Health literacy of the higher education community in a European country: a cross-sectional survey.

Stress recovery is a central topic in environmental psychology because everyday environments can either intensify or alleviate psychological and physiological strain. Restorative environment research has shown that natural environments can support affective, attentional, physiological, and neurophysiological recovery, as explained by Stress Recovery Theory and Attention Restoration Theory (; ; ; ; ; ; ). This evidence has been supported by studies using photographs, simulated nature, and virtual reality environments, with outcomes including mood, perceived restoration, cardiovascular responses, skin conductance, and EEG activity (; ; ; ; ; ; ; ; ; ; ; ). University-aged young adults are a relevant group for such research because they often face academic, employment-related, and social pressures (; ). However, much of this work still focuses on broad environmental categories, such as natural versus urban scenes, or on the general presence of natural elements. Less attention has been given to where specific landscape components appear within the visual field. Rural landscapes are a meaningful but underexplored context for restorative environment research. Unlike many urban green spaces or single-category natural scenes, rural scenes often combine water bodies, forests, agricultural fields, roads, buildings, and settlement edges within the same visual field. Previous studies have suggested that rural natural perception, place attachment, vegetation, blue–green elements, and landscape coherence may contribute to restorative experience, preference, and emotional responses (; ; ; ; ). However, rural landscapes should not be treated simply as less urbanized natural environments. Their restorative potential may depend on how ecological, agricultural, and residential elements are visually organized. Existing restorative landscape studies have mainly compared broad environmental categories or the presence of specific elements, such as water, vegetation, buildings, or roads (; ). Recent research has further shown that visual composition and vegetation proportion can be associated with attention and attractiveness ratings, although neural responses may be weaker or more complex (). Unlike , which examined urban green-space composition and affective responses, this study uses depth-controlled AI-generated rural scenes to examine stress recovery and exploratory foreground–midground associations. These findings suggest that composition matters, but most studies still describe composition through element presence, semantic class, or image proportion. Less attention has been given to the depth position of landscape elements. A water body in the foreground may affect texture, brightness, accessibility, and visual preference, whereas vegetation, fields, or buildings in the midground may contribute more to spatial coherence, enclosure, extent, and the sense of being away. Therefore, a critical unresolved question is not only whether a rural scene contains restorative elements, but how these elements relate to responses when they appear in different visual depth layers. Recent work has increasingly used controlled images, virtual environments, and computationally generated scenes to examine environmental perception and affective responses. is particularly relevant because it used image-based and computational approaches to evaluate affective responses to urban green spaces. It shows the growing value of scalable and controlled visual stimuli for landscape perception research. However, the present study differs from that work in three main respects. First, it focuses on rural scenes in which ecological, agricultural, and settlement elements co-occur, rather than on urban green spaces. Second, it embeds controlled stimuli in a stress-induction and recovery paradigm, rather than focusing only on affective evaluation. Third, it uses depth-estimation-guided editing to examine exploratory foreground, midground, and background-context associations. Therefore, this study does not claim that AI-generated landscape evaluation is entirely new; rather, it extends related work by linking controlled rural scene construction with physiological, EEG-based, and preference responses from a visual-depth perspective.