Multimodal Analysis of Emotional and Visual Responses to Urban Green Spaces in Virtual Environments

Abstract

Urban green spaces (UGSs) play a vital role in promoting emotional well-being and supporting public health. Emotional responses to landscapes are closely linked to visual behavior, reflecting how individuals engage with environmental features. This study proposes and applies a multimodal sensing framework to examine how the types and proportions of landscape elements influence emotional and visual responses in immersive virtual environments. Sixty participants experienced 360° panoramic simulations of three UGS types—urban parks, community gardens, and ecological wetlands—featuring seven landscape elements: trees, shrubs, flowering trees, grass, herb plants, water landscapes, and artificial objects. Data were collected using electrodermal activity (EDA), pulse rate (PR), eye-tracking metrics, and self-reported emotional ratings. Visual behavior indices were analyzed in relation to landscape composition. Results show that both visual behavior and emotional responses were significantly influenced by landscape elements. Herb plants enhanced attention and positive emotions across compositions. Flowering trees exhibited a dual effect—eliciting physiological arousal while promoting relaxation. Water landscapes captured immediate attention and evoked positive emotional experiences, whereas artificial objects required greater visual effort but yielded limited emotional benefits. These findings provide actionable, data-driven insights for the design of emotionally supportive UGSs in high-density environments.