research offers valuable insights for future applications, highlighting the potential of multimodal physiological signals in reducing the need for extensive EEG channels without compromising, or even improving, accuracy. This development could pave the way for more practical, accessible solutions in areas like mental health monitoring, user experience optimization, and human-computer interaction.