Facial obstructions and baseline correction shape affective computing's detection of emotion-behavior relationships

Area of Science:

• Psychology
• Computer Science
• Human-Computer Interaction

Background:

• Affective computing (AC) is increasingly used to study emotions in decision-making.
• Methodological validity of AC in capturing spontaneous emotions and their behavioral relevance is debated.
• The impact of facial obstructions and baseline correction on AC accuracy is unclear.

Purpose of the Study:

• To evaluate the predictive validity of AC-based facial emotion measures in charitable decision-making.
• To compare AC (FaceReader) with electromyography (EMG) and self-report (SR) measures.
• To examine the effects of facial obstructions and baseline correction on AC accuracy.

Main Methods:

• Participants (N=88) viewed dog images and made donations.
• Facial expressions were recorded using FaceReader (FR).
• Group 1 (n=43) also had EMG electrodes; Group 2 (n=45) did not. FR results were compared with EMG and SR measures, analyzing effects of obstructions and baseline correction.

Main Results:

• Facial obstructions (EMG electrodes) reduced AC accuracy.
• Baseline correction improved convergence between FR and SR emotions and strengthened emotion-donation links.
• Stronger negative emotions correlated with higher donations; FR valence and happiness linked to donations, especially without obstructions.

Conclusions:

• AC can capture emotion-behavior relationships comparable to EMG and SR under optimal conditions.
• Facial obstructions decrease AC accuracy, while baseline correction enhances it.
• Methodological factors like obstructions and baseline correction are critical for multimodal emotion research.