Affective Response Dataset for Virtual Workspaces: Based on Color Stimuli and Multimodal Physiological Signals.

In the context of post-pandemic remote work normalization and the emergence of the metaverse, virtual workspaces have attracted significant attention as critical digital infrastructure with promising application prospects. While virtual workspaces enable efficient task performance, compared with traditional ones, the lack of emotional connection between humans and machines adversely affects participants' mental health. The emergence of affective computing has made it possible to endow virtual workspaces with "affective intelligence". Therefore, this study aims to clarify the relationship between color and participants' emotions in virtual workspaces through an experiment involving 48 participants, and eight virtual workspaces were constructed, incorporating four color conditions (red, blue, yellow, and green) and two workspace types (shared and single). Data were synchronously collected using the Positive and Negative Affect Schedule (PANAS), a questionnaire item on arousal, electrodermal activity (EDA), and heart rate variability (HRV). The results successfully established specific associations between colors and emotions: red with "anxious", yellow with "happy", and blue with "calm". Although no specific emotion word was identified for green, this study successfully achieved the emotion classification of virtual workspaces and constructed a corresponding dataset. These findings provide a theoretical foundation for the development of affective computing models.